UNIVERSITAS INDONESIA

EFFECT OF PERCEIVED PERSONALIZATION AND SELF

EXPRESSIVE BRAND TOWARDS PERCEIVED QUALITY AND
BRAND LOYALTY OF LOCAL FOOTWEAR BRANDS ON
INSTAGRAM

UNDERGRADUATE THESIS

ADHIWIRA INDRABRATA
1606898192

FACULTY OF ECONOMICS AND BUSINESS

INTERNATIONAL UNDERGRADUATE PROGRAM
DEPOK
JULY 2020
 UNIVERSITAS INDONESIA

EFFECT OF PERCEIVED PERSONALIZATION AND SELF

EXPRESSIVE BRAND TOWARDS PERCEIVED QUALITY AND
BRAND LOYALTY OF LOCAL FOOTWEAR BRANDS ON
INSTAGRAM

UNDERGRADUATE THESIS
Submitted as one of the requirements to obtain the title of Sarjana
Ekonomi

ADHIWIRA INDRABRATA
1606898192

FACULTY OF ECONOMICS AND BUSINESS

INTERNATIONAL UNDERGRADUATE PROGRAM
DEPOK
JULY 2020
 STATEMENT OF ORIGINIALITY

This thesis represents my own work and any idea or except from other writers in
this paper have been acknowledged in accordance to academic standards and
reference procedures.

Name : Adhiwira Indrabrata

NPM : 1606898192

Signature :

Date : 8 July 2020

ii
Universitas Indonesia
 VALIDATION PAGE

This Undergraduate Thesis is proposed by:

Name : Adhiwira Indrabrata

Student Number : 1606898192

Major : International Business Management

Thesis title : “ Effect of Perceived Personalization and Self-Expressive

Brand toward Perceived Quality and Brand Loyalty of
Local Footwear Brand on Instagram”

This thesis was successfully defended in front of the Board of Examiners and
was accepted as part of the requirement to obtain the title of Sarjana Ekonomi
from the International Undergraduate Program, Faculty of Economics and
Business Universitas Indonesia.

BOARD OF EXAMINERS

Thesis Advisor : Dr. Ir. Tengku Ezni Balqiah, M.E. (.....................)

Examiner 1 : Prof. Dr. Adi Zakaria Afiff, S.E., M.B.A. (.....................)

Examiner 2 : Ira Iriyanty S.E., M.S.M. (.....................)

Established in : Depok

Date : 8 July 2020

iii
Universitas Indonesia
 ACKNOWLEDGEMENTS

Praise the authors pray to Allah SWT, because thanks to Allah SWT guidance and
grace the writer can finish this thesis well. The topics chosen regarding the context
of personalised advertisement discussed in this undergraduate thesis was chosen
because of the author personal interest and enthusiasm for the current technological
developments in social media and his high enthusiasm for the current development
of fashion brands in Indonesia, particularly footwear brands. The researcher hopes
this thesis can contribute and be implemented within the related industries. Then,
the writing of this thesis was carried out in order to fulfil one of the requirements
to obtain Sarjana Ekonomi degree from the International Undergraduate Program,
Faculty of Economy and Business Universitas Indonesia. From the very beginning
of this thesis writing process, the author has received a lot of support, assistance,
and guidance from various people. Therefore, the researcher would like to
acknowledge and thank:

1) Beloved family, Ayah, Ibu, Gita, Aila that always support and pray for the
writer’s success upon pursuing his study.
2) My thesis advisor, Dr. Ir. Tengku Ezni Balqiah, M.E. that always guide and
give constructive feedbacks during the writer’s process of finishing this
undergraduate thesis and furthermore insights that the writer’s learn during
consultation session and writing process of this undergraduate thesis.
3) This thesis board of examiners, Ms. Ira Iriyanty S.E., M.S.M. and Prof. Dr.
Adi Zakaria Afiff, S.E., M.B.A. for the constructive critics and feedbacks
that the writers acquire to improve this undergraduate thesis.
4) All the lecturers and academic staff in FEB UI that give the writer valuable
knowledge and experience during the writer study in FEB UI. And, to Mr.
Rahmat Aryo Baskoro, S.E., M.M., CFP., AEPP as the writer’s academic
advisor and head of Fastur Unit in FEB UI that constantly giving his time
and help during the writer’s study and time in Indonesia Capital Market
Student Studies and Management Student Society FEB UI.

iv
Universitas Indonesia
5) Mr. Fahrudin and all International Undergraduate Program academic staffs
that helped the writer on academic and the writing process of this
undergraduate thesis.
6) Natashah Amellie that always give the writer positive influence and
encourage the writer to be a better person each day. Thank you for being
here, for being my support system, thank you for everything.
7) Fellow brothers within Yakuyza Kukusan, especially Christo Pierre, Arif
Yosia, Yusuf Rabiultsani, Rangga Yusuf, and Adek bagus Rangga that
creates so much memories during the writer’s study in FEB UI and the
writing process of this undergraduate thesis feel unique and burden free.
8) Beloved “Fight Club” members that have been “fighting” together since the
days of putih abu-abu, especially Fatur and Arwan my fellow FEB UI
brothers, Deno, Eky and Faiz, Falah, Gagas, Faiz, and Ibnu. Thank you for
all the support and laughs.
9) Christo Pierre and Arif Yosia that helped the writer on so many things in
life and especially constructive advice and support during the writing
process of this undergraduate thesis and moreover the preparation during
thesis defence.
10) Fellow friends in Bybyt Unggul , thank you for all the unique and fun
memories during the writer’s study in FEB UI.
11) Irfan Sutarsa, Fitra, Fadli, Aufar, my friends since junior high school, thank
you for all the support, friendship, and the laughs.
12) LDR Group especially Sukit, Kura-Kura, Opung, and Bocil, thank you for
this surprising yet beautiful friendship that we create and all the laughs,
memories, the writer wishes this friendship will last despite of its name.
13) Mergers group, Andrea, Atikah, Irsa, Jeje, Hafizh, Rifqi, Aryo, Falah, Ariq
and Arwan, thank you for all the support and help since high school up until
now.
14) Korea Simple House members, Qidsa, Ewing, Reynaldi, Sultan and
Ibrahim. Thank you for the knowledge, laugh, and experience during the
writer’s study in FEB UI and times in Hanyang University, South Korea.

v
Universitas Indonesia
15) Operation Division The 16th Indonesia Capital Market Student Studies
(ICMSS), Adek Bagus, Westy, and Hilmy. Thank you for the valuable
experience, support and friendships that the writer gets to experience.
16) Fellow directors of Operation Division The 17th Indonesia Capital Market
Student Studies (ICMSS) especially Risa, Rangga, and Arman, and
associates, Favian, Tebe, Rani, and Robo, thank you for all the experience,
laugh, tears during those memorable times.
17) Project Division of Management Student Society FEB UI, especially Yusuf
and Sultan, Leandra, Aurora, and Naufal. Thank you for the valuable
experience.
18) Fellow internship friends in Bank Indonesia especially Aldi, Bariq, Kamil,
Fadiah, Fatur, and Galang which made the internship time very interesting
and unique.
19) Dr. Trang Phuc Tran, one of the authors of this undergraduate thesis
reference journal, thank you for the insights and help.
20) Fellow friends on Faculty of Economics and Business Universitas
Indonesia especially batch 2016. It is an honour to get to know all these
superb people that will be leading this nation later in the future. Thank you
for all the experience that can never be replaced.
21) With all due respect, the writers would like to acknowledge and thank all
the parties that cannot be mentioned each one that gave support during the
writer’s effort in finishing this undergraduate thesis.

Depok, 8th of July 2020

Adhiwira Indrabata

vi
Universitas Indonesia
APPROVAL PAGE OF FINAL ASSIGNMENT PUBLICATION
FOR ACADEMIC PURPOSES

As a Universitas Indonesia civitas academica, I the undersigned,

Name : Adhiwira Indrabrata

Student ID (NPM( : 1606898192

Study Program : International Business Management

Department Faculty :International Undergraduate Program Faculty of

Economics and Business

Type of Work : Undergraduate Thesis

For the development of knowledge and science, I hereby agree to give to

Universitas Indonesia a non-exclusive royalty-free right to my academic paper
entitled:

“EFFECT OF PERCEIVED PERSONALIZATION AND SELF-EXPRESSIVE

BRAND TOWARDS PERCEIVED QUALITY AND BRAND LOYALTY OF
LOCAL FOOTWEAR BRANDS ON INSTAGRAM”

Along with the related materials (if necessary). With the non-exclusive royalty-free
right, Universitas Indonesia has the right to store, format/reformat, manage in
database form, maintain, and publish this undergraduate thesis under the condition
of listing my name as the writer and the copyright holder.

I hereby declare that the above statement is true.

Created in : Depok

Date : 8 July 2020

Declared by :

(Adhiwira Indrabrata)

vii
Universitas Indonesia
 ABSTRACT

Name : Adhiwira Indrabrata

Study Program : International Business Management
Title : Effect of Perceived Personalization and Self- Expressive
Brand towards Perceived Quality and Brand Loyalty of
Local Footwear Brands on Instagram
Counsellor : Dr. Ir. Tengku Ezni Balqiah, M.E.

Along with the increasing level of internet penetration in Indonesia, the current
developments encourage people to familiarize and adopts variety of internet-based
products and service especially Instagram as social media in Indonesia. Amidst
these opportunities, businesses that are particularly preferred in current lifestyle like
local footwear brand can take advantage by being present on Instagram to foster
their relationship with their consumers through personalised advertisement. This
quantitative research is constructed to analyse the effectiveness of personalised
advertisement by proposing key driver namely perceived personalization and self-
expressive brands that predicts consumer brand engagement, brand attachment and
its effect towards perceived quality and brand loyalty. A total of 386 respondents
that are Instagram users and consumers of four local footwear brand were collected
using purposive sampling method and the data was processed using Partial Least
Squares – Structural Equation Modelling (PLS-SEM). The findings suggest that
perceived personalization and self-expressive brand within personalised
advertisement affect the consumer’s brand engagement and brand attachments.
Whereby, consumer brand engagement and brand attachment play an important role
on affecting the perceived quality. Lastly Brand attachment and perceived quality
also found to influence the consumer’s loyalty toward a brand.

Key words: Personalised advertisement, Consumer Behaviour, Marketing

Communication, Perceived Personalization, self-expressive brand, perceived
quality, brand loyalty

viii
Universitas Indonesia
 ABSTRAK

Nama : Adhiwira Indrabrata

Program Studi : International Business Management
Judul : Analisis Pengaruh Perceived Personalization dan Self-
Expressive Brand terhadap Perceived Quality dan Brand
Loyalty Pada Merek Sepatu Lokal di Instagram
Pembimbing : Dr. Ir. Tengku Ezni Balqiah, M.E.

Seiring dengan meningkatnya penetrasi internet di Indonesia, peningkatan tersebut

juga turut mendorong orang-orang di Indonesia untuk menggunakan berbagai
macam produk dan layanan berbasis internet khususnya media sosial seperti
Instagram. Di tengah peluang yang tersedia ini, bisnis yang sekarang sedang
diminati untuk menunjang gaya hidup seperti merek sepatu lokal dapat
memanfaatkan peluang tersebut dengan menggunakan fitur iklan yang
dipersonalisasi yang ada di Instagram. Penelitian kuantitatif ini dilaksanakan untuk
menganalisis pengaruh iklan yang dipersonalisasi melalui variabel perceived
personalization dan self-expressive brand yang mana diprediksi dapat
mempengaruhi perceived quality dan brand loyalty merek sepatu lokal yang
beriklan di Instagram. Sebanyak 386 responden yang merupakan pengguna
Instagram dan juga konsumen dari empat merek sepatu lokal diperoleh
menggunakan metode purposive sampling dan dianalisis menggunakan permodelan
Partial Least Squares – Structural Equation Modeling (PLS-SEM). Hasil penelitian
ini menunjukkan bahwa perceived personalization dan self-expressive brand dalam
konteks iklan yang dipersonalisasi berpengaruh terhadap consumer brand
engagement dan brand attachment yang dimana kedua faktor ini juga
mempengaruhi perceived quality, dan juga hasil penelitian ini menunjukkan brand
attachment dan perceived quality mempengaruhi brand loyalty.

Kata Kunci: Personalisasi Iklan, Perilaku Konsumen, Marketing Communication,

Perceived Personalization, self-expressive brand, perceived quality, brand loyalty

ix
Universitas Indonesia
 TABLE OF CONTENTS

STATEMENT OF ORIGINIALITY ................................................................. ii

VALIDATION PAGE....................................................................................... iii
ACKNOWLEDGEMENTS .............................................................................. iv
APPROVAL PAGE OF FINAL ASSIGNMENT PUBLICATION FOR
ACADEMIC PURPOSES ................................................................................ vii
ABSTRACT .................................................................................................... viii
ABSTRAK ......................................................................................................... ix
TABLE OF CONTENTS ................................................................................... x
LIST OF TABLES .......................................................................................... xiv
LIST OF FIGURES ......................................................................................... xv
LIST OF ATTACHMENT ............................................................................. xvi

CHAPTER I ....................................................................................................... 1
INTRODUCTION ............................................................................................. 1
1.1 Research Background ........................................................................... 1
1.2 Problem Definition ............................................................................... 9
1.3 Research Purpose ............................................................................... 10
1.4 Research Implications......................................................................... 10
1.5 Research Scope .................................................................................. 11
1.5.1 Research Respondents .................................................................. 11
1.5.2 Geographical Scope of Research ................................................... 11
1.6 Research Structure .............................................................................. 12

CHAPTER II ................................................................................................... 14
LITERATURE REVIEW ................................................................................ 14
2.1 Consumer Behaviour and Decision Making ........................................ 14
2.2 Marketing Communication ................................................................. 16
2.2.1 Advertising ................................................................................... 18
2.3 Digital Marketing ............................................................................... 19
2.4 Social Media and Instagram ............................................................... 20
2.4.1 Personalised Advertisement on Instagram ........................................... 20

x
Universitas Indonesia
 2.5 Brand ................................................................................................. 21
2.6 Perceived Personalization ................................................................... 23
2.7 Self-Expressive Brand ........................................................................ 24
2.8 Consumer Brand Engagement ............................................................ 25
2.8.1 Cognitive Processing (Cognitive).................................................. 27
2.8.2 Affection (Emotional) ................................................................... 27
2.8.3 Activation (Behavioural)............................................................... 27
2.9 Brand Attachment............................................................................... 28
2.9.1 Affection ...................................................................................... 29
2.9.2 Connection ................................................................................... 30
2.9.3 Passion ......................................................................................... 30
2.10 Perceived Quality ............................................................................... 30
2.11 Brand Loyalty..................................................................................... 31

CHAPTER III .................................................................................................. 33

RESEARCH METHODOLOGY .................................................................... 33
3.1 Research Model .................................................................................. 33
3.2 Research Hypotheses ................................................................................ 35
3.2.1 The Influence of Perceived Personalization on CBE and BA......... 35
3.2.2 The Influence of Self-Expressive Brand on CBE and BA .............. 36
3.2.3 The Influence of Consumer Brand Engagement on PQ and BL ..... 37
3.2.4 The Influence of Brand Attachment on PQ and BL ....................... 38
3.2.5 The Influence of Perceived Quality on BL .................................... 38
3.3 Data Collection Method ...................................................................... 39
3.4 Research Design ................................................................................. 40
3.5 Variable Operationalization ................................................................ 41
3.6 Questionnaire Development................................................................ 50
3.7 Sampling ............................................................................................ 52
3.8 Data Analysis Method ........................................................................ 52
3.8.1 Wording Test Questionnaire Analysis ........................................... 52
3.8.2 Pre-test Questionnaire Analysis .................................................... 53
3.8.3 Frequency Distribution Analysis and Descriptive Statistics ........... 53
3.8.4 Structural Equation Modelling (SEM) with PLS-SEM .................. 54

xi
Universitas Indonesia
 3.8.4.1 Hierarchical Component Models: CBE and BA ..................... 56
3.8.4.2 Measurement Model Analysis (Outer Model) ....................... 57
3.8.4.3 Structural Model Analysis (Inner Model) .............................. 60

CHAPTER 4..................................................................................................... 63
ANALYSIS AND DISCUSSION ..................................................................... 63
4.1 Wording Test and Pre-test Questionnaire Analysis ............................. 63
4.2 Frequency Distribution Analysis......................................................... 66
4.2.1 Gender .......................................................................................... 67
4.2.2 Age ............................................................................................... 67
4.2.3 Level of Last Pursued Education or Current Pursued Education .... 68
4.2.4 Employment Status ....................................................................... 69
4.2.5 Domicile ....................................................................................... 70
4.2.6 Monthly Spending ........................................................................ 71
4.2.7 Marital Status ............................................................................... 72
4.2.8 Purchased Brands ......................................................................... 72
4.2.9 Recently Bought Brands ............................................................... 73
4.2.10 Length as Followers of the Brand Instagram Accounts .................. 74
4.2.11 Social Media Used ........................................................................ 74
4.2.12 Average Time Spent on Social Media per Day .............................. 75
4.3 Descriptive Statistics .......................................................................... 76
4.4 Measurement Model Analysis (Main-Test) ......................................... 81
4.4.1 Reflective Measurement Model Evaluation ................................... 81
4.4.1.1 Internal Consistency Evaluation ............................................ 82
4.4.1.2 Convergent Validity and Discriminant Validity Evaluation ... 83
4.5 Structural Model Analysis .................................................................. 88
4.5.1 Collinearity Evaluation ................................................................. 88
4.5.2 Coefficient of Determination (R2) Evaluation ............................... 89
4.5.3 Predictive Relevance (Q2) Evaluation ........................................... 90
4.5.4 f 2 Effect Sizes, q2 Effect Sizes, & SRMR Evaluation .................... 91
4.5.5 Size and Significance of Path Coefficient (Direct Effect) Evaluation
93
4.6 Hypotheses Testing Result .................................................................. 95

xii
Universitas Indonesia
CHAPTER V .................................................................................................. 105
CONCLUSION .............................................................................................. 105
5.1 Conclusion ....................................................................................... 105
5.2 Managerial Implication..................................................................... 106
5.3 Research Limitations ........................................................................ 111
5.4 Recommendation for Future Research .............................................. 112
REFERENCES .............................................................................................. 113

xiii
Universitas Indonesia
 LIST OF TABLES

Table 1.1 Indonesia Footwear Market Size ........................................................... 2

Table 1.2 Instagram and Facebook Audience Growth........................................... 4

Table 3.1 Variable Operationalization ................................................................ 41

Table 3.2 Evaluation of PLS-SEM Measurement Model .................................... 57
Table 3.3 Evaluation of PLS-SEM Structural Model .......................................... 60

Table 4.1 Result of Reliability and Validity of Pre-test Questionnaire ................ 64

Table 4.2 Result of Reliability and Validity of Perceived Personalization
after PP4 Removal ............................................................................................. 66
Table 4.3 Gender ............................................................................................... 67
Table 4.4 Age .................................................................................................... 68
Table 4.5 Level of Last Pursued Education/Current Pursued Education .............. 69
Table 4.6 Employment Status............................................................................. 69
Table 4.7 Domicile ............................................................................................ 71
Table 4.8 Monthly Spending .............................................................................. 71
Table 4.9 Marital Status ..................................................................................... 72
Table 4.10 Purchased Brands ............................................................................. 73
Table 4.11 Recent Bought Brands ...................................................................... 74
Table 4.12 Length as Followers of the Referred Brand on Instagram .................. 74
Table 4.13 Social Media Used............................................................................ 75
Table 4.14 Average Time Spent on Social Media per Day.................................. 76
Table 4.15 Descriptive Statistics ........................................................................ 76
Table 4.16 Internal Consistency Evaluation Result with Repeated Indicator
Approach ........................................................................................................... 82
Table 4.17 Internal Consistency Evaluation Result with Two Stage Approach ... 83
Table 4.18 Convergent Validity and Discriminant Validity Result with Repeated
Indicator Approach ............................................................................................ 84
Table 4.19 Convergent Validity and Discriminant Validity Result with Two-Stage
Approach ........................................................................................................... 86
Table 4.20 Fornell-Larcker Criterion for Discriminant Validity Evaluation ........ 87
Table 4.21 Heterotrait-Monotrait Ratio (HTMT) Evaluation .............................. 88
Table 4.22 Collinearity Evaluation with Inner VIF Value ................................... 89
Table 4.23 Evaluation on Coefficient of Determination ...................................... 89
Table 4.24 Evaluation of Predictive Relevance with Q2 Cross-Validated
Redundancy ....................................................................................................... 90
Table 4.25 Evaluation of f2 Effect Sizes ............................................................. 91
Table 4.26 Evaluation of q2 Effect Sizes ............................................................ 92
Table 4.27 Result of Direct Path Coefficient through Bootstrapping Procedure .. 93
Table 4.28 Summary of Hypotheses Testing Result............................................ 95

xiv
Universitas Indonesia
 LIST OF FIGURES

Figure 3.1 Research Model ................................................................................ 33

xv
Universitas Indonesia
 LIST OF ATTACHMENT

Attachment 1 Research Questionnaire .............................................................. 128

Attachment 2 SMART-PLS Result................................................................... 138
Attachment 3 IBM SPSS 23 PRE-TEST RESULT ........................................... 142

xvi
Universitas Indonesia
 CHAPTER I

INTRODUCTION

1.1 Research Background

Along with the rapid growth of Internet penetration level in Indonesia that
covers 56% of the populations, the digital environment is also rapidly evolving
where now it presents variety of digital based products and service to be chosen
(Hootsuite, 2019). The products vary from Entertainment Streaming Applications
(ESAs), ride hailing service, food delivery services, banking, and social media
(Hootsuite, 2019). The Improvement within the digital environment create changes
as according to the Technological Readiness Report, Indonesia’s score index on
how prepared Indonesia is to face technological changes improved from the
previous rank in 2017 of 69 with the score of 3.25 to rank 67 with the score of 4,375
(Economist Intelligence Unit, 2018) in which signals improvement that Indonesia
is currently embracing digitization.
Presented with variety of digital based products and services and
digitization at hand, one of the most used digital products by people in Indonesia
is social media that currently have active users with size around 56% out of the
total populations of Indonesia (Hootsuite, 2019). Social media as a form of online
social networking services offers an enchantment in terms of information
communication technologies as individuals are able to communicate easily and
could seek out other individuals opinion about specific topics through social media
such as about products and service to satisfy their role as customer to seek rich
information which creates opportunities for firms to engage in social media and
adapts and shifts their digital marketing strategies that can closely match the
customers desire to provide customers the needed information based on their
preferences and habits while also gather feedbacks, data and moreover promoting
two-way interaction as a form of engagement to strengthen relationship between
firms and the customer (Tiago & Veríssimo, 2014).

1
Universitas Indonesia
 2

Furthermore, Social media are a rich platform for brands that seeks to
stimulate consumer’s hedonic goals. (Spiller, Tuten & Carpenter, 2011). In which
when the consumption of the goods is driven by an affective and sensory experience
(Hirschman and Holbrook, 1982). Hedonic objectives are often associated to
shopping in which the reason to conduct shopping is to satisfy the goal to reach
enjoyment, pleasure, social experience, and entertainment (Babin et al., 1994).
Which are commonly found within the context of fashion shopping and are
associated with adventure shopping motivation to satisfy the individuals desire to
be different from the ordinary and idea shopping motivation that drive the
customer’s willingness to learn new styles and noticing the current products and
innovations to keep up with the trend in fashion (Kang & Park-Poaps, 2010).
Regarding fashion, fashion in Indonesia is considered to be one of the highest
shopped category in online shopping that reached 2.3 billion USD spent by the
consumer (Hootsuite, 2019).

Table 1.1 Indonesia Footwear Market Size

Indonesia Footwear Market Sizes

Unit Year Sales (Retail RSP) in Current Prices

2015 34,125.6

2016 36,716.4

IDR billion 2017 37,930.2

2018 41,656.6

2019 44,616.5

Source: Euromonitor International (2020)

Within the fashion subsector, footwear is one of the highlighted sub-sectors

that experience a growth in demands. From table 1.1, it can be seen that the retail
sales of footwear sector is experiencing an average of 7% growth each year where
in 2019 it was recorded to be worth 44,6 billion rupiah and are estimated to worth
48,2 billion rupiah in 2020 (Euromonitor International, 2020). Seeing that the
accumulated sales revenue within the footwear market always experiencing a

Universitas Indonesia
 3

growth each year, it can be inferred that there is a growing domestic demand for
footwear in Indonesia, presenting an opportunity within the market for businesses
within footwear industries. Within the industries itself, Kementrian Perindustrian
Republik Indonesia stated that the local footwear industries also experienced a
significant growth in production especially within the period of 2017-2018. Within
2018, the local footwear industries were able to reach peak growth in seven years
period of 9,42% in which also surpass significantly the 2017 growth that only reach
2,22% (Kemenperin, 2019). And within the course of 2018, Indonesia’s footwear
industry export value is also growing to 4,13% reaching 5,11 billion USD from the
previous year value of 4,91 billion USD (Kemenperin, 2019).
Presented with such market opportunities, according to an article published
by Badan Ekonomi Kreatif Indonesia, fashion industry which include footwear is
also considered to be one of the biggest contributing subsector within Indonesia
creative economy and within the top 10 of the most purchased item on e-commerce
platform (Munaf & Pesik, 2019). From these article it can be inferred that there is
also an outgoing preference toward local footwear within the Indonesia market, in
which is also corroborated by consumer insight report published by Deloitte South
East Asia from a 1500 representative samples that local clothing and footwear
category are improving in terms of purchase preference with 96% of respondents
expressed this preference from the previously 86% from 2018 with reasons
predominantly being affordable (Wee Meng, 2020). Even so, local footwear is still
in heavy competition with foreign brands within the market in which are more
preferred in terms of their perception of quality and moreover trust compared to
local footwear brand (Wee Meng, 2020). Despite of the opportunities, based on this
consumer insight survey, indeed there are challenges for local footwear brands in
terms of their quality perception and trusts and commitment of the consumers. On
tackling these marketing problems, previous research by Shanahan et al., (2019)
imply that the current digital platform promotes opportunities for brands to employ
digital marketing efforts by promoting their products and service while also
strengthening their consumer relationship with the brand through personalised
advertising. Past studies by Shanahan et al. (2019) on Facebook found personalised

Universitas Indonesia
 4

advertising are beneficial if the advertisement is perceived to be personalised and

generate positive outcomes toward the brand perception and brand attitude.

Table 1.2 Instagram and Facebook Audience Growth

Category Instagram Facebook

Total audience 62 Million 130 million

Advertisement Audience Aged 13+ 18 Million (29%) 79 million (61%)

Quarterly growth of advertising reach + 5,1 % 0%

Percentage of male audience 31 million (51%) 72 million (56%)

Percentage of female audience 30 million (49%) 57 million (44%)

Source: (Hootsuite, 2019)

Possessing similar function in the context of personalised advertisement,

Instagram as a photo and video sharing social media also provide the options for
brands to air personalised advertisement. In Indonesia, Instagram are one of the
highest used social media in Indonesia aside from Facebook in terms of audience
as can be seen on Table 1.2. Even though Facebook indeed possessed a bigger
number of audiences, its audience already stopped growing while Instagram as of
2019 possessed a total of 62 million active monthly advertising audience with a
constantly growing audience with a rate of 5,1 % in a quarterly basis. Among the
62 million of monthly active users in Indonesia, 29 % of the populations is aged 13
or above and considered to be highly reachable with adverts that is promoted by
brands using Instagram (Hootsuite, 2019). Furthermore, 93,9% of internet users
in Indonesia are connected to the internet through their mobile devices in form of
smartphone (APJII, 2017) which on most mobile phone users in Indonesia,
Instagram is ranked within the top three in the category of installed mobile
application with the most active users (Hootsuite, 2019). Instagram as a visual-first
social media are also beneficial for brands as according to a research conducted by
Facebook (2019) using survey responses from 21,000 respondents, Instagram are
typically used by users to seek inspiration from people and brands while shopping
online. Moreover, Instagram is used to understand what the current trend is and

Universitas Indonesia
 5

conduct research before buying. In terms of online behaviour, 87% out of the
people surveyed also said that after they seek for information on Instagram about
certain brands or products, they took several actions such as following the brand,
visiting the website and even conduct purchase activity. The research also states
that people they surveyed perceived a brand as popular (78%), creative (77%),
entertaining (76%), and relevant when they have presence on Instagram. A survey
by Ipsos also corroborate a similar findings in which that 81% of Instagram users
in Indonesia uses Instagram to learn about a brands that they interested in, and
moreover 76% of Instagram users also claimed that after discovering the brands on
Instagram they conduct purchase from that corresponding brand that they discover
(Ipsos, 2018). These shows that the current landscape of local footwear industries
and opportunities presented by social media particularly Instagram in Indonesia will
present growth in the upcoming days, therefore this research will emphasize toward
a business problem of local footwear brands that seek to leverage their presence on
Instagram by using Instagram advertisement as a personalised advertising scheme.

Brands that advertise in Instagram can be identified using a key indicator of

the word “sponsored” that is commonly found above the posts (Instagram, 2020).
Instagram advertisement are a posts or stories that a registered business account
pays to be promoted to Instagram users’ feeds (Newberry, 2020) on Hootsuite
article. Instagram advertisement within its original looks are similar with regular
Instagram posts that can be formatted with stories, photo, video, or carousel
presentation where the only differences lies with the “sponsored” label identified
within the contents. Instagram advertisement as a personalised advertising scheme
also allow brands to choose its placement according to their specific targeted
audience profile based on the online data that the brands possessed such as online
behaviour (search query; internet activity), and demographic information (Cooper
2015; Newberry, 2020). Among numbers of local footwear brands that show up in
current days as the preferences toward local products are improving, according to
an article of Bisnis Indonesia written by Andriani (2019) there are numbers of
Indonesian made footwear brands that is very popular within the domestic market
and are also able to reach international sales to foreign country. These brands are
originally launched in Bandung namely Brodo, Geoff Max, NAH Project, and

Universitas Indonesia
 6

Portee Goods. Among these four brands, Brodo, NAH Project, and Geoff Max also
acquire a significant search query trends from the period of February 2019 –
January 2020 in google trends that represents Indonesia’s interests toward these
brand popularity as stated in article written by Krisdamarjati (2020) on Litbang
Kompas research article. Furthermore, these select of brands are also identified to
still be airing Instagram advertisement in recent days that fit the requirement of the
business problem this research try to investigate. Following the business problems
in which covers opportunities present in both digital and local footwear landscape
of Indonesia also the challenge that local footwear brands currently facing regarding
their quality perception and how to strengthen the consumer’s trusts to build brand
loyalty, this research will cover the marketing problem in which whether
personalised advertising scheme can be leveraged to stimulate consumer brand
engagement and brand attachment in which can predicts positive quality perception
and brand loyalty of the brands.

Within the research conducted by Shanahan et al., (2019) it is found that

personalised advertising scheme in Facebook by Nike is perceived to be
personalised in which this factor leads to the stimulation of engagement and
emotional attachments of the consumers toward the brand. From the development
of interactive experience within the engagement process and also emotional
attachment, it is found that these two aspect also impact the brand’s quality
perception and loyalty in which that the author asserts that therefore the
personalised advertising scheme can lead to positive outcomes for brands as long
as it is perceived to be personalised or match with the preferences of the consumers.
According to Li, (2016) Personalization can be defined as delivering a
personalized advertisement to individuals based on their exclusive preferences.
Where within the context of a company and its customers personalization can be
described as how a company create personal communication approach following
the customer’s stated or implied interests (Roberts & Zahay, 2012). In addition to
that, customers find personalization carried through advertisement are appealing as
it provides information that helps them fulfil their needs (Boerman et al., 2017).
Within the context of social media, personalization has been applied into the online
environment following the opportunity presented by recent advances in

Universitas Indonesia
 7

technologies (Tran, 2017). on a research conducted by Shanahan et al (2019)

regarding the effect of personalized advertising towards consumer brand perception
and attitudes in Facebook found that positive perception of personalised
advertisement content drives higher consumer brand engagement and stronger
brand attachment in which affect perceived quality and brand loyalty positively.
Then to understand the effect of personalized ads carried out by brands it is
interesting to analyse the phenomenon through the consumer perspective that is
exposed to personalised advertisement on social media (Shanahan et al, 2019).
Several research regarding personalization by (Sunikka & Bragge, 2009), (Pletikosa
Cvijikj & Michahelles, 2013) and DelVecchio et al (2006); Fetscherin and Heinrich
(2015) generates a similar general conclusion regarding personalization effect in
which it strengthens consumers attitudes towards a brand in terms of engagement
and loyalty to the brand. For example, in the context of Facebook as social media,
the platform offer users the opportunity to interact with the ads by liking, sharing,
or viewing the ads to their friends on Facebook (Dehghani & Tumer, 2015).
Nikhashemi et al. (2017) defined perceived quality as consumer’s
perception that a brand’s product or service is reliable and dependable according to
their preference, purchase choice, and satisfaction. According to a research carried
out by (Hollebeek et al., 2014), brands tend to have a higher consumer loyalty and
profitability when brand’s interact with their consumer through social media or
public. Here, a consumer may decide to purchase a certain product from a brand
because of the favourable relationship with a certain brand because of the pre or
post purchase process engagement made by the (Chaudhuri & Holbrook, 2001).
Prior study also found that personalized advertisement can also improve
consumer’s participation and commitment to a certain brand (Zhang et al., 2015).
Personalization then allows consumers to develop and improve an emotional
attachment towards a brand (Shanahan et al., 2019). When an attachment towards
a brand are emotionally based, the feelings toward a brand drive purchase and
repurchase behaviours (Whan Park et al., 2010). Brand attachment usually predicts
customer repurchase behaviours (Chaudhuri & Holbrook, 2001). Therefore
personalization in social media can assist the development of brand attachment and
brand loyalty. (Shanahan et al., 2019). Regarding brand loyalty, a consumer may

Universitas Indonesia
 8

purchase a product if that product gives unique value that no other product can
imitate (Chaudhuri & Holbrook, 2001). Following a positive response towards a
brand a consumer can feel happy and develop brand loyalty (Chaudhuri &
Holbrook, 2001). Based on the explanation above, this research will also study the
effect of perceived quality towards brand loyalty.
Furthermore, as a modification, the researcher also adds self-expressive
brand toward the model to investigate the effect toward consumer brand
engagement and brand attachments from the model developed by this research core
reference journal by Shanahan et al. (2019). Self-expressive brand is added based
on the inspiration from previous study by Algharabat et al. (2020) and Lee and
Workman (2015). In addition to that, considering this research are carried to study
the effect of brand’s personalised advertisement on Instagram. Instagram
advertisement allows the advertisement to be incorporated with the same functions
of interactions within an organic content of social media such as like, comment, and
share. Therefore, on the social network, such interactions effort by user in social
media asserts the level of engagements in which they consider the brand to be one
of their part of online self-expression (Algharabat, 2017). In the context of
personalised advertising in social media (Facebook), these engagements are
typically seen through whether the brand congruent signals stimulate the users to
like, comments as interactions that can show their online extended self to others
through the advertisement (Tran, 2017).
According to consumer culture theory, symbolic and experiential features
of a brand that consumer’s experience with a brand can enhance one’s self-identity
(Brodie et al., 2013) therefore asserting the conceptual relationship of self-
expressive brand and consumer brand engagement. Relationship theory also imply
that brand relationships facilitate oneself with the opportunities that he or she can
improve the definition of themselves by the formation of value, worth, social
identity and peer evaluations that indicate social impression formations (Fournier,
1998). Social impression formation is affected by cues displayed and used by
brands carried with a proper linguistic power to express and relate to one’s personal
life through the cues within marketing activities (Lee and Workman, 2015). Self-
concept theory also asserts that individuals can maintain their self-concept by using

Universitas Indonesia
 9

a brand attributes such as promotional messages (Graeff, 1996). As the

development of the digital environment is become richer, through the backbone of
Web 2.0, self-concept is expanding towards the digital environment (Belk, 2013)
where technology enable brands and customer to easily project their online persona
through a simple post, tweets, like, or comment through a social media (Belk, 2013).
According to Aaker (2009) brand attributes that is seen to be congruent with one
self-concept can strengthen the brand’s connection toward its consumer as an it is
able to assist an individual to express themselves. In addition to that, Self-
expressive brand can help develop affective aspects such as brand attachment that
could potentially lead to brand loyalty (Carroll & Ahuvia, 2006). Past research also
shows that brands reflect the way consumers think in which provide them some
kind of extension of themselves (Belk, 1988) where usually the brand symbolic
design is used by consumer to shape his or her self-identities in which can impact
affective behavioural components (Carroll & Ahuvia, 2006). Ball and Tasaki
(1992) also describe that the stronger a brand can reflect oneself through the
individual roles, relationship, accomplishment, and experiences are more likely to
have a stronger attachment to the individual consumer that perceived the brand as
somewhat “personalized” which may lead to an increased commitment toward the
brand. Therefore, self-expressive brand is added on this research to measure
whether the brand’s communicated attribute within the ads is seen to be congruent
with the consumer’s self that extend towards their self-expression.

1.2 Problem Definition

Referring to the research background above, this research will be focusing on a core
problem at investigating the effect of perceived personalization and self-expressive
as a driver that could lead to perceived quality and brand loyalty in the context of
personalization of advertisement carried on Instagram by local footwear brands as
the chosen entity within the fashion industry due to the current improvement its
experiencing. Aside from the modification to the industries and addition of variable
namely self-expressive brand, this research is referring to a prior research
conducted by (Shanahan et al., 2019). Therefore, the proposed research question
that will be investigated are as follows:

Universitas Indonesia
 10

1) Does perceived personalization influence consumer brand engagement and

brand attachment of a local footwear brand on Instagram?
2) Does self-expressive brand influence consumer brand engagement and brand
attachment of a local footwear brand on Instagram?
3) Does consumer brand engagement and brand attachment influence brand
loyalty and perceived quality of a local footwear brand on Instagram?
4) Does perceived quality influence brand loyalty of a local footwear brand on
Instagram?

1.3 Research Purpose

The purpose of carrying out this research is to answer the proposed research
question above in the context of personalized advertisement on Instagram and
structured as follows:

1) Gain insights on whether social media perceived personalization influence

consumer brand engagement and brand attachment of a local footwear brand on
Instagram
2) Gain insights on whether self-expressive brand influence consumer brand
consumer brand engagement and brand attachment of a local footwear brand on
Instagram
3) Gain insights on whether consumer brand engagement and brand attachment
influence brand loyalty and perceived quality of a local footwear brand on
Instagram
4) Gain insights on whether perceived quality influence brand loyalty of a local
footwear brand on Instagram

1.4 Research Implications

There are several benefits or implication for future research or business

practitioners as follows:

1) For Future Research

Hopefully this research can provide understanding, insight, or guidance for a
similar or a more in-depth research soon. In addition to that, hopefully this

Universitas Indonesia
 11

research can also fil the research gap and be used as reference for a research
regarding personalization and its consequence on social media context.

2) For Business Practitioners

Hopefully this research can contribute or give insights for businesses, especially
local footwear brand regarding the effect of personalization and self-expressive
brand within the context of personalised advertisement carried on Instagram on
their brand towards consumer brand engagement, brand attachment, perceived
quality and brand loyalty. Hopefully provided with insights on this research,
brand and business practitioners in a similar field can understand the
phenomenon better to strengthen the brands relationship with the consumer and
therefore develop and choose a strategy best suited for them.

1.5 Research Scope

1.5.1 Research Respondents

This research will choose respondents from both gender group of male and
female and are within the age group of 18-34 years old which according to
Asosiasi Penyelenggara Jasa Internet Indonesia is the age group with the
highest percentage of internet user by 49,52% and accounts from the total
population of Indonesia (APJII, 2019) and accounts for 71,80% from the
total of Instagram users in Indonesia (Statista, 2020). Since the research will
be investigating the phenomenon on the customers of several local footwear
brands that deliver personalized advertisement in Instagram, the
respondents then will be focused on those who uses Instagram that follow
(Brodo/Portee Goods/Geoff Max/NAH Project) Instagram page and owns
at least one product from the stated brands.

1.5.2 Geographical Scope of Research

Data collection of this research will cover the geographical domain of Java
island where according to the data from Asosiasi Penyelenggara Jasa
Internet Indonesia, Java accounts for the highest concentration of internet

Universitas Indonesia
 12

users compared to other area in Indonesia by having 55% from the total of
internet user in Indonesia (APJII, 2019).

1.6 Research Structure

In terms of the report writing structure, this research will be divided into five parts
as follows:

Chapter 1 – INTRODUCTION

Chapter 1 will introduce and explain the research background or why this research
is proposed, problem statement, research purpose, and the expected contribution
from this research. In addition to that, chapter 1 will also explain the research scope
starting from the research participants or respondents, geographical scope of the
research, and the timeline of research.

Chapter 2 – LITERATURE REVIEW

Chapter 2 will describe various theory and conceptualization used within the
research. So, the conceptualization will cover the concept of consumer behaviour,
digital marketing, marketing communication, brand, social media, personalised
advertising and moreover the concept of each variables namely perceived
personalization, self-expressive brand, consumer brand engagement, brand
attachment, perceived quality, and brand loyalty.

Chapter 3 – RESEARCH METHODOLOGY

Chapter 3 will cover the description of the research model, proposed hypothesis,
research approach, variable operationalization, type of data, sampling method, data
collection method, and data processing and analysis method that will be used on
this research.

Chapter 4 – ANALYSIS AND DISCUSSION

Chapter 4 will be utilized by the researcher to present the result of data processing
included with the researcher analysis and discussion to answer the problem
statement and research question described previously on chapter one.

Universitas Indonesia
 13

Chapter 5 – CONCLUSION AND RECOMMENDATION

Chapter 5 as the last part of the structure will be utilized by researcher to draw
conclusions from this research and formulate recommendation that can be
considered by related parties or businesses within local footwear industries in terms
of decision making when formulating a marketing strategy within the field of digital
marketing by considering the investigated result of consumer behaviour within this
research. In addition to that, chapter 5 will also describe the research limitation and
suggestion for future research in the similar field that can be considered by other
researcher in near future.

Universitas Indonesia
 CHAPTER II

LITERATURE REVIEW

2.1 Consumer Behaviour and Decision Making

Consumer behaviours is rooted from four disciplines, psychology that

studies human mind and the mental factors that affect behaviour such as needs,
perception, personality, learned experiences and attitudes, Sociology that studies
the development, structure, functions, and problem within human society,
anthropology which compares human societies, culture and its development and
also communication as the process of exchanging information personally or through
a medium by using persuasive strategies (Schiffman et al., 2005). Consumer
behaviours highlight process of searching, purchasing, using, evaluating, and
disposing of products and services by consumers that is done in response of
satisfying their desire or needs (Schiffman et al., 2005).

The process done by the consumers is called decision making process,

where upon making a decisions, consumers as decision maker and problem solver
usually faced between a conditions of routinized response behaviour or extended
problem solving (Howard & Sheth, 1969; schiffman et al, 2005; Solomon, 2009).
These conditions influence how consumers proceed toward the decision making
process (Howard, 1977). Upon engaging extended problem solving, consumers will
go through the step of decision making thoroughly from problem recognition up to
post purchase evaluation. On the other hand, routinized response behaviour will
only require a consumer to make minimal effort to come up with a decisions
(Howard, 1977; Alba and Hutchinson, 1988).

When categorizing whether decision process follows extended problem

solving or habitual decision making, a consumer is faced with several antecedent
condition such as the level of involvement of a certain product, the degree of
perceived risk, information processing activity, the consumers learning model on

14
Universitas Indonesia
 15

cognitive learning or behavioural learning, and the marketing actions that the
consumers come across for example a product display as environmental cues or
advertising through web sites, salespeople, etc (Solomon, 2009). Overall, consumer
decision making process includes three stages which are the input, process, and
output stages of decision making (Schiffman et al., 2005). Within input stage, there
are two influencing factors that is faced by the consumer as decision maker, first is
the firm’s marketing mix (i.e. product, price, promotion, and place where the
product is sold), second is the sociocultural influences such as the consumer’s
relatives, cultural and subcultural entities that influence the consumers through
different kind of communication methods (Schiffman et al., 2005).

Second, the process stage that highlight how consumers make decisions that
is influenced by psychological factor such as motivation, perception, learning,
personality, and attitudes (Schiffman et al., 2005). These psychological factor or
internal influences affect those aspect within the input stage on how consumer’s
process their recognition of need, search of information, and evaluation of
alternatives that in turn can become a part of consumer’s psychological factors
through the process of learning (Schiffman et al., 2005). These internal influence
then create a drive on consumers to conduct the decision making process, for
example motivation can be the internal state that activates consumers to satisfy their
needs through a goal-oriented behaviours (Solomon, 2009). Then information
search can be affected by what is called perception in which is the process on how
people select, organize, and interpret information (Solomon, 2009). The acquired
information then can affect learning or a change in behaviour caused by the
acquired information or experience in which can be described into two forms of
behavioural learning that is caused by the consumers connections with external
events or stimuli, and cognitive learning that highlight consumers mental processes
in solving problems through acquiring information (Solomon, 2009).

Personality also play a part as it is the set of unique psychological

characteristics that influence the way consumers as a person respond toward a
situation within the environment, and mainly a consumers buy products that are
extensions of their personalities or a products that reflect his or her self-concept
(Solomon, 2009; Carroll and Ahuvia, 2006). Attitudes or a person’s evaluation

Universitas Indonesia
 16

toward a person, object or issue also influence the decision-making process through
three components. Those three components are affect or emotional responses,
cognition or beliefs and knowledge a consumer have toward a products, and
behaviour or consumers intention to do something (Solomon, 2009).

The last process is the output stage that covers the post-decision activities
which comprises of consumer purchase behaviours on whether or not the consumers
conduct purchase activity on their product choice and also post-purchase evaluation
that may result on no-repurchase or re-purchase that can further be traced to trust
and loyalty (Schiffman et al., 2005). These sequence however can be different in
terms of order, and duration between conventional offline environment and digital
environment (Court et al., 2009; Edelman and Singer, 2015). For example, on
digital environment customers can gather information from a search engine or read
other customer’s review from an online forum, or further back, even the initial idea
of demand to purchase could be stimulated by just seeing a post on social media
(Kannan & Li, 2017). Therefore, Consumer behaviours is a field that study
consumers process of searching, purchasing, using, evaluating, and disposing of
products and services to satisfy their needs and desires (Schiffman et al., 2005).

2.2 Marketing Communication

Stemming from Communication theory, Communication is commonly

defined as the transmission of information from one party to another (Shannon &
Weaver, 1949). Within the concept of communication, communication can be
defined as a verbal act, reduction of uncertainty effort, and a process (Dance, 1970).
As a verbal act, communication is the exchange of thought or idea (Hoben, 1954).
Communication also arises as an effort to comply the needs on reducing
uncertainty, act effectively or projecting the ego of human (Barnlund, 1963). As a
process, communication can be defined as the process of transmitting information,
ideas, emotions, skills through symbols, words, pictures, figures and other media
(Berelson & Steiner, 1964). In transmitting information, the transmission itself can
also be defined as communication when a transmission conducted by more than one
participating parties transferred something in the process, for example, messages
that leave one party possession towards the other participating parties (Ayer, 1955).

Universitas Indonesia
 17

As a definition however communication cannot be defined as one sole definitive

concept but moreover a chain of similar terms that include the term communicative
or synonymous term with communication (Grove, 1998).

Within marketing, communication can be described toward the presentation

of communication model in which it explains the process of “meaning” being
transferred from a source toward a receiver (Solomon, 2009). The communication
model explains how organizations as the source create message through the process
of encoding ideas to a message that will be transmitted to the receiver, that is carried
through a medium or communication vehicle in a hope that the consumer
understand what the message the marketer intends to deliver, through the process
of decoding (Solomon, 2009). Throughout the communication model a “noise” is
also considered as part of the process that may interfere the process of achieving
effective communication or block messages (Solomon, 2009). If the message
received by the receiver and are decoded, the communication loop may continue as
the source might receive a feedback or reaction from the consumer as receiver in
which can assist marketer to fine-tune the message (Solomon, 2009).

According to Solomon (2009), every element of the marketing mix is a form

of communication where marketing communication is included as one part of the
marketing Four Ps, namely Promotion, in which promotion mix includes the
communication elements that the marketer controls such as advertising, sales
promotion, public relations, personal selling, and direct marketing. Promotion is
then the coordination of marketing communication effort that is intended to
influence attitudes or behaviours (Solomon, 2009). Marketing communication has
four distinctive roles in which it is intended to inform consumer about new goods
or services, remind them to continue using a particular brand, persuade them to
choose a specific brand over the other, and moreover to build relationship with the
customers (Solomon, 2009).

As of now, marketing communication has turned into what is called

Integrated marketing communication (IMC) where marketers believe that
promotional strategy should include diverse form of marketing communication
(Solomon, 2009). Integrated marketing communication therefore is defined as a

Universitas Indonesia
 18

concept of marketing communications planning which recognizes the added value

of a comprehensive plan that evaluates the strategic role from variety of
communication disciplines such as general advertising, direct response, sales
promotion and public to achieve clarity, consistency and maximum impact in terms
of communications (Caywood et al., 1991).

2.2.1 Advertising

Advertising is a method of marketing that provides information

about a product, service, and business itself with a purpose to build image
and stimulate purchase (Park et al., 2008). Initially advertising was carried
using traditional advertising media such as newspaper or magazines then
towards tv and radio advertisement in which are considered to be somewhat
outdated as it possessed several downside such as repetition, clutter effect
and low involvement level (Park et al., 2008). Nowadays, online advertising
is much more preferred as it is carried over telecommunication lines
(internet) which possessed the ability to conduct ease of targeting,
personalised contents and interactivity even accessibility in mobile devices
and let the advertisers to collect data to evaluate the targeting accuracy to
improve the advertisement personalization (Park et al., 2008). All and all,
the current digital days provide opportunities for advertisers to conduct such
personalization through online behavioural advertising in which can be
defined as online profiling and behavioural targeting (Bennet 2011). In a
way to adjust advertisement to previous online activity of the consumers by
monitoring consumer’s online behaviour such as web browsing data, search
histories, media consumption (post or video watched), purchases, click
towards call to action button, etc, here firms collect the data to basically
create individual profiling to target the advertisement and create a
personally relevant advertisement (Boerman et al., 2017). Therefore, online
behavioural advertising is defined as the practice to monitor people’s online
behaviour and using the collected information and data to show them the
individually targeted advertisement (Boerman et al., 2017).

Universitas Indonesia
 19

2.3 Digital Marketing

Marketing as a whole concept is defined as “activity, set of institutions, and

processes for creating, communicating, delivering, and exchanging offerings that
have value for customers, clients, partners, and society at large (AMA, 2020).
Following the emergence of Internet, the concept of marketing has gone through a
fundamental change in which it adds a technology-specific definitions to marketing
towards building and maintaining customer relationships through online activities
that facilitate the exchange of ideas, products, and services that satisfy the customer
(Mohammed et al., 2004). In response to this change, marketing now involves
digital technologies such as the web, e-mail, databases, mobile devices, and digital
TV in order to achieve profitable result and customer retention through a
multichannel buying process (Chaffey, 2008)Through the use of internet as the
technologies then digital marketing is defined as the utilization of digital
technologies to create an integrated, targeted and measurable communication to
acquire and retain customer while also building deeper relationships with the
customers. (Smith, 2007; Wymbs, 2011).

Aided by Digitals technologies that is changing rapidly and are also causing
a shift in environment where firm operates and create a reduction in Information
asymmetries between customers and sellers (Kannan & Li, 2017). firms are now
able to foster their digital relationship and allow conversation with its customer
through reaching the customer digital touch points that is bridged by digital
conversation interfaces (Alkhateeb et al., 2008). Facilitated through the creation of
Web 2.0, this technical infrastructure enables the creation of interactive websites
by firms that encourage and facilitate participation from the users (Tapscott &
Williams, 2007). Through the foundation of Web 2.0 social media as a product
facilitate as conduits that transmit content through the interactions between
individuals and organization (Kietzmann et al., 2011) where it allows consumers as
users to participate and generate value through the consumer-generated media (i.e.
reviews, posts) that promotes the brand (P. Berthon et al., 2008; Muñiz Jr & Schau,
2011). As of now, according to American Marketing Association, digital marketing
is defined as the utilization of digital or social channels to promote a brand whilst

Universitas Indonesia
 20

also reaching the consumers and are predominantly executed through social media,
search engine, internet, mobile device and other channels (AMA, 2020).

2.4 Social Media and Instagram

Social media is defined as a group internet-based application that is built

based on the technological foundations of Web 2.0 that enables the user to create
or exchange user-generated content (Kaplan & Haenlein, 2010). As a conduit,
social media as a channels use internet and web-based technologies in which
transform broadcast media that are initially monologues (one to many) into social
media dialogues that are putting emphasis on many to many interactions (Kaplan
& Haenlein, 2010). Social media also varies in type, first there are social networking
site that enable users to write and access short messages such as Twitter, image
sharing and video sharing social media and social media networks in which allows
users to add friends, update their individual profile and send messages to friends (P.
R. Berthon et al., 2012). As of the current development, social media are
experiencing a shift toward their forms into social networks in which incorporates
collective features from the available social media (P. R. Berthon et al., 2012).

Social networking sites are an application that enables its users to connect
and invites their friends and colleagues to access the profiles, and sending instant
message to each other through the creation of personal profile (Kaplan & Haenlein,
2010). As a social networking site, Instagram is defined as a service in which are
available for smartphones and other mobile devices that allow users to share photo
and videos with other users (Gibbs et al., 2015). As a core photo and video sharing
social networking site, Instagram allows user to choose different format and also
adding written descriptions toward the posted photograph or videos (Gibbs et al.,
2015).

2.4.1 Personalised Advertisement on Instagram

Shanahan et al. (2019) on their research stated that personalised

advertisement is an advertisement that is shown to users based on their prior
activity on the internet such as interest represented by their prior search
queries and internet activities. The available online data within the online

Universitas Indonesia
 21

environment is also an opportunity as now business or brand can assess

customer preferences, behaviour, habits, interests through social media
(Shanahan et al. 2019). Among the difference setting investigated between
Facebook (prior research) and Instagram (this research) Instagram
advertisement are a posts or stories that a registered business account pays
to be promoted to Instagram users’ feeds (Newberry, 2020) on Hootsuite
article. Instagram advertisement within its original looks are similar with
regular Instagram posts that can be formatted with stories, photo, video, or
carousel presentation where the only differences lies with the “sponsored”
label identified within the contents and like, comment, and share features
but with additional options to add call-to-action button such as learn more,
catalogue sales conversion, and conversion to the brand’s website. allows
Instagram user to interact with the advertisement (Newberry, 2020).

Within the execution of delivering Instagram advertisement, brand

can also choose for its placement options to be automatic or manual in which
allows the brand to pick and choose the format of the advertisement and also
the specified target audience that will receive the advertisement based on
their online behaviour such as search query and internet activity and also
demographic information accompanied by the options to conduct detailed
targeting options using keyword of interest that can virtually be anything
which can represent the audience interests (Cooper 2015; Newberry, 2020)

2.5 Brand

According to Hankinson (1995), brand can be described into at least six

different meanings based on a category that includes visual, perceptual, positioning,
added value, image, and personality. Within visual definition, brand is a
distinguished form of name or symbol such as logo or trademark to identify and
differentiate sellers of goods and service from their competitors (Aaker, 1991).
Perceptually a brand can also be described using three category that must exist
within a brand, this includes an appeal of senses, an appeal of reason and an appeal
towards emotion (Doyle et al). This three-constituent part of a brand is the most
identifiable in which it contributes to a brand ability on maintaining or

Universitas Indonesia
 22

strengthening the overall brand performance in the marketplace (Hankinson &

Cowking, 1995).

Positioning definition of a brand can be described as the way consumers

distinguish brands and the way that brands is able to communicate their goods and
service to earn a position within the consumers mind through a practice of
advertising to gain impressions through brand name(Duckworth, 1991; Ries et al.,
1986). Defining brand from added-value perspective, a brand is defined as an
identifiable product, service, person, or place in which is augmented to a way that
the buyer perceives unique added values that relatively match to their needs (de
Chernatony, 1992). Through the headings image, a brand is described to have a
congruity between the brand’s own image and the actual or self-image of the user
which led to the brand’s to be more enjoyable when used (Gardner & Levy, 1955).
This explanation then led to the concept of brand image that are not uniform
(Hankinson & Cowking, 1995). Then, brand image is defined as a symbol that is
created by the advertiser in effort to differentiate the brand psychologically through
the associated symbols (Frazer, 1983) . Bullmore (1984) on the other hand defined
brand image as what people think and feel about the brand and that the brand has
inter-relationship with the user and are differ individually.

Brand defined from the perspective of its personality through Brand

Personality Inventory is a unique combination of functional attributes, symbolic
values that is identified by the target consumer (Alt & Griggs, 1988). Therefore,
through combining all these different definitions as part of brand definition, brand
proposition concept is developed that describe a brand from its personality,
positioning, target consumer (Hankinson & Cowking, 1995). Therefore, a brand is
a product or service that is made distinctive through its positioning relative to the
competition, and are possess unique combination of functional attributes, symbolic
values that is identified by the target consumer as the brand’s personality
(Hankinson & Cowking, 1995).

Universitas Indonesia
 23

2.6 Perceived Personalization

The term personalization is defined from a theoretical framework of

Elaboration Likelihood Model (Li, 2016; Petty et al., 2000). Elaboration Likelihood
Model describe that components of communication such as the source, message,
recipient, and context can be transformed into to a persuasive communication in
which it can affect people’s information processing and attitude change in different
form and at different level of ability in elaborating the message (Petty et al., 2000).
Communication itself is defined as through communication the transmission of
information from one party to other party (Shannon & Weaver, 1949). According
to the Elaboration Likelihood Model, there are two information-processing route,
central and peripheral based on the degree of high or low level of elaboration (Petty
& Cacioppo, 1986). The central route indicates when people carefully assess all
available information within a message before they shows an attitude while
peripheral route happened when people use minimal effort to assess the information
and generate a faster attitude changes (Petty & Cacioppo, 1986). In essence, through
the process of communication, a persuasive communication is a personalization
effort to create a match between a message to a certain characteristics of the
message recipient in order to be more persuasive than a generic message (Petty et
al., 2000). Similar conditions also happened in the study of personalization as a
process by (Kramer, 2007). that personalization hinges on the compatibility of the
message that match the recipient’s characteristics (i.e. preferences), and that the
ability of the recipient to recognize the message will also influence on whether the
expected favourable personalization effect is achieved. Since personalization
hinges on matching the message toward preference, marketers rely on the
“measured” preferences or assumed preferences to form the personalized message
since individuals often have a poorly defined preferences thus not guaranteeing
that the actual personalized message can reach the recipients “true” preferences (Li,
2016). This conditions argue that a favourable personalization effect are likely to
be achieved by what is called a perceived personalization (Li, 2016).
Different kind of communication form also affect how people perceive on
whether there is a message match or mismatch (Seok Sohn et al., 2012). For
example, personalization has been applied through the practice of direct marketing

Universitas Indonesia
 24

letters in the conventional offline environment up to the tailoring of customized

message content by a marketers on digital environment (Ross, 1992; Vesanen,
2007). Then as a concept, personalization covers the issue of personalized
marketing, personalized marketing output, and value for both the customer and
marketer (Vesanen, 2007). The objective of personalization is that the recipient or
consumers are expected to pay closer attention, and thus process it better and more
likely to use the information for decision making (Rimer & Kreuter, 2006). By
personalizing, the personalized message tend to be memorable, more likeable, and
persuasive compared to non-personalized message (Noar et al., 2009). Therefore,
Personalization is defined as delivering a personalized message to individuals based
on their exclusive preferences (Li, 2016).

2.7 Self-Expressive Brand

Self-expressive brands is the degree on which a brand express or reflect a

consumer’s inner self and how it enhances his or her self within the social context
(Carroll & Ahuvia, 2006). Through being “self-expressive”, a brand can encourage
consumers to develop attachments toward the brand in which it can positively lead
to brand loyalty (Carroll & Ahuvia, 2006). Based on Aaker (2009), a self-expressive
brand strengthens the connection of customer towards a brand, where it is expressed
on how the brand symbolize certain status to the customer. According to Belk
(1988) consumers of a brand often think about the brands as part or extension of the
consumer’s self and uses the symbolic design of the brands to build their self-
identities. Fournier (1998) describe that self-expressive brand indicates consumer’s
willingness to spread messages about themselves to other within the social media
context. Loureiro et al. (2010) also define self-expressive brand as a fit between the
brand and the consumers self-concept and how the brand allows the consumers to
express themselves.

Self-expressive brand has also been discussed within the self-concept theory
(Graeff, 1996). Based on the self-concept theory, individuals or consumers of a
brand can uphold their self-concept through the brands and products of that brand
that they own. While the consumers consider that their ideal or actual self-concept
are in match with the brand’s personality, the development of self-congruence

Universitas Indonesia
 25

between the individuals and with the brand will be fostered (Wallace et al., 2017).
Self-expressive brand also provides important aspect for social integration such as
being a symbol of personal accomplishment in which can differentiate themselves
from other and thus shows their individuality (Escalas and Bettman, 2005;
Schembri et al., 2010). Brands is linked to the consumer’s self when a brand is able
to assist consumers to achieve a self-motivated goals (Escalas & Bettman, 2005).
Within the context of social media, previous research by Jensen Schau and Gilly
(2003) implies that social networks assist by providing platform for consumers to
project their “ideal-self” when expressing themselves with the brand. Previous
research by Tran et al. (2020) posits that nowadays brands can be established within
the social media environment (Facebook) in which by taking advantage of
personalised advertisement on Facebook, after the consumer’s search a brand and
then exposed to the advertisement. The findings from the research by Tran et al.
(2020) also suggest several positive outcomes of brand self-congruence in which
personalised advertising that are congruence with the consumer self will lead to
positive impacts such as, consumer-brand identification, engagements and
relationships and lastly it can predicts loyalty.

2.8 Consumer Brand Engagement

Engagement is when people pursue a goal that fits their orientation so that
they experience a stronger reaction to the activity (Avnet & Higgins, 2006). An
engagement experienced by a customer is a motivational state that is caused by
interactive, co-creative customer experience towards an object (e.g. a brand)
(Brodie et al., 2011). Consumer brand engagement can be determined when a
consumer include important brand as part on how a consumer’s view themselves
among other (Sprott et al., 2009). Consumer brand engagement is also focused
within the area of cognitive, emotional, and behavioural of consumers during a
specific interaction with a brand within the offline consumer brand engagement
context (Brodie et al., 2011). Building on the digital environment of social media
on this research, previous research by Eigenraam et al. (2018) implies that digital
customer engagement is interactions such as talking, learning, or working activities
related to a brand within the online environment. With the presence of social media,

Universitas Indonesia
 26

Pillai and Chauhan (2015) also implies on their research about social commerce on
Facebook, social media as the platform enable brands to conduct a real-time
customer engagement that can foster relationships between the customer and the
brand. Within its conceptualization, the current advancement of research on
consumer brand engagement is quite limited and there are differences on measuring
consumer engagement as an unidimensional concept or multidimensional due to the
variety of field being studied for consumer brand engagement (Brodie et al., 2011;
Hollebeek et al., 2014; Leckie et al., 2016). Previous research by Dessart et al.
(2016) implies that on developing customer engagements, there are two main
aspects that became the pillars of customer engagement which are brand
engagements and community engagements where in details consist of affective
aspects such as enthusiasm and enjoyment, behavioural aspects represented with
sharing, learning, and moreover endorsing the brand while lastly cognitive aspects
that consist of customer’s attention and absorption.
Since this research will be focused on consumer brand engagements on
social media, particularly Instagram as the platform that enable brand to deliver
personalised advertisement and its effect toward consumers of local footwear brand,
referring to Hollebeek et al. (2014), from their research that studies consumer’s
perspective of engagements toward a brand studied on social media, consumer
brand engagement is defined as how a consumer is able to cognitively valence and
react based on their emotional and behavioural aspects toward all activities
conducted by brands on social media platforms, in which are more focused to reflect
the interactively generated nature of consumer brand engagement in terms of
interactive experience on social media that a consumer experience with a particular
brands (Hollebeek et al., 2014). In which that it is viewed as an experience when
consumer is being somewhat cognitively, emotionally, and/or behaviourally
engaged (Brodie et al., 2011; Hollebeek et al., 2014).
Consumer brand engagement indeed possessed conceptual relationship with
emotional brand attachment but are conceptually distinct since brand attachment
that are emotionally led are found as the potential consequences of engagements
which is a result from a consumer’s specific interactive experiences processes with
a focal brand (relational consequences) (Brodie et al., 2011). This rationale is

Universitas Indonesia
 27

derived with a focus of consumer brand engagement concept proposed by

Hollebeek et al. (2014) that are derived from the core definition of consumer brand
engagement that focus on the notions of consumer’s interactive and co-creative
experience as a process and are rooted from the domain of relationship marketing
(Brodie et al., 2011; Hollebeek et al., 2014). Following the definition of Consumer
Brand Engagement by Hollebeek et al. (2014) consumer brand engagement
represents several of key dimensions within the concept in which are,

2.8.1 Cognitive Processing (Cognitive)

Cognitive processing address how the consumer’s process their thoughts

when interacting with the focal brand (Hollebeek et al., 2014). Cognitive processing
is the degree to which consumers absorbed or being concentrated (attentiveness)
towards the brand or brand related attributes within the specific brand interactions
(Hollebeek et al., 2011). From the perspective of social exchange theory, consumers
are predicted to reciprocate positive thoughts towards an object such as brand upon
receiving specific benefits from the brand relationship (two-way interactions)
(Blau, 1964;Pervan Bove, and Johnson, 2009).

2.8.2 Affection (Emotional)

Affection address the level of consumer’s positive brand related feelings

when interacting with the focal brand (Hollebeek et al., 2014). This dimension of
consumer brand engagement reflects the extent of individuals emotional investment
within specific brand interactions (Hollebeek et al., 2011). Within social exchange
theory, this degree of affect level represents the passion which reflect the consumers
perceived brand-related benefits and their favourable brand related affect during
specific brand interactions (Hollebeek et al., 2011).

2.8.3 Activation (Behavioural)

activation which address the level of consumer’s energy, efforts, and time
spent when interacting with the focal brand (Hollebeek et al., 2014). This dimension
of consumer brand engagement reflects the consumers’ willingness to invest
significant time or effort during their interaction with their selected brands
(Hollebeek et al., 2011). From social exchange theory perspective, activation

Universitas Indonesia
 28

reflects how consumer reciprocate their perceived brand-related benefits in a degree

of positive, high dynamic energy and time spent within their interactions with the
brand in which also represents the behavioural facets of consumer brand
engagement (Hollebeek et al., 2011; Vivek., 2009).

2.9 Brand Attachment

Attachment following the Attachment Theory was founded in the context

of parent and child relationships Bowlby (1979). According to Bowlby (1979)
attachments is specified as an emotionally led bond between a person toward a
specific object. Within the context of this research attachments will be focused on
analysing consumer behaviours related to their attachment to a market entity like
brands instead of the physical products (Fournier 1998; Keller 2003; Schouten and
McAlexander 1995). Previous research that conceptualize attachments toward a
brand implies that attachments came in different level of strength which are
influenced with different level of feelings on connections, affection and love, lastly
passion toward the objects (e.g. brand) (Thompson et al., 2005). As addressed with
previous research by Thompson et al. (2005), emotional attachments strength level
are associated with three main behaviours, first are the degree of closeness between
the person and attachment objects which indicates securities, then how the
individual reacts toward environmental stressors on findings the attachment objects
as their safe haven from the associated stress, and lastly reflective psychological
behaviours that depicts separation distress when the individuals are about to be
separated with the corresponding attachment objects.
. Attachment is the representation of the consumer’s experiential, symbolic
and emotional needs (Park et al., 2006). Previous research also highlight that
consumer’s attachment toward a particular brand’s are influenced by the perception
of the consumer’s on whether the brand are socially visible and reflective toward
the consumer’s self and the consumer’s effort as individual that are accompanied
with the perception toward the brand being personalised for the consumers (Ball
and Tasaki, 1992). In similar vein, previously discussed concept of brand
attachment follows three main aspects namely connection, affection and love, and
passion (Thompson et al., 2005) in which by the research of Mikulincer and Shaver

Universitas Indonesia
 29

(2007) highlight the degree of the brand’s linkage toward the self represents a “hot”
affect toward the brand. Ball and Tasaki (1992) also implies that the strength of
emotional attachment toward a brand is formed by the degree of association
between the objects toward a significant events or people that the individual
consumers may found attachment to in which indicates that the strength of
attachment between the person and the brand is subject to a length of experience
and dependence between the person and the objects.
In the context of attachment toward a brand, brand attachment is defined as
the strength of connection between a brand with the consumer (Whan Park et al.,
2010). This connection between the consumers and the brand is characterized by
feelings of affection, connection, and passion (Thompson et al., 2005). affection is
described as consumer’s feeling of peace, love, and friendliness to a particular brand
(Thompson et al., 2005), connection refers to the feelings of bond towards a brand
(Thompson et al., 2005), whereas passion is described as a delightful and
captivated feeling that consumer experience with a certain brand (Thompson et al.,
2005). It is argued that when a consumer is attached toward a brand is more likely
to show commitment towards it and are willing to make sacrifices for the brands
(Thomson et al., 2005; Park et al., 2006).

2.9.1 Affection

Affection is consumer’s feeling of peace, love, and friendliness to a

brand (Thompson et al., 2005). Affection represents emotions such as love,
friendly and peaceful feelings in which characterize their feelings toward
special objects or in the notion of brand is a specific consumption object
(Thompson et al., 2005). Affection can also be viewed in behaviours
representation in which often, affection developed to certain object (e.g.
brands) symbolize affection that the individual has or had toward a person
or memories that is recalled upon having experience with the object or
brands. (Thompson et al., 2005).

Universitas Indonesia
 30

2.9.2 Connection

Connection is consumer’s feeling of bonds or being joined with a

brand (Thompson et al., 2005). Connection represents feelings such as being
personally connected, bonded and attached towards a brand. In terms of its
representation in a form of behaviour, connection represents behaviours to
maintain proximity toward the attachment object (e.g. brands) that reveal
the level of attachments (Thompson et al., 2005). Connection is derived
from the human nature that often seek physical or psychological protection
from the attachment object and moreover it can also be related towards the
distress it can result from separation threat from the attachment object
(Bowlby, 1980; Hazan & Zeifman, 1999).

2.9.3 Passion

Passion is consumer’s feeling of delight and captivation toward a

brand (Thompson et al., 2005) and therefore represents multiple facets of
being passionate, delighted, and captivated towards a specific object (e.g.
brand). In which represents behaviour impulse behaviour in terms of their
reactions toward the brand for example in terms of impulsive purchase and
exceeding allocation of budget to purchase the goods as a form of
representation that the individuals forego his or her self-interest to obtain
the specific attachment object (Thompson et al., 2005).

2.10 Perceived Quality

Quality is often represented by a comparison of product quality and its

nominal price. Based on this concept, the bundle of attributes within a product
represent the level of quality that can provide values toward the consumers
(Lancaster, 1971). In defining perceived quality, the concept is closely related to
the user-based approach of Garvin (1983) in which it states that perceived quality
is different from objective quality as it is made within the consumer’s own
subjective evaluations of quality (Zeithaml, 1988). Zeithaml (1988) also identifies
that perceived quality is one of brand value’s component in which with a high
perceived quality it will assist consumers to choose a brand rather than other

Universitas Indonesia
 31

alternatives. Previous research also highlight that a high perceived quality is the
result of consumer’s long-term experience with the brands where it enables the
consumers to recognize the brands differentiation and superiority relative to other
brands (Zeithaml 1988). In line with previous research by Zeithaml (1988), Prior
research founds that producers are competing to produce high quality products to
create good perception and experience for their customer. (Kemp, 2005; Kyoon and
Ah Park, 2007; Raj and Roy, 2015).
Furthermore, it is also implied that brands create good perception on their
quality towards their customer in order to display superiority in which allows
consumers compare their brand with the alternatives (Aaker, 1991). Therefore,
upon defining perceived quality, this research follows the definition by Zeithaml
(1988) which implies perceived quality as a consumer judgement regarding product
or service superiority and excellence and a positive perception towards a brand
feature such as their product or service that are closely linked to the consumers
preference, satisfaction and purchase choice. Within the field of advertising, Yoo
et al. (2000) also identifies that a brand that is highly advertised usually is perceived
with high quality. In the online social media context, particularly personalised
advertising in social media, research by Shanahan et al. (2019) found that
personalised advertising positively improves consumer’s perception of quality in a
brand, with a tendency to perceive the brand as high quality. In similar vein, social
media is a significant source of information for consumers in which brands can
reflect their product quality and also enabling the consumer’s to promote the
perceived quality to other consumers through their own user generated content (Li
and Bernoff, 2011; Schivinsky and Dabrowski, 2015).

2.11 Brand Loyalty

Brand loyalty can be defined as a deep commitment to rebuy or repatronize

a product or service in consistent manner in the future. In which it causes repetitive
purchase toward a product or service from a same brand despite of other situational
influence or marketing that can cause switching behaviour (Oliver, 1999). Loyalty
can be categorized as two type of loyalty which are behavioural and attitudinal
loyalty, in which behavioural loyalty is when a repeated purchase toward a specific

Universitas Indonesia
 32

brand is performed whereas attitudinal brand loyalty is the degree of commitment

toward a unique value associated with the brand (Aaker1991;Assael 1998; Day
1969; Jacoby and Chestnut 1978; Jacoby and Kyner 1973; Oliver 1999).

Attitudinal loyalty is also described as a consumer’s commitment toward a

focal brand and its intention to repurchase the brand’s product (Russell-Bennett et
al., 2007). Within this research, attitudinal loyalty is a key determinant to be
analysed in which it tries to capture the overall brand loyalty towards a specific
brand rather and its actual purchase toward a brand in behavioural context related
purchase frequency (Yoo & Donthu, 2001). Brand loyalty is also defined as a
conditioned positive response towards a brand after a purchase has been made
(Chaudhuri & Holbrook, 2001). Brand loyalty may be measured using a key
determinants of brand trust and brand affect in which two of these construct leads
to brand loyalty (Fournier, 2013)

Universitas Indonesia
 CHAPTER III

RESEARCH METHODOLOGY

3.1 Research Model

Figure 3.1 Research Model

Source: Adapted from (Shanahan et al., 2019)
The research model used on this research is an adaptation of a research
model developed by Shanahan et al. (2019). The researcher replicates and
structured the research model by connecting all variables, dimensions, and the
representing indicators based on the acquired theories as represented in figure 3.1.
The model explains that exogenous construct that acts as independent variables
consisting of perceived personalization and self-expressive brand whilst Consumer
brand engagement and brand attachments that acts as both endogenous variable and
intervening variable which are proposed to cause consequences toward both
endogenous variables namely perceived quality and brand loyalty within the path
model that represents the impacted brand perceptions and attitude by the consumers
(Shanahan et al., 2019). The model were developed based on the current social
media settings and gap in which that the previously stated author of the core
reference journal develop the model with intent to test the effect of personalization
effects on consumer brand perception and attitudes which are consumer brand
engagement, brand attachment, perceived quality, and brand loyalty.

33
 34

This model proposed that personalization effort carried in social media

advertisement strengthen user’s attitude towards a brand and their perception of and
loyalty to the brand itself through the initially developed positive perception toward
advertisement personalization that are perceived to match the users’ needs and/or
preferences. This research also adds one variable namely self-expressive brand that
inspired from the research by Algharabat et al. (2020) and Lee and Workman.
(2015). The rationale behind adding self-expressive brand toward the initial model
developed by Shanahan et al. (2019) is by considering this research are carried
within the nomological environment of social media particularly Instagram to study
the effect of its personalised advertisement, the current social network personalised
advertisement allows the advertisement to be incorporated with the same functions
of interactions within an organic content of social media such as like, comment, and
share. Therefore, on the social network, such interactions effort by user in social
media asserts the level of engagements in which they consider the brand to be one
of their part of online self-expression (Algharabat, 2017). Following rationale by
Carroll and Ahuvia (2006), customer often uses the brand symbolic design to shape
and reflect their self-identities. This rationale is also in line with the consumer
culture theory that asserts symbolic and experiential features experienced by the
consumer during interaction with a brand that can enhance one’s self-identity
(Brodie et al., 2013) which justify the conceptual relationship of self-expressive
brand and consumer brand engagement. Furthermore, relationship theory which
asserts that brand relationships facilitate oneself with the opportunities to enhance
someone’s definition of the self by the formation of value, worth, social identity
and peer evaluations that indicate social impression formations (Fournier, 1998).
Social impression formation is affected by cues displayed and used by brands
carried with a proper linguistic power to express and relate to one’s personal life
through the cues within marketing activities (Lee and Workman, 2015).
Following the notion of Instagram advertisement as a form of personalised
advertisement, these marketing activities as a cues are the personalised
advertisement carried by the selected brands on this research which provide both
interactive experience for the automatic implementation of interaction function for
users to express themselves in the social media, and in which social media,

Universitas Indonesia
 35

particularly Instagram as visual-first photo and video sharing social media allow
brands to enhance the relationship with their consumer’s by displaying virtual
marketing cues of their brands within the advertisement to enhance the consumer’s
self-identity thus allowing themselves to express their ideal-self in social media
environment. Therefore, the researcher adds self-expressive brand toward the
research model.

3.2 Research Hypotheses

3.2.1 The Influence of Perceived Personalization on CBE and BA

Following the findings that perceived personalization is dependent

on how the personalized message matches the customer’s preferences (Li,
2016). Several researches also point out that a personalized message
significantly affect the way consumer sees a brand. For example, (Vesanen,
2007) found that when an advertisement is personalized according to
consumer preferences, personalization can help differentiate the ads from
spam in which it benefits consumer brand engagement. Research by
Shanahan et al., (2019) also highlight that perceived personalization is
predictor of consumer brand engagement in social media. In addition to that,
personalized advertisement are also beneficial as it allows brand or
organization to foster its relationship with the customers and therefore
achieve a higher customer engagement (Maslowska et al., 2016).
Brands also began to use personalized advertisement using social
media to foster their customer brand attachment (Whan Park et al., 2010).
Past research also found that advertisement that is based on consumer
preferences, or categorized as personalised message like purchase
recommendations, personal price offerings can positively impact
attachment toward the brand (Zhang et al., 2015). Research by Shanahan et
al., (2019) also found that when an organization utilize consumer preference
data such as past purchase and recent browsing history to create a
personalized advertisement the perceived personalized message within the
advertisement can positively impact the formation of consumer emotional

Universitas Indonesia
 36

attachment towards the brand. Referring to the past research findings above,
several hypotheses can be developed such as.

H1: Perceived personalization positively influence consumer brand

engagement

H2: Perceived personalization positively influence brand attachment

3.2.2 The Influence of Self-Expressive Brand on CBE and BA

Self -expressive brand can be defined as the degree of how consumer

perceive a specific brand reflects his or her inner self (Carroll & Ahuvia,
2006). Past research founds that consumers on a particular brand have a
tendency to think about a particular brand as their extension of self through
the brand’s symbolic design that helps consumer to construct their self-
identities (Belk, 1988). Sprott et al. (2009) also founds that consumer’s
perception on brands that is self-expressive partly impact their engagement.
The same general conclusion was also highlighted by France et al. (2016)
in which the research carried found that there is a positive relationship
between self-congruity of a brand and Consumer Brand Engagement
Research carried by (Algharabat et al., 2020; Leckie et al., 2016) also found
that self-expressive brand is a direct antecedents that predicts consumer
brand engagement.

In addition to that, Kim & Sherman (2007) founds that when a brand
can be congruent with one’s individual self-concept and in which it allows
them to express himself or herself, it can lead to increased commitment
toward the brand. Previous research by Lee and Workman (2015) also
highlight that there is a positive relationship between self-expressive brand
and brand attachment. Lee and Workman (2015) also stated that a brand can
help customer to better define and enhance their inner self and image
through product promotion in which also help the consumers to acquire the
products easily. Therefore, in the context of personalised advertisement, the
researcher hypothesizes

Universitas Indonesia
 37

H3: Self-Expressive brand positively influence consumer brand

engagement

H4: Self-Expressive brand positively influence brand attachment

3.2.3 The Influence of Consumer Brand Engagement on PQ and BL

Nikhashemi et al (2017) stated that consumer brand engagement is

related to perceived quality as the purpose of enhancing consumer brand
engagement and perceived quality are to stimulate a positive response
through targeting consumer preferences. Previous research founds that
consumer brand engagement in social media can have a positive effect
toward the brand perceived quality as consumers are given a chance to voice
their opinions regarding the quality of the products or service from a
brand(Shanahan et al., 2019). In similar vein, Previous research by
Schivinski & Dabrowski (2015) highlight that user-generated social media
brand communication effort can foster a high brand perceived quality
through user-generated social media brand communication where it acts as
source of information where consumers can voice the brand perceived
quality.

In addition to that, a research by previously mentioned Shanahan et

al., (2019) also found that consumer brand engagement has a positive impact
toward the formation of deep commitment toward a brand (brand loyalty)
in the context of personalised advertisement on Facebook. Schivinski &
Dabrowski (2015) research also found that user-generated social media
brand communication has a positive impact on brand loyalty. It is also
inferred that brands that is able to stimulate their consumers to interact with
the brand through social media tend to have a better customer loyalty and
higher profitability (Hollebeek et al., 2014). Therefore, from several past
research finding above, the researcher develops several hypotheses as
follows.

H5: Consumer brand engagement positively influence perceived

quality

Universitas Indonesia
 38

H6: Consumer brand engagement positively influence brand loyalty

3.2.4 The Influence of Brand Attachment on PQ and BL

Referring to a prior research, consumers attachment towards a brand

are based on their emotion in which it drives purchase behaviours (Whan
Park et al., 2010). Previous research by Shanahan et al. (2019) highlight that
there is a positive relationship between emotionally led brand attachment
towards the judgemental consideration on quality of consumers toward a
brand. Zeithaml (1988) also describe that perceived quality is subjective in
which are closely related on the consumers unique needs and the length of
experience between the consumers and the brand.

Next, previous research also found that attachment is a predictor of

repurchase behaviours (Chaudhuri & Holbrook, 2001). Where in similar
conclusion, brand attachment drives repurchase behaviour through its two
factors which are brand connection and brand prominence which translates
to the extent of attachment between a brand and its customer and positive
feelings accompanied by memories toward a brand.(Whan Park et al.,
2010). Prior research also identify that brand attachment influence how
consumer behave during purchase activity, in which as consumers get more
attached towards a brand it is more likely that consumers will purchase the
products from the same brand in the future (Akçura et al., 2004; Franz‐
Rudolf, 2006; Lee & Workman, 2015; Shanahan et al., 2019) Therefore,
referring to the findings above, several hypothesis are developed as follows:

H7: Brand attachment positively influence perceived quality

H8: Brand attachment positively influence brand loyalty

3.2.5 The Influence of Perceived Quality on BL

Following the definition of brand loyalty as a deep commitment to

conduct repetitive purchase toward a product or service from a brand
(Oliver, 1999), a prior research identify that brand loyalty is often caused
when a consumer is satisfied with the quality of a product or service in

Universitas Indonesia
 39

which a certain product or service generate unique value that other brand
could not imitate (Chaudhuri & Holbrook, 2001). Similar findings also
highlight that there is a positive relationship between the quality subjective
judgement of a brand toward brand loyalty as indicated with repetitive
purchase behaviour (Erdoǧmuş & Büdeyri-Turan, 2012; Shanahan et al.,
2019).

H9: Perceived quality positively influence Brand Loyalty

3.3 Data Collection Method

Researcher collect two type of data which are primary and secondary data.
Secondary data that is used by the researcher is used to construct research
background and to help the researcher understand the current condition of social
media and local footwear brand in the current days. In terms of primary data, the
researcher will use a primary data that will be collected directly by the researcher
in order to answer the research problem (Malhotra, 2010). Primary data is
categorized into two forms which are qualitative data and quantitative data. On this
research, researcher will be using quantitative data. The data collection method for
this research is self-administered survey. Under self-administered survey method,
a questionnaire will be distributed online through social media where respondent
will be directed to the questionnaire page in a form of Google Form to be filled
individually by each of the respondents.
In terms of questionnaire items, this research will use a structured question
with a close-ended type of question for all the questions. In addition to that, the
questionnaire items will require the respondents to fill in their demographic data.
To measure the variables this research will adopt measurement scales from previous
studies. Perceived personalization will be measured using an adaptation from one
of the component of 8C Framework specifically customization by (Srinivasan et
al., 2002), Self-expressive brands adopted from (Carroll & Ahuvia, 2006),
consumer-brand engagement scales will be adapted from (Hollebeek et al., 2014),
then brand attachment adapted from (Thompson et al., 2005), lastly perceived
quality and brand loyalty measurement scale is adapted from (Yoo & Donthu,
2001). Since all the constructs will be asked in the questionnaire using 7-point

Universitas Indonesia
 40

Likert scale, the responses will be recorded following the scale (1) will be “strongly
disagree” and up to scale (7) being “strongly agree”.

3.4 Research Design

Malhotra (2010) on the literature book of Marketing Research define

research design as a structure to carry out a marketing research. Research design is
used to determine the necessary procedure in order to acquire information that will
be used to answer the research problem within the research. This research will use
a conclusive research design in purpose of measuring the investigated phenomenon
that has been defined by the researcher through testing the proposed hypothesis and
the relationship between each variable. Therefore, it is essential for this research to
have a structured research process, representative samples, and quantitative type of
data.
Referring to Malhotra (2010), this research will be conducted as descriptive
research in form of cross-sectional design research in which the data collected for
this research will be done towards a population and sample on one period.
Descriptive research is used to test the relationship between perceived
personalization and self-expressive brand as independent variable towards
consumer brand engagement and brand attachment. In addition to that, the research
will identify how consumer brand engagement and brand attachment effect towards
its outcomes namely perceived quality and brand loyalty in the context of local
footwear brand that mainly uses Instagram as its digital marketing tools.

Universitas Indonesia
 41

3.5 Variable Operationalization

Table 3.1 Variable Operationalization

Scale: Likert 1-7

Variable Definition Dimension Code Indicator Source
Advertisement
of a brand on
Instagram
creates
PP1 purchase
recommendati
on that
matches
oneself needs
Advertisement
Personalization of a brand on
Instagram
is defined as
enables the
delivering a PP2
possibilities to
personalized Srinivasan
acquire a
Perceived advertisement to tailormade et al.,
Personalization individuals products (2002) in
N/A
based on their Shanahan
(PP) exclusive The overall et al.
preferences advertisement (2019)
of a brand on
Instagram is
PP3
tailored
(Li, 2016)
according to
oneself
situation

Feeling of
being a unique
customer due
to the
PP4
exposure of
advertisement
from a brand
in Instagram

Universitas Indonesia
 42

Table 3.1 Variable Operationalization (Continued)

Scale: Likert 1-7
Variable Definition Dimension Code Indicator Source
Beliefs that an
advertisement
from a brand
PP5
in Instagram is
customized to
oneself needs

Advertisement
of a brand on
Instagram
symbolizes
SEB 1
what kind of
person
someone really
is
Advertisement
of a brand on
Self-Expressive SEB 2 Instagram
brand is a Carroll
reflect oneself
consumer &Ahuvia,
personality
(2006) in
perception of
Advertisement Algharabat
the degree to
of a brand on et al (2020)
Self-Expressive which a brand
N/A Instagram is
Brand (SEB) reflects his or SEB 3
viewed as an
her inner self
extension of
the inner self

(Carroll and Advertisement

Ahuvia, 2006) of a brand on
Instagram
SEB 4
mirror the real
version of
oneself
Advertisement
of a brand on
Instagram
SEB 5
contributes to
the image of
oneself

Universitas Indonesia
 43

Table 3.1 Variable Operationalization (Continued)

Scale: Likert 1-7
Variable Definition Dimension Code Indicator Source
Seeing
advertisement
of a brand
Consumer brand Cognitive ECP1
causes oneself
engagement is Processing
to think about
defined as a Dimension the brand
consumer’s (ECP)
positively Thinking
The about the
valenced consumer’s brand a lot
cognitive, level of
Consumer when seeing
emotional, and brand-related ECP2
Brand an
behavioural thought advertisement
Engagement
brand-related processing of that brand Hollebeek
(CBE) activity during, and on Instagram et al.
or related to a elaboration (2014) in
consumer and Stimulation of
within their Shanahan
brand interest to
interaction et al (2019)
learn more
interactions with a brand about the
(Hollebeek et (Hollebeek et ECP3 brand after
al., 2014) seeing
al., 2014)
advertisement
of that brand
on Instagram

Universitas Indonesia
 44

Table 3.1 Variable Operationalization (Continued)

Scale: Likert 1-7
Variable Definition Dimension Code Indicator Source
Positive
feelings when
exposed to
EAF1
advertisement
Affection from a brand
on Instagram
Dimension
(EAF) Seeing
advertisement
The
from brand on
consumer’s EAF2
Instagram
degree of Hollebeek
causes oneself et al.,
positive to be happy (2014) in
brand-related
Good feelings Shanahan
affect within
when exposed et al (2019)
consumers
and brand to
EAF3
advertisement
interaction
from a brand
(Hollebeek et on Instagram
al., 2014)
Proud when
seeing
EAF4 advertisement
from a brand
on Instagram

Universitas Indonesia
 45

Table 3.1 Variable Operationalization (Continued)

Scale: Likert 1-7
Variable Definition Dimension Code Indicator Source
Spending a lot
of time seeing
an
advertisement
EAC1 of one brand
on Instagram
compared to
Activation
other type of
Dimension advertisement
(EAC)
Opening the
Consumer’s advertisement
level of from a brand
energy, Hollebeek
to view more
et al.,
efforts, and EAC2 whenever
(2014) in
time spent seeing the
Shanahan
within advertisement
et al (2019)
consumers of that brand
and brand on Instagram
interaction The brand
(Hollebeek et being
al., 2014) advertised on
Instagram is
one of the
EAC3
brands oneself
usually try to
find after
seeing the
advertisement

Universitas Indonesia
 46

Table 3.1 Variable Operationalization (Continued)

Scale: Likert 1-7
Variable Definition Dimension Code Indicator Source
Feelings
characterized
by affection
toward a brand
AAF1 when seeing
an
advertisement
of that brand
on Instagram
Affection
Dimension Feelings
(AFF) characterized
by friendliness
Definition: toward a brand
Brand AAF2
attachment is Warm when seeing
the strength of advertisement Thompson
feeling a
Brand of that brand et al.,
bond connecting consumer
Attachment on Instagram (2005) in
the brand with has toward a
(BA) the self brand Feelings of Shanahan
reflected by love toward a et al (2019)
(Whan Park et brand when
affection,
al., 2010) AAF3 seeing
love, and
friendliness advertisement
of that brand
(Thompson on Instagram
et al., 2005)
Feeling
characterized
by peace
towards a
AAF4 brand when
seeing
advertisement
of that brand
on Instagram

Universitas Indonesia
 47

Table 3.1 Variable Operationalization (Continued)

Scale: Likert 1-7
Variable Definition Dimension Code Indicator Source
Feelings
characterized
by a sense of
personal
connection
Connection ACC1 towards a
Dimension brand when
(ACC) seeing
advertisement
Definition:
of that brand
Consumer’s on Instagram
feelings of Thompson
Feelings of
being joined et al.,
attachment
with the (2005) in
toward a brand
brand Shanahan
ACC2 when seeing
reflected by et al (2019)
advertisement
connection, of that brand
bond, and on Instagram
attachment
Feelings
(Thompson characterized
et al., 2005) by bond
toward a brand
ACC3
when seeing
advertisement
of that brand
on Instagram

Universitas Indonesia
 48

Table 3.1 Variable Operationalization (Continued)

Scale: Likert 1-7
Variable Definition Dimension Code Indicator Source
Feeling
characterized
Passion by passion
toward a brand
Dimension APS1
when seeing
(APS) advertisement
of that brand
Definition: on Instagram
Consumer’s Feelings
intense and characterized
Thompson
aroused by delight
et al.,
positive towards a
(2005) in
feelings APS2 brand when
Shanahan
toward a seeing
et al (2019)
brand advertisement
reflected by of that brand
passion, on Instagram
delight, and Feeling
captivation captivated
toward a brand
(Thompson
APS3 when seeing
et al., 2005)
advertisement
of that brand
on Instagram

Universitas Indonesia
 49

Table 3.1 Variable Operationalization (Continued)

Scale: Likert 1-7
Variable Definition Dimension Code Indicator Source
Perceiving a
high quality
PQ1
towards a
brand

perceive that
the likely
Perceived quality quality of a
PQ2
brand is
is a subjective
extremely
judgement
high
regarding
superiority and perception
excellence of a that the Yoo and
product or service likelihood of Donthu
Perceived PQ3 a brand (2001) in
with a positive N/A
Quality (PQ) would be Shanahan
perception that is
functional is et al (2019)
linked to
consumers very high
preference, perceive that
satisfaction and the
purchase choice likelihood of
PQ4
a brand to be
(Zeithaml, 1988) reliable is
very high.

Perception
that a brand
PQ5 must be of
very good
quality

Universitas Indonesia
 50

Table 3.1 Variable Operationalization (Continued)

Scale: Likert 1-7
Variable Definition Dimension Code Indicator Source
The
Brand loyalty is consideration
BL1 that oneself is
a deep
loyal towards
commitment
a brand
that causes
repetitive Placing a
purchase toward brand as a first Yoo and
a product or BL2 choice than Donthu
Brand Loyalty any other (2001) in
service from a N/A
(BL) brand Shanahan
same brand
et al (2019)
despite of other Consideration
situational on not
influence or purchasing
marketing other brands if
BL3
effort. a specific
brand is still
(Oliver, 1999) available at the
store

(Source: Researcher’s Output)

3.6 Questionnaire Development

The questionnaire that will be used on this research will be formed with a
structured question consisted of four part:
1) Introduction
The researcher will first conduct introduction starting from his name, university
of origin, and the faculty completed by the researcher’s study concentration.
Here, the researcher will briefly explain the research and purpose of the
research. The researcher will also ask the respondents to answer the
questionnaire according to their own individual experience.
2) Screening Question
In order to filter the respondents, the researcher will set a screening questions
so that the respondents that will fill the questionnaire fits with the requirement
of respondent characteristics. The screening process will be conducted through

Universitas Indonesia
 51

a question on whether the respondents is among the age group of 18-34 years
old and whether the respondents uses Instagram accounts, followers on one of
the footwear brand on Instagram, owned a products from one of the brands and
live within the designated area of Java island.
3) Measurement question
On this part, the questionnaire will be divided into five sub-section where each
section will be covering each of the variables present on this research from
perceived personalization, self-expressive brand, consumer-brand engagement,
brand attachment, perceived quality, and brand loyalty. The questionnaire item
on this research will be using structured question with close-ended type of
question and are using 7-point Likert-scale with interval data scale for all the
measurement question except the question within the respondents profile
section that will be asking about the demographical data of the respondents. The
measurement scale of perceived personalization is adapted from (Srinivasan et
al., 2002), measurement scale of consumer brand engagement is adapted from
(Hollebeek et al., 2014), brand attachment from (Thompson et al., 2005)
perceived quality and Brand Loyalty from (Yoo & Donthu, 2001). The
additional variable as part of the updated conceptual model, namely self-
expressive brand will be adapting a measurement scale from (Carroll & Ahuvia,
2006). All measurement scale follows wording adjustments made within the
core reference of this research from Shanahan et al., (2019) to measure the role
of personalized advertisement (Instagram ads) in Instagram in which that the
respondents ever come across with (sees and interact with). In addition to that
self-expressive brand also follows the same adjustment in which the
respondents will refer towards the personalised advertisement particularly the
contents to measure the self-expressiveness in which reflected by the consumers
in Instagram (social media). All the measurement item uses 7-point Likert scale
in which the researcher uses scale (1) that represents strongly disagree up to the
scale of (7) that represents strongly agree responses.

Universitas Indonesia
 52

4) Respondent Profiles

On this part, the researcher will be asking the respondents regarding their brief
profile that includes gender, age, current or last education level, profession,
domicile, marital status, monthly spending, brands that they have purchased
(can be more than one brands if the respondents own products from the brands
selection), social media platform aside from Instagram that they use, and lastly
their average time spent on social media.

3.7 Sampling

Malhotra (2010) states that sampling is done by conducting a survey to a

subgroup of the population determined, which is called the sample. In the selection
of respondents, Non-probability sampling techniques (Malhotra, 2010) will be used
in this study. The researcher personal assessment will be the basis for selecting
respondents. The research will use convenience sampling considering the
accessibility of the researcher with the respondent while also considering the
requirements of the respondent that meets the criteria and fit the research context
(Malhotra, 2010). In this case, the determination of criteria is also called the
purposive sampling method. The researcher uses this technique to filter the
respondents that is in accordance with research scope and respondent profiles.
Respondents from this study, as discussed in the section on the scope of the study
are Indonesian residents who live in Java island, and have criteria aged 18 to 34
years old. The minimum number of respondents for this study will refer to Hair et
al. (2006) whereby the minimum respondents needed for this research will be the
number of items multiplied by five. Within this research, there is 38 items that will
be measured. Therefore, the minimum number of respondents required is 190
respondents.

3.8 Data Analysis Method

3.8.1 Wording Test Questionnaire Analysis

The questionnaire that will be distributed will first be adapted and

translated by the researcher from the previous research so that respondents
will understand clearly the context of the questions from each item of the

Universitas Indonesia
 53

questionnaire. To avoid error on interpreting the questions, it is essential for

the words and sentence to be clear and concise and understandable by
respondents. (Malhotra, 2010). Therefore, the researcher will implement a
wording test to ensure respondents understand all the questionnaire items
and to identify if there is ambiguity on the questionnaire items.

3.8.2 Pre-test Questionnaire Analysis

Pre-test will also be done by the researcher by distributing the

questionnaire towards 30 respondents that fits the research sample
characteristics from the research study populations. Pre-test will be done to
measure validity and reliability of each construct used on this research.
From a validity test, the researcher will be able to identify how accurate the
construct is being represented by each question that covers all research
variable within the questionnaire (Malhotra, 2010). Validity test is done by
ensuring that the value of Kaiser-Meyer-Olkin (KMO ≥ 0,5), Bartlett’s test
of Sphericity with the requirement ≤ 0,5, and the value of factor loading
from each indicator should be ≥ 0,5, If the requirements are fulfilled, then
the variable used on the research is valid (Malhotra, 2010).

Within the pre-test, the researcher will also conduct reliability test
to measure how consistent the measurement scale used and to identify if the
measurement tools are reliable for data analysis. The reliability test will be
carried out by seeing whether the value of Cronbach’s Alpha is ≥ 0,6 where
the measurement scale can be valued as reliable (Malhotra, 2010).

3.8.3 Frequency Distribution Analysis and Descriptive Statistics

Frequency distribution analysis will also be conducted by counting

the number of respondent and analysing the frequency distribution from
demographical information of respondents such as gender, age, current or
last education level, marital status, monthly spending, length of time
respondents been using Instagram as social media, latest brands the
respondents purchase, purchased brands, length of time respondents
followed the brand on Instagram, and length of time spent on social media.

Universitas Indonesia
 54

The analysis is conducted to measure how many respondents can be

associated towards certain characteristic where later it can be shown in a
form of percentage from the overall respondents (Malhotra, 2010). In
addition to that, the researcher will also conduct descriptive statistical
analysis from each of variable and dimension regarding the responses on the
questionnaire based on the minimum, maximum, standard deviation, mean,
and grand mean of each variable to generalize the responses from each of
the respondents to the questionnaire main questions(Malhotra, 2010).

3.8.4 Structural Equation Modelling (SEM) with PLS-SEM

According to Hair et al. (2017), structural equation modelling or

SEM is a second-generation multivariate data analysis method. Multivariate
data analysis involves the application of statistical methods that
simultaneously analyse several variables that represent measurements
related to individuals, companies, events, activities, situations, and so on.
This study uses SEM because there are several relationships between
variables that need to be tested simultaneously, the need for structural model
testing and measurement and the modification of the research model with
the addition of variables. There are two types of SEM, namely covariance-
based SEM (CB SEM) and variance-based SEM (PLS-SEM). This research
will be carried out using variance-based modelling SEM Partial Least
Squares - Structural Equation Modelling (PLS-SEM) using Smart-PLS
application. The use of PLS-SEM requires certain reasoning to explain the
reasons behind its use (Henseler & Chin, 2010).

According to Hair et al. (2012), in marketing disciplines, there are

three main reasons for research using PLS-SEM, namely abnormal data
distribution, small sample size, and there are indicators with formative
measurements. Referring to the type of measurement scale Malhotra (2010),
in this research there is one type of data processed, namely interval scale
data which is categorized as a reflective measurement for questions using
7-point Likert scale. In the SEM concept, there are two types of variables in
the study, namely latent variables or variables that cannot be observed, and

Universitas Indonesia
 55

observed or measured variables. To measure latent variables, the

measurement is done by measuring indicators associated with these latent
variables, which are the observed variables. The latent variables in this study
are perceived personalization, self-expressive brand, consumer-brand
engagement, brand attachment, perceived quality, and brand loyalty. Then,
the indicators of each of these latent variables in the form of question
indicators in the questionnaire are observable variables. There is different
treatment for Consumer-brand engagement and brand attachment as these
latent variables are second order construct. The treatment will consist of
averaging all of the first-order indicators and the average score will be used
as indicators in the structural analysis after the removal of low or cross-
loading items or in other words, by carrying the process of two-staged
approach in treating reflective-reflective endogenous higher order construct
(Hair Jr et al., 2016).

Hair et al., (2017) stated that within the SEM model analysis, there
are two types of models, namely the measurement model that models each
latent variable to be the basis for the formation of observable variables, as
well as the structural model that describes the relationship between the latent
variables to examine the relationship. Based on the guidelines of Hair et al.
(2017), researchers must follow the process of processing data using PLS-
SEM, that is, after making the model specifications, proceed with the
evaluation of outer model then the evaluation of inner model. Inner model
or structural model shows the relationship between the constructs that is
evaluated. Meanwhile, the outer model, also known as the measurement
model, is used to evaluate the relationship between indicator variables and
related constructs.

Furthermore, Hair et al., (2017) also imply that the result of data
analysis conducted using PLS-SEM should be evaluated through the series
of sequence in order acquire an accurate result from the data analysis, the
sequence need to be followed from evaluating the quality of measurement
model and the structural model of PLS-SEM by seeing the metric that
indicates the predictive ability of the model. Regarding the evaluation

Universitas Indonesia
 56

sequence, the sequence follows a similar process of evaluation of CB-SEM

in which the important metrics for PLS-SEM measurement model
evaluation is reliability, validity, convergent validity, and discriminant
validity. For evaluating the structural model, the important evaluation
measures are R2 (variance explained), f2 (effect size), Q2 (predictive
relevance), and the significance of the path coefficient (Hair et al., 2017).
PLS-SEM are different than CB-SEM in which it relies on the evaluation of
variance instead of covariance on CB-SEM. Therefore, upon using PLS-
SEM the measurement of covariance-based Goodness-of-Fit (GoF) could
not be performed in PLS-SEM. Therefore, within the structural model
evaluation, PLS-SEM estimate the parameter by maximizing the value of
explained variance within the endogenous latent variable, where on the
other hand, on CB-SEM it estimates the structural model parameter by
minimizing the differences between sample covariance that is represented
by the theoretical model (Hair et al., 2017).

3.8.4.1 Hierarchical Component Models: CBE and BA

Hierarchical models or higher order constructs is defined as

constructs that involves more than one dimensions, where each
dimension captures some proportion of the second-order latent
variables (Becker et al., 2012). Scholars imply that the use of higher-
order constructs allow for more theoretical parsimony and also help
to shrink the model complexity (Wetzels et al., 2009). Within the
hierarchical model the indicators and the constructs either could be
represented as reflective or formative model and as of now, there are
four different combinations of hierarchical component model form
(Ringle et al., 2012; Jarvis et al., 2003). These are: reflective-
reflective, type I model; reflective formative, type II; formative-
reflective, type III; and formative-formative, type IV.

Furthermore, there are three approaches to assess

hierarchical orders in PLS-SEM have been proposed in the
literature: first, the repeated indicator approach (Lohmöller, 2013);

Universitas Indonesia
 57

second, two-stage approach (Ringle et al., 2012) and third, the

hybrid approach (Becker et al., 2012). Within the repeated indicator
approach, the indicators of the first order variables are used in the
second or higher order constructs. While the two-stage approach, the
initial repeated order indicator approach is first used to obtain the
latent variable scores within the first stage and then the first order
constructs (dimensions) are subsequently used as indicators for the
respective constructs in the separate second stage analysis (Becker
et al., 2012; Ringle et al., 2012). The hybrid approach splits the
indicators of first order variable and uses one half to estimate the
first order variable and the other half to estimate the second order
construct (Becker et al., 2012).

The repeated indicator approach followed by the two-stage

approach is used to measure both consumer brand engagement and
brand attachment construct which are an endogenous reflective-
reflective (Type I) second order model. The consideration on using
the two-stage approach on this research is due to the recent
recommendation that when the higher order construct is an
endogenous variable within the path model, the repeated indicator
followed by two staged approach thus is recommended to estimate
the second order construct (Cheah et al., 2019; Sarstedt et al., 2019).

3.8.4.2 Measurement Model Analysis (Outer Model)

Table 3.2 Evaluation of PLS-SEM Measurement Model

Reflective Measurement Model Evaluation (Outer Model)

• Internal Consistency (Cronbach’s alpha, Composite reliability)

• Convergent Validity (Indicator reliability, average variance
extracted)
• Discriminant Validity (Cross loading, Fornell-Larcker, HTMT)

(Source: Hair et al., 2017; Henseler et al., 2010)

Universitas Indonesia
 58

Measurement Model or outer model assessment is done to

identify whether the observable variable (indicators) is accurate to
measure latent variable or those variables that are not observable.
After the outer and inner models are formed in the model
specification stage, the next step is to run the PLS-SEM algorithm
(Henseler et al., 2012) to evaluate the feasibility of the model by
looking at the level of reliability and validity of the measurement
constructs in the outer model. By evaluating the outer model,
researchers can believe that the constructs that form the basis for
assessing the relationships of the inner models are measured and
represented accurately by their items or indicators. The
measurement model was tested with the PLS algorithm function in
the Smart-PLS application.

According to Hair et al., (2017), the evaluation of models on

reflective measurement variables is done by evaluating internal
consistency, convergent validity, and discriminant validity. The
results of the evaluation of internal consistency or reliability are used
to measure the level of consistency of the measurement by
evaluating Cronbach's alpha values of each latent variable that needs
to be ascertained to have a value of ≥ 0,6 (Malhotra, 2007) or ≥ 0,7
(Tenenhaus et al., 2005) to be considered reliable. Cronbach's alpha
values indicate the reliability estimation based on the
interrelationships between indicators of the observed variables.
Then, the level of consistency of the measurement model also needs
to be evaluated by looking at the value of composite reliability which
must be ≥ 0,7. Composite reliability is obtained by combining all
variance and covariance values in a composite indicator variable
related to the construct which is then divided by the total variance in
the composite. Cronbach's alpha is a conservative measure of
reliability which often results in relatively low-reliability values.

Conversely, the composite reliability tends to overestimate

its value, resulting in a relatively higher reliability estimate.

Universitas Indonesia
 59

Therefore, researchers need to consider both criteria. According to

Hair et al (2016), the actual reliability value usually lies between the
Cronbach's alpha value which represents the lower bound value and
composite reliability which represents the upper limit value. After
that, validity testing is then performed by evaluating the convergent
validity and discriminant validity of the construct. Convergent
validity is related to the extent in which a measurement correlates
with other measurements of the same construct. Items that are
indicators of measurement of a reflective construct must be
convergent or share a high proportion of variances (Hair Jr et al.,
2016). To evaluate the convergent validity of the reflective
construct, the researcher evaluates the reliability indicator by
looking at the outer loading value and the average variance extracted
(AVE).

To be considered feasible, the outer loading value must be ≥

0,70 and the average variance extracted (AVE) value ≥ 0,50, AVE
is the total mean value (grand mean value) of the squared loading
value of a set of items or indicators (Hair et al., 2014), and AVE is
equivalent or equal to the communality of a construct. Then
discriminant validity represents the extent to which constructs differ
empirically from other constructs, in other words, constructs
measure what they want to measure. The method used to assess the
level of discriminant validity is by evaluating the cross loading value
of each indicator and by evaluating the Fornell & Larcker criteria
(Fornell & Bookstein, 1982). Discriminant validity which is
evaluated by looking at the value of cross loading (Jörg, 2009) is
done by ensuring the fulfilment of requirement within the outer
loading value in which each indicator must have the highest value,
which means that it has the strongest correlation with its own latent
construct, not with other latent constructs. Then, the Fornell &
Larcker criteria are evaluated by ensuring that the construct is valid,

Universitas Indonesia
 60

that is, with the condition that the construct shares more variance
with its own construct than with other constructs.

As an additional measure, Henseler et al. (2015) advices to

evaluate heterotrait-monotrait ratio (HTMT) within the correlation
found on the model. Heterotrait-monotrait (HTMT) ratio is the
average value of indicator correlation in among constructs that
measures different constructs. (heterotrait-heteromethod) relative
toward the geometric mean from the average indicator correlation
that measures the same construct (monotrait-heteromethod). The
approach of evaluating HTMT is done by estimating the degree of
correlation between two or more different constructs within the
model. The value of correlation disatttenuation between two or more
construct that is close or reach 1 indicates the lack of discriminant
validity. Therefore, the cut off value for HTMT on all correlation
should not reach the value of one (1).

3.8.4.3 Structural Model Analysis (Inner Model)

Table 3.3 Evaluation of PLS-SEM Structural Model

Structural Model Evaluation (Inner Model)

• Collinearity (inner VIF value)

• Coefficients of Determination (R2)
• Predictive Relevance (Q2)
• f 2 effect sizes, q2 effect sizes, SRMR
• Size and significance of path coefficients

(Source: Hair et al., 2017; Henseler et al., 2010)

Structural model analysis is done to get a conceptual

representation of each of the relationship between latent variables
(Malhotra, 2010) and is done with the structural model feasibility
test. The first step in testing the structural model is the identification
of collinearity problems that may occur in the structural model by
seeing the inner Variance Inflation Factor (VIF) value in which the

Universitas Indonesia
 61

requirement should be ≤ 0,50, Collinearity needs to be evaluated to

ensure there is no intercorrelation or collinearity between variables
(Hair et al., 2017). Then, the next step is to evaluate each
relationship within the structural model or path coefficient that
represents the construct hypothesized relationship. The path
coefficient has a standard value between –1 (negative influence) and
+1 (positive effect). The closer the estimated value of the path
coefficient to the value “0”, the weaker the relationship. The
significance of each relationship is then evaluated by looking at the
T-Value because this study uses a one-tailed test with a significance
level of 5%, so the critical value is 1.65 (Hair et al., 2017).

Next, the assessment of the model feasibility is evaluated

based on its ability to predict endogenous constructs by looking at
the coefficient of determination (R2). The determination coefficient
shows how much the variation of the independent variables can
explain the dependent variable in the research model and shows the
level of accuracy of the model predictions. R2 values range from 0
to 1, with a value of 1 representing the level of perfect predictive
accuracy (Hair et al., 2011; Jörg, 2009). In addition to evaluating the
value of R2 as a predictor of accuracy, researchers also evaluate the
Stone-Geisser Q2 value (Stone, 1974). Q2 measure is an indicator of
the predictive power of an out-of-sample model or the relevance of
a prediction. When a PLS model path has a predictive relevance
level, the path will accurately predict data that is not used in the
estimation of the model. In structural models, the relevance of
predictions on the model path is indicated by a Q2 value greater than
zero (Q2 ≥ 0) (Hair et al., 2017).

Then, the f2 effect size makes it possible to assess the

contribution of exogenous constructs to R2 values of endogenous
latent variables. Whereas, the q2 effect size makes it possible to
assess the contribution of exogenous constructs to the Q2 value of
endogenous latent variables (Hair et al., 2017). In addition to testing

Universitas Indonesia
 62

the theory represented within the model, an evaluation of

standardized root means square residual (SRMR) is used to measure
the square difference between the observed correlations and the
model implied correlations matrix. In which allow the researcher to
assess the discrepancies between observed and expected correlations
to validate a PLS-SEM model (Henseler & Sarstedt, 2013). In terms
of evaluating SRMR, SRMR values of less than 0,10 (Hu & Bentler,
1999) indicate the feasibility of a good PLS-SEM model.

Therefore, a structural model analysis conducted in this

research is purposed to provide a conceptual representation from the
relationship of each latent variable (Malhotra, 2010). This research
conducts two kind of structural model analysis which are the
feasibility of the structural model, and hypothesis testing or causal
analysis that is proposed by the researcher

Universitas Indonesia
 CHAPTER 4

ANALYSIS AND DISCUSSION

4.1 Wording Test and Pre-test Questionnaire Analysis

Regarding the questionnaire item used on this research, the item used are
adapted from various reference journal written in English in which then forward
translated by the researcher to Bahasa Indonesia. During the process of
questionnaire measurement development, the researcher translated the items by first
consulting the items to the original writer of the reference journal used within this
research. The consultation was done in order to ensure there is no misinterpretation
during the translation process especially for the constructs of Consumer Brand
Engagement and Brand Attachment in which the original author operationalize the
indicators to fit the context of advertisement personalization. After gaining
clarification on regards to the wording adjustment made to fit the context of
advertisement personalization, the original author operationalizes the words for
consumer brand engagement as consumer and brand interactions upon “seeing” the
personalised advertisement on social media and brand attachment also follow the
same adjustment. After the researcher complete the translation process, the
researcher then conducts wording test to 10 respondents by consulting directly
whether the item is understandable or not. Most of the input acquired from this
wording test stage is regarding the formation of the sentence and the selection of
representative word that is much more understandable in Bahasa Indonesia. Then,
the result of wording test is used by the researcher to fix and improve the structure
of the sentence that represent the item within the questionnaire.

After the questionnaire is seen consistent and understandable, the researcher

continues to pre-test stage by distributing the questionnaire toward 30 respondents
that fits the criteria of this research administered in google form. The criteria of the
pre-test respondents are those who own the products from the brands, followed the
brand on Instagram, minimum aged 18 and maximum aged 34, and are currently
living in Java Island area. Pre-test is done to evaluate reliability and validity of each

63
Universitas Indonesia
 64

measurement items and variables by using IBM SPSS Statistics 23. The
questionnaire item is deemed to be reliable if the value of Cronbach’s Alpha ≥ 0,6
and is valid if the value of Kaiser-Meyer-Olkin (KMO) ≥ 0,5, Bartlett’s Test of
Sphericity ≤ 0,05, MSA ≥ 0,5, and factor loading within component matrix ≥ 0,5,
Therefore, the result of reliability and validity test is described below:

Table 4.1 Result of Reliability and Validity of Pre-test Questionnaire

Reliability Test Validity Test

Variable/
Item
Dimension
Cronbach’s Bartlett's Factor
Result KMO MSA Result
Alpha Test Loading
PP1 0,465 0,688 Valid
PP2 0,691 0,815 Valid
PP PP3 0,673 Reliable 0,507 0,000 0,483 0,685 Valid
PP4 0,335 0,406 Invalid
PP5 0,574 0,754 Valid
SEB1 0,849 0,911 Valid
SEB2 0,875 0,833 Valid
SEB SEB3 0,916 Reliable 0,880 0,000 0,872 0,906 Valid
SEB4 0,903 0,849 Valid
SEB5 0,915 0,830 Valid
ECP1 0,653 0,873 Valid
ECP ECP2 0,831 Reliable 0,661 0,000 0,609 0,921 Valid
ECP3 0,774 0,800 Valid
EAF1 0,851 0,839 Valid
EAF2 0,935 0,933 Valid
CBE EAF 0,852 Reliable 0,747 0,000
EAF3 0,843 0,831 Valid
EAF4 0,714 0,674 Valid
EAC1 0,714 0,757 Valid
EAC EAC2 0,797 Reliable 0,605 0,000 0,596 0,871 Valid
EAC3 0,565 0,933 Valid
AAF1 0,836 0,878 Valid
AAF2 0,907 0,784 Valid
AAF 0,895 Reliable 0,829 0,000
AAF3 0,778 0,922 Valid
AAF4 0,829 0,899 Valid
ACC1 0,780 0,943 Valid
BA
ACC ACC2 0,941 Reliable 0,765 0,000 0,797 0,939 Valid
ACC3 0,724 0,956 Valid
APS1 0,799 0,945 Valid
APS APS2 0,947 Reliable 0,754 0,000 0,692 0,968 Valid
APS3 0,785 0,947 Valid
Source: Researcher’s IBM SPSS 23 Output

Universitas Indonesia
 65

Table 4.1 Result of Reliability and Validity of Pre-test Questionnaire

(Continued)
Reliability Test Validity Test
Variable/
Item Cronbach’s Bartlett’s Factor
Dimension Result KMO MSA Result
Alpha Test Loading

PQ1 0,873 0,905 Valid

PQ2 0,817 0,716 Valid
PQ PQ3 0,920 Reliable 0,846 0,000 0,866 0,893 Valid
PQ4 0,815 0,937 Valid
PQ5 0,851 0,898 Valid
BL1 0,737 0,914 Valid
BL BL2 0,897 Reliable 0,746 0,000 0,794 0,895 Valid
BL3 0.715 0.923 Valid
Source: Researcher’s IBM SPSS 23 Output

Referring to the result of reliability and validity testing of the pre-test

questionnaire as shown on table 4.1 above, it is found that almost all latent variable
is considered valid based on its KMO value > 0,5, Bartlett’s Test of Sphericity <
0,05, and its indicators factor loading that is > 0,5. However, it is also found that
one of the indicators of latent variable, perceived personalization, coded PP4, does
not met the validity requirement since its factor loading is below 0.5 (0,406).
Therefore, referring to the validity issue, further analysis on perceived
personalization validity measures will be further discussed on table 4.2 below.

Regarding the reliability test, all latent variables within this research can be
considered to have a good reliability. This is due to the result shown on table 4.1
where the output show all latent variable acquire a value of Cronbach’s Alpha
within the range of 0,673 up to 0,947 in which the value is above 0,6. However,
since there are validity issues discovered on one indicator of perceived
personalization latent variable (PP4), it can be said that all latent variable on this
research is reliable for the valid construct and indicators but still require a further
analysis on perceived personalization that will be discussed below.

Universitas Indonesia
 66

Table 4.2 Result of Reliability and Validity of Perceived Personalization after

PP4 Removal

Reliability Test Validity Test

Variable/
Item
Dimension
Cronbach’s Bartlett's Factor
Result KMO MSA Result
Alpha Test Loading
PP1 0,623 0,792 Valid
PP2 0,613 0,835 Valid
PP PP3 0,721 Reliable 0,631 0,000 0,696 0,679 Valid
PP4 n/a n/a Removed
PP5 0,617 0,681 Valid
Source: Researcher’s IBM SPSS 23 Output

Continuing the discussion of reliability and validity of the pre-test

questionnaire, there is one validity issue discovered on one indicator of perceived
personalization latent variable (PP4) that acquire a factor loading that is below the
suggested limit to be considered valid that is above 0.5. Therefore, a further analysis
is conducted by removing PP4 and conduct analysis on its reliability and validity
once again. After the removal of PP4, it is found that perceived personalization
latent variable and its indicator now can be considered valid as the acquired value
of Kaiser-Mayer-Olkin (KMO) > 0.5, Bartlett’s test of Sphrecity < 0.05 and its
indicator factor loading > 0.5 has been met. It is also found that after the removal
of PP4 from the reliability analysis of perceived personalization, perceived
personalization latent variable remains reliable and its reliability measure of
Cronbach’s Alpha also improve to 0,721 from the previous of 0,673. Therefore,
based on the statistical result of pre-test, PP4 are further removed from the main
PLS-SEM analysis due to validity indices that does not meet the requirement on
this pre-test phase.

4.2 Frequency Distribution Analysis

Within this research, respondents are required to answers certain questions

in regards of their individual profile. These questions cover information such as
gender, current age, level of current or last education, employment status, domicile,
their amount of monthly spending, length of time they have been following the
Brand’s Instagram account, number of brands they have purchased, their recent

Universitas Indonesia
 67

purchased brand, social media account they own, and the average time spent per
day on using social media. The total respondents gathered and are validated to be
analysed on this research are 386 respondents. Therefore, the frequency distribution
analysis are as follows.

4.2.1 Gender

Based on table 4.3, the researcher learns additional information

regarding the consumer base and followers from four of the local footwear
brands selected on this research. Based on gender, from the total of 386
respondents participate on this research, there is gender dominance from
males which counts for 304 respondents or 78,80% that fill in the
questionnaire. As the respondents are consumers and followers of the four
selected local footwear brand that advertise on Instagram, this finding could
indicate that four of the brands listed are more preferred by those that are
males. Furthermore, the rest being 21,20% from the total respondents or 82
individuals informed that they are females.

Table 4.3 Gender

Gender Frequency Percentage

Male 304 78,80%
Female 82 21,2%
(Source: Researcher’s Output)

4.2.2 Age

Based on the demographic criteria of age, respondents are

dominated by those within their early twenties (18-26 years old) that were
born within the range of 1994-2002 with 325 respondents or 89,90% from
the overall respondents. Based on age group distribution, this section
provides insights that the selected four brand listed for this research are
more preferred by those within the age range of 18-26 years old as these
groups found to be the largest base of age groups that own the brand’s
products and also follows the brand Instagram accounts. As for the rest,
there are 39 respondents falls within the age range of 27 up to 34 years old
(10,10%). Details can be seen on table 4.4.

Universitas Indonesia
 68

Table 4.4 Age

Age Frequency Percentage

18 36 9,30%
19 44 11,40%
20 51 13,20%
21 61 15,80%
22 52 13,50%
23 40 10,40%
24 24 6,20%
25 17 4,40%
26 22 5,70%
27 6 1,60%
28 6 1,60%
29 9 2,30%
30 4 1,00%
31 8 2,10%
32 2 0,50%
33 3 0,80%
34 1 0,30%
(Source: Researcher’s Output)

4.2.3 Level of Last Pursued Education or Current Pursued

Education

Based on their level of last pursued education as presented on table

4.5, research respondents of this research are dominated by those who are
lastly pursuing or currently conducting study on bachelor’s degree level
with the amount of 276 individuals or 71,50% from the total of 386
respondents. Then, the rest of 73 respondents or 18,90% from the total
samples is lastly pursuing high school or of the similar level of education,
23 respondents among the Diploma level, and 14 within the level of master’s
degree.

Universitas Indonesia
 69

Table 4.5 Level of Last Pursued Education/Current Pursued Education

Level of Education Frequency Percentage

S1 276 71,50%
SMA/Sederajat 73 18,90%
Diploma 23 6,00%
S2 14 3.60%
(Source: Researcher’s Output)

4.2.4 Employment Status

Regarding the employment information presented on table 4.6,

majority of respondents that participate on this research are largely engaged
as students by the amount of 241 respondents or 62,40% from the total
respondents. Then followed by private sector or state-owned employees by
the amount of 80 respondents or (20,70%). Furthermore, the rest of the
respondent’s employment are relatively varied in terms of sectors where 26
or (6,70%) respondents served as Civil Servant, 14 respondents or (3,60%)
self-employed as entrepreneur. Among the participants of this research,
there are also respondents (2,30%) working with a freelance scheme of
employment and eight or (2,10%) respondents worked within variety of
sector (others) such as online ride-hailing service driver (3 person), lawyer
(1 person), laborers (3 person), and one person work as a journalist. Lastly,
it is also found that eight respondents or (2,10%) informed to be currently
unemployed that participate on this research.

Table 4.6 Employment Status

Employment Status Frequency Percentage

Students 241 62,40%
Private Sector or State-Owned
80 20,70%
Enterprises Employees
Civil Servant 26 6,70%
(Source: Researcher’s Output)

Universitas Indonesia
 70

Table 4.6 Employment Status (Continued)

Employment Status Frequency Percentage

Entrepreneur 14 3,60%
Freelance 9 2,30%
Unemployed 8 2,10%
Others 8 2,10%
(Source: Researcher’s Output)

4.2.5 Domicile

Regarding the current living are or domicile of this research

participants. The respondents are asked to answer according to their city
location and state or province locations. Since this research set the
geographical domain on Java Island, then the distribution from the provided
answers within the questionnaire are classified according to the province or
administratively considered provinces where the respondents are living
within by the researchers. From the gathered data presented on table 4.7,
majority of the respondents are living on West java area and DKI Jakarta
region where in details, 161 respondents or (41,70%) that are residents of
West Java Provinces, and 100 respondents or (25,90%) are residents of DKI
Jakarta. Among the distributions of both provinces, within West Java,
majority of the respondents (153 respondent) informed to be living on
Bandung, Bekasi, Bogor, and Depok city. From this data, the dominance of
research participants number coming from both provinces are probably
caused as all the brands listed are commonly found on outlets within both
provinces and are initially launched from Bandung, West Java (Andriani,
2019). Furthermore, 42 respondents or (10,90%) are residents of Banten
Province and 40 respondents or (10,40%) are residents of East Java
Province. Lastly, 26 respondents (6,70%) and 17 respondents (4,40%) are
residents from Central Java Province and Special Region of Yogyakarta
respectively.

Universitas Indonesia
 71

Table 4.7 Domicile

Provinces Frequency Percentage

West Java 161 41,70%
DKI Jakarta 100 25,90%
Banten 42 10,90%
East Java 40 10,40%
Central Java 26 6,70%
Special Region of Yogyakarta 17 4,40%
(Source: Researcher’s Output)

4.2.6 Monthly Spending

Within this section, researcher will provide and analyse information

regarding the respondent’s monthly spending on Rupiah currency. Based on
the information given by respondents presented on table 4.8, The findings
indicate the amount of money in Rupiah respondent’s able to spend on
monthly basis are dominated by two groups of respondents. Majority of the
respondents with 151 individuals (39,10%) from the total respondents spend
roughly less than Rp1.500.000 on monthly basis which then followed with
132 respondents (34,20%) spent Rp1.500.000 - Rp3.000.000 per month.
Furthermore, there are 63 respondents (16,30%) informed that they spent
Rp3.000.000 - Rp5.000.000 per month and 31 respondents (8,00%) spent
roughly Rp5.000.000 - Rp10.000.000 within a month. And lastly, there are
nine respondents or 2,30% from the total respondents that informed to have
a monthly spending with the amount above Rp10.000.000.

Table 4.8 Monthly Spending

Monthly Spending Frequency Percentage

< Rp1.500.000 151 39,10%
Rp1.500.000 - Rp3.000.000 132 34,20%
Rp3.000,000 - Rp5.000.000 63 16,30%
Rp5.000.000 - Rp10.000.000 31 8,00%
> Rp10.000.000 9 2,30%
(Source: Researcher’s Output)

Universitas Indonesia
 72

4.2.7 Marital Status

Within this section the researcher also asked the respondents

regarding their marital status as an additional information that is in-line with
the age grouping to understand more about the respondent’s profile. Based
on the questions and grouping as presented on table 4.9, majority of
respondents by 361 participants or 93,50% from the total respondents
answered that they are single or never been married before. The very high
dominance of single marital status here can indicate that respondents which
are currently not within a legally recognized relationship (single) are more
likely to be interested to own local footwear brand product’s and follow the
brand Instagram account as part of their lifestyle. Furthermore, there are
also a small percentage of 21 respondents or 5,40% from the total
respondents that informed they are currently married and four respondents
or 1,00% from the total respondents answered that they were married.

Table 4.9 Marital Status

Marital Status Frequency Percentage

Married 21 5,40%
Was Married 4 1,00%
Single 361 93,50%
(Source: Researcher’s Output)

4.2.8 Purchased Brands

From table 4.10 below, the researcher also asked questions regarding
what the brands respondents have purchased. Within the questions, the
respondents can give more than one answers if they have purchased any of
the brands from the selection. Based on the frequency distribution analysis,
there are two brands that is found to be the most purchased brands from the
four selection of brands. The frequency distribution analysis found that
Brodo Footwear and Geoff Max is the local footwear brand are the most
purchased and owned by respondents with 182 responses (47,20%) and 171
responses (44,30%) respectively from the total respondents. Then, followed
by NAH Project with 108 responses or 28,00% and Portee Goods with 76

Universitas Indonesia
 73

responses (19,70%) from the total pool of respondents that participate on

this research.

Table 4.10 Purchased Brands

Purchased Brands Frequency Percentage from Total Sample

Brodo 182 47,20%
Geoff Max 171 44,30%
NAH Project 108 28,00%
Portee Goods 76 19,70%
(Source: Researcher’s Output)

4.2.9 Recently Bought Brands

Upon answering the rest of questionnaire, the researchers also asked

the respondents to select one brand that they purchase most recently to be
used as reference to answer the questionnaire about the brand’s personalised
ads. The recently bought brands is used as reference for respondents to
answer the questionnaire in which they own the brand’s product and follow
the brand Instagram accounts. From the findings presented on table 4.11,
the researcher learns additional information that majority of participants on
this research are those who recently purchase Geoff max and Brodo brands
product. Geoff Max with 131 participants (33,90%), followed by Brodo
Footwear with 129 respondents or 33,40% from the total respondents. The
result of this distribution also prove additional insights in which Geoff max
and Brodo are the two dominating brands as there is a high count of
participants that have bought or own the products from these brands from
the previously discussed most purchased brands distribution As for the rest,
there are 74 or 19,20% from the total respondents refer to NAH Project and
lastly 52 or 13,50% from the total respondents was referring to Portee Goods
to answer the questionnaire.

Universitas Indonesia
 74

Table 4.11 Recent Bought Brands

Recently Bought Brands Frequency Percentage

Brodo 129 33,40%
GEOFF Max 131 33,90%
NAH Project 74 19,20%
Portee Goods 52 13,50%
(Source: Researcher’s Output)

4.2.10 Length as Followers of the Brand Instagram Accounts

Within this section, the researcher also asked the respondents on

how long they have followed the brands that they recently purchased. There
are four different grouping range that is established by the researcher as
shown on table 4.12. Interpreting the results, the researcher found that
majority of respondents by 282 individuals (73,10%) have followed the
brands for more than six months until the questionnaire is administered with
a dominance within a group of respondents that followed the brand
Instagram for 7-12 months period by 142 respondents (36,80%). This
distribution provides insight as the findings indicate that majority of the
participants already become a follower of the brands Instagram account that
they refer to for quite some time. Details can be seen on table 4.12.

Table 4.12 Length as Followers of the Referred Brand on Instagram

Length as Followers Frequency Percentage

< 6 months 104 26,90%
7 – 12 months 142 36,80%
13 – 24 months 78 20,20%
> 24 months 62 16,10%
(Source: Researcher’s Output)

4.2.11 Social Media Used

Within the questions regarding social media that the respondents use
aside from Instagram as one of the main requirements to fill the
questionnaire, respondents can answer to more than one social media that
the respondent uses. From table 4.14 below, it can be inferred that all the
respondents are validated in which they use Instagram as the main

Universitas Indonesia
 75

requirements to participate on this research. Furthermore, researcher also

learn additional information that respondents are somewhat familiar with
the current advancement on social media, as large proportions of
respondents are found to be using social media listed within the multiple
options starting with messenger application, Twitter, Facebook, and
moreover LinkedIn. The following social media listed follows several of the
top used social media ranked by (Hootsuite, 2019). Details can be seen on
table 4.13 below.

Table 4.13 Social Media Used

Percentage from Total

Social Media Used Frequency
Sample
Instagram* 386 100,00%
Messenger
App:WhatsApp/Line/WeChat 319 82,60%
Twitter 251 65,00%
Facebook 248 64,20%
LinkedIn 146 37,80%
Instagram is Required to Fill Questionnaire*

(Source: Researcher’s Output)

4.2.12 Average Time Spent on Social Media per Day

Lastly, upon the questions regarding the average time spent on social
media, the frequency distribution analysis as shown on table 4.14 shows that
138 respondents or 35,80% from the total respondents spent an average
range of 3 – 5 hours per day using social media. Then 103 respondents or
26,70% spent an average range of 1 – 3 hours per day. With the large
proportion of respondents that informed they spent their time using social
media from the range of 3-5 hour per day, this can indicate that the
respondents spend a lengthy time when using social media such as
Instagram which also makes them to more likely be exposed with the
brand’s personalised advertisement on Instagram. In addition to that, 82
respondents (21,20%) answers that they spent an average range of 5 – 7
hours a day using social media, while 17 respondents or 4,40% answers that
they use social media not longer than 1 hour on daily basis.

Universitas Indonesia
 76

Table 4.14 Average Time Spent on Social Media per Day

Time Spent on Social Media Frequency Percentage

< 1 Hour 17 4,40%
1 – 3 Hour 103 26,70%
3 – 5 Hour 138 35,80%
5 – 7 Hour 82 21,20%
> 7 Hour 46 11,90%
...(Source: Researcher’s Output)

4.3 Descriptive Statistics

Descriptive analysis on this research is performed in order to give a general

picture of respondent’s answers on each of the questionnaire items. This analysis is
performed by measuring the minimum value (min) and maximum value (max) from
each answer, the standard deviation value, average value of each indicator (mean),
and the total mean of the represented constructs. Upon measuring the minimum and
maximum value, since the questionnaire item on this research uses Likert 1-7 scale,
the minimum value that each item can have is one (1) with the maximum is (7).
Therefore, with the gathered respondent’s number or N (sample) of this research
that are 386, the descriptive analysis is as follows:

Table 4.15 Descriptive Statistics

Total
Variable/Dimension Item Min. Max. Standard Deviation Mean
Mean
PP1 1 7 0,99 5,75
PP2 1 7 1,02 5,68
PP 5,53
PP3 1 7 1,18 5,25
PP5 1 7 1,25 5,43
SEB1 1 7 1,50 4,48
SEB2 1 7 1,46 4,50
SEB SEB3 1 7 1,46 4,45 4,52
SEB4 1 7 1,53 4,29
SEB5 1 7 1,50 4,90
(Source: Researcher’s Output)

Universitas Indonesia
 77

Table 4.15 Descriptive Statistics (Continued)

Standard Total
Variable/Dimension Item Min. Max. Mean
Deviation Mean
ECP1 1 7 1,15 5,68
ECP ECP2 1 7 1,25 5,47 5,61
ECP3 1 7 1,24 5,66
EAF1 1 7 1,26 5,09
EAF2 1 7 1,28 5,14
CBE EAF 5,01
EAF3 1 7 1,21 4,94
EAF4 1 7 1,40 4,87
EAC1 1 7 1,59 4,37
EAC EAC2 1 7 1,49 5,02 4,90
EAC3 1 7 1,29 5,32
AAF1 1 7 1,43 4,24
AAF2 1 7 1,38 5,11
AAF 4,60
AAF3 1 7 1,46 4,67
AAF4 1 7 1,39 4,38
ACC1 1 7 1,55 4,38
BA
ACC ACC2 1 7 1,50 4,53 4,43
ACC3 1 7 1,50 4,38
APS1 1 7 1,27 5,04
APS APS2 1 7 1,22 4,99 5,09
APS3 1 7 1,27 5,23
PQ1 2 7 1,03 5,87
PQ2 1 7 1,10 5,82
PQ PQ3 2 7 1,02 5,80 5,87
PQ4 3 7 1,01 5,85
PQ5 1 7 1,05 6,00
BL1 1 7 1,45 4,69
BL BL2 1 7 1,54 5,12 4,67
BL3 1 7 1,74 4,18
*N = 386
(Source: Researcher’s Output)

Based on Table 4.15 above that presents the result of descriptive

analysis on each items within the latent variables on this research, Perceived
Personalization acquire a total mean of 5,53, in which can be categorized as
relatively high, based on the analysis, it can be derived that respondents

Universitas Indonesia
 78

perceive that the advertisement delivered by the brand are personalised for
them. Here it can also be seen that two out of four-indicator used, PP1 and
PP2 acquire a mean value of 5,75 and 5,68 respectively which means that
in average, respondents perceive that the ads help them to acquire a product
that is recommended and suited for them. The standard deviation value
however is mostly above one except for PP1 in which indicates that the
answer given regarding each indicator within perceived personalization
constructs are also relatively varied from each individual respondent and
based on the means from each indicator there are a tendency for respondents
to answer “agak setuju” across all items.

Self-Expressive brand variable acquire a total mean value of 4,52 in

which can be categorized that the score is relatively not high given the total
mean value almost close to four. Then, it can also be seen that within this
variable, the standard deviation value is very high indicating that the
answers given by the respondents are varied. This also indicates that
respondents have different considerations when it comes to whether the
corresponding brand’s personalised advertisement reflect their inner-self,
personalities, and that it also contributes to how the respondent self-image
is.

Next, Cognitive Processing dimension of Consumer Brand

Engagement acquire a total mean value of 5,61 in which can be categorized
as relatively high. Even so, the standard deviation for all items are relatively
high which reach above one which means the answers from respondents are
varied. Therefore, it can be inferred that majority of respondents considers
to be stimulated to process thoughts or think about the brand when they
come across the brand’s personalised ads in Instagram. Out of all three
indicators used, ECP1 and ECP3 acquire the highest mean value of 5,68 and
5,66 in which are also considered to be high and ECP3 also provide some
interesting finding as most of the respondents also is stimulated to learn
more about the brand when seeing the personalised advertisement on
Instagram.

Universitas Indonesia
 79

Then, Affection dimension of Consumer Brand Engagement acquire

a total means of 5,01 in which are relatively not high. Then, within this
dimension it can also be seen that the standard deviation is high indicating
that the answers are also varied across respondents when it comes to
acknowledging a positive feeling toward a brand upon seeing the brand’s
personalised ads. From all the three-indicator used within this dimension all
three indicators namely EAF1, EAF2, EAF3 acquire an almost similar
means that indicates respondents answer the questions in between the value
of four (neutral) up to five (agak setuju).

Third, activation dimension of Consumer Brand Engagement

acquires a total means that is lower when compared to other dimension
within Consumer Brand Engagement which is 4,90 while also acquired a
high standard deviation that indicates variety of answers from respondents.
Especially EAC1, or indicator number one acquire the lowest mean of 4,37
and highest standard deviation among other indicator in the dimension while
the others (EAC2 and EAC3) still maintain to acquire a mean value of five
(5). Here, the researcher can infer that when it comes to the effort, activity,
and time spent toward a brand’s personalised ads in Instagram, respondents’
answers are somewhat varied and are close to neutral in responses to
interacting and spending more time with the brand through its personalised
ads.

The first dimension of brand attachment namely Affection acquire

a relatively low total means value of 4,60 and among indicators three
indicators namely AAF1, AAF3, AAF4 acquire a relatively low individual
means that is close to four (neutral) while AAF2 means are relatively high
as it reaches the average of five in regards to the responses within likert
scale 1 – 7. In terms of standard deviation, all four-indicator used also
acquire a high standard deviation indicating that the answers from each
respondents are varied when it comes to acknowledging their warm feeling
toward a brand as reflected with feeling of affection, friendliness, love, and
peace upon seeing the corresponding brand’s personalised advertisement.

Universitas Indonesia
 80

Second, connection dimension of brand attachment acquire quite

low total means value of 4,43 across all variable or dimension within the
questionnaire. Among the three indicators namely ACC1, ACC2, ACC3
also acquire a very high standard deviation value in which indicates that the
responses given by respondents are very varied. The means from each
indicator used are also low which indicates that majority of the respondents
are neutral when it comes to their feeling of personal connection,
attachment, and bond toward a certain brand upon seeing the personalised
ads of that brand on their Instagram.

Third, passion dimension of brand attachment acquires a relatively

high total means value of 5,09 and are the highest among three dimension
of brand attachment. Across three indicators used in passion dimension
namely APS1, APS2, APS3 all three acquire an individual means that is
close to five (5) that is also relatively high. The value of standard deviation
among the indicators are lower in relative to other standard deviation across
the brand attachment construct although the answers are still considerably
varied across respondents. Since the standard deviation value exceed the
value of one across all indicators, it can be inferred that there are variety of
responses given to respondents regarding their intense positive feelings such
as feeling passionate, happy, and captivated toward a brand upon seeing the
brand personalised advertisement on Instagram.

Next, perceived quality variable acquire the highest total mean

across all variable and dimensions within this research. Here, the total
means of perceived quality from five indicators achieve the value of 5,87.
That indicates respondents tend to perceive that the corresponding brand
that they choose are superior and excellent in terms of quality that is fitting
to their preference. Regarding the standard deviation on each of individual
indicators namely PQ1, PQ2, PQ3, PQ4, PQ5, the value is still above one
(1), however not as high as other items. Here, PQ3 and PQ4 also acquire a
minimum value of 2 and 3 respectively. Where lastly PQ5 also acquire an
individual mean value of 6,00 that is highest among all indicators.

Universitas Indonesia
 81

Lastly, brand loyalty acquires a relatively low total mean value of

4,67. From each individual indicators namely BL1, BL2, BL3 the individual
mean value of each indicator are relatively low being BL1 (4,69) and BL3
lowest by 4,18 where BL2 are relatively high with an individual means of
5,12. Regarding the standard deviation acquired, each of indicator used
within brand loyalty acquire a very high deviation in which indicates a high
variety of answers from respondents in regard to whether they acknowledge
themselves to be loyal and are in a deep commitment toward that brand.

4.4 Measurement Model Analysis (Main-Test)

4.4.1 Reflective Measurement Model Evaluation

Since all the variable namely Perceived Personalization, self-

expressive brand, consumer brand engagement, brand attachment,
perceived quality, and, brand loyalty is measured reflectively including
consumer brand engagement and brand attachment which is a hierarchical
second order construct that is measured as reflective-reflective. Six of these
variables and its dimensions within the measurement model stage will be
evaluated by seeing its internal consistency, convergent validity, and
discriminant validity. The assessment of hierarchical order with the repeated
indicator approach is shown in Table 4.16 and 4.18 to evaluate the internal
consistency of each lower order components and its cross loadings. After
the initial repeated indicator output is evaluated and there are no issues
regarding its reliability, and low cross loadings , the two-stage approach is
implemented on separate model analysis by applying the first order
dimensions into single items or as “indicators” of the second order variable
(Ringle et al., 2012), namely consumer brand engagement and brand
attachment.

Universitas Indonesia
 82

4.4.1.1 Internal Consistency Evaluation

Table 4.16 Internal Consistency Evaluation Result with Repeated Indicator

Approach

Variable Dimension Cronbach’s Alpha Composite Reliability

PP N/A 0,767 0,851
SEB N/A 0,937 0,952
ECP 0,806 0,886
CBE
EAF 0,891 0,924
EAC 0,764 0,864
AAF 0,885 0,921
BA
ACC 0,933 0,957
APS 0,897 0,935
PQ N/A 0,883 0,915
BL N/A 0,834 0,900
(Source: Researcher’s Smart-PLS Output)

Within the repeated indicator approach, the above table

presents that there are no reliability issues from each corresponding
dimensions and variable within the measurement model analysis.
The result from table 4.16 reveals the evaluation of internal
consistency or reliability within the measurement model in which
the value of composite reliability must be 0,7 or above and
Cronbach’s Alpha with the minimum value of 0,7. As described
within the table above, all of the variables and dimensions for the
second-order constructs exceed the cut off value while also maintain
values above 0,8 for both Composite reliability and Cronbach’s
Alpha except for the Cronbach’s Alpha value of Perceived
Personalization and Activation dimension of Consumer Brand
Engagement that are 0,767 and 0,764 respectively but are still within
the recommended Cronbach’s Alpha value requirements that should
be above 0,7.

Universitas Indonesia
 83

Table 4.17 Internal Consistency Evaluation Result with Two Stage Approach

Variable Cronbach's Alpha Composite Reliability

PP 0,767 0,851
SEB 0,937 0,952
CBE 0,785 0,874
BA 0,886 0,929
PQ 0,883 0,915
BL 0,834 0,900
(Source: Researcher’s Smart-PLS Output)

After evaluating the repeated indicator internal consistency,

reliability and its convergent validity to assess if there are low
loadings within the lower order components level, the separate two-
stage approach result then is analysed. From the two-stage approach
as presented on table 4.17 above, there are also no reliability issues
from each latent variable within the measurement model analysis.
The result represented on above table shows that all variables
successfully met the requirement to be considered reliable where the
minimum value of composite reliability is 0,7 and Cronbach’s alpha
with the minimum value of 0,7. Based on the table 4.18 above,
almost all variable maintain both Cronbach’s alpha and composite
reliability value above 0,8 except for perceived personalization and
consumer brand engagement Cronbach’s alpha that is 0,767 and
0,785 respectively.

4.4.1.2 Convergent Validity and Discriminant Validity

Evaluation

To evaluate measurement model validity, the assessment is

done by evaluating convergent validity that is by seeing the value of
Average Variance Extracted (AVE) that must satisfy the cut off
value ≥ 0,50 and indicator reliability by seeing the value of outer
loading that must be ≥ 0,70.

Universitas Indonesia
 84

Table 4.18 Convergent Validity and Discriminant Validity Result with

Repeated Indicator Approach

Item AVE AAF ACC APS BL EAC EAF ECP PP PQ SEB

AAF1 0,858 0,707 0,576 0,521 0,493 0,598 0,383 0,389 0,307 0,589
AAF2 0,808 0,608 0,574 0,499 0,538 0,516 0,443 0,369 0,406 0,460
0,745
AAF3 0,907 0,684 0,705 0,584 0,581 0,659 0,440 0,424 0,462 0,589
AAF4 0,877 0,666 0,581 0,511 0,447 0,594 0,366 0,401 0,362 0,527
ACC1 0,725 0,944 0,641 0,562 0,467 0,591 0,378 0,354 0,315 0,600
ACC2 0,882 0,749 0,948 0,659 0,543 0,503 0,593 0,462 0,402 0,365 0,596
ACC3 0,700 0,925 0,628 0,535 0,487 0,593 0,417 0,398 0,349 0,595
APS1 0,680 0,663 0,916 0,591 0,595 0,621 0,507 0,441 0,467 0,517
APS2 0,829 0,659 0,629 0,921 0,602 0,558 0,643 0,441 0,439 0,459 0,486
APS3 0,589 0,574 0,894 0,531 0,571 0,562 0,515 0,402 0,506 0,449
BL1 0,612 0,614 0,597 0,889 0,462 0,544 0,366 0,352 0,461 0,526
BL2 0,750 0,446 0,412 0,560 0,869 0,431 0,472 0,372 0,353 0,517 0,391
BL3 0,525 0,468 0,476 0,839 0,333 0,403 0,199 0,290 0,320 0,402
EAC1 0,545 0,548 0,512 0,436 0,784 0,535 0,379 0,361 0,245 0,510
EAC2 0,680 0,491 0,399 0,537 0,409 0,849 0,469 0,451 0,330 0,372 0,403
EAC3 0,440 0,336 0,513 0,335 0,840 0,458 0,514 0,321 0,415 0,353
EAF1 0,617 0,572 0,604 0,505 0,536 0,884 0,469 0,490 0,359 0,570
EAF2 0,602 0,559 0,613 0,497 0,548 0,907 0,496 0,418 0,390 0,522
0,754
EAF3 0,577 0,580 0,566 0,433 0,477 0,863 0,416 0,399 0,371 0,522
EAF4 0,592 0,478 0,540 0,481 0,488 0,816 0,388 0,401 0,353 0,490
ECP1 0,403 0,408 0,448 0,302 0,432 0,425 0,876 0,436 0,430 0,381
ECP2 0,721 0,419 0,364 0,472 0,298 0,435 0,426 0,862 0,380 0,423 0,288
ECP3 0,380 0,366 0,443 0,334 0,518 0,451 0,808 0,447 0,442 0,376
PP1 0,308 0,232 0,336 0,279 0,290 0,304 0,366 0,754 0,368 0,296
PP2 0,364 0,330 0,382 0,307 0,329 0,416 0,390 0,804 0,377 0,371
0,588
PP3 0,381 0,340 0,333 0,307 0,305 0,374 0,392 0,734 0,312 0,370
PP5 0,349 0,342 0,387 0,285 0,327 0,406 0,373 0,775 0,272 0,435
PQ1 0,387 0,282 0,456 0,453 0,332 0,333 0,394 0,316 0,846 0,295
PQ2 0,386 0,337 0,464 0,409 0,324 0,309 0,411 0,380 0,832 0,282
PQ3 0,682 0,385 0,318 0,393 0,447 0,330 0,371 0,412 0,339 0,845 0,317
PQ4 0,375 0,318 0,447 0,449 0,360 0,398 0,452 0,375 0,859 0,302
PQ5 0,302 0,250 0,401 0,310 0,390 0,338 0,439 0,380 0,743 0,310
(Source: Researcher’s Smart-PLS Output)

Universitas Indonesia
 85

Table 4.18 Convergent Validity and Discriminant Validity Result with

Repeated Indicator Approach (Continued)
Item AVE AAF ACC APS BL EAC EAF ECP PP PQ SEB
SEB1 0,550 0,536 0,453 0,484 0,456 0,505 0,327 0,410 0,302 0,886
SEB2 0,561 0,579 0,481 0,465 0,487 0,556 0,380 0,455 0,356 0,897
SEB3 0,799 0,553 0,561 0,481 0,457 0,453 0,536 0,345 0,463 0,329 0,927
SEB4 0,560 0,586 0,467 0,459 0,444 0,571 0,339 0,428 0,305 0,919
SEB5 0,584 0,575 0,496 0,431 0,439 0,536 0,440 0,402 0,330 0,840
(Source: Researcher’s Smart-PLS Output)
From the analysis conducted using repeated indicator
approach in Smart-PLS for convergent validity and discriminant
validity on table 4.18, it is found that all the variables and lower-
order components within the second-order constructs (consumer
brand engagement and brand attachment) surpass the minimum
value of Average Variance Extracted (AVE) that should be ≥ 0,50,
Connection dimension of brand attachment variable acquire the
highest value for 0,882, Passion dimension of Brand Attachment
(0,829), Self-expressive brand variable (0,799). Affection
dimension of consumer brand engagement (0,754), Brand loyalty
(0,750), affection dimension of brand attachment (0,745) and lastly
cognitive processing dimension of consumer brand engagement
(0,721). In addition to that, perceived quality (0,682), activation
dimension of consumer brand engagement (0,680). Lastly,
perceived personalization (0,588) acquire AVE value that is not as
high as the other variable or dimension but are still above the
minimum value of 0,50 in order to be considered valid.

Aside from the evaluation of AVE, the outer loading value

from all variables and dimensions above also met the cut off value
≥ 0,70, thereby the corresponding items within the variables and
dimensions have establish convergent validity which means that all
the instruments used within the variables or dimensions are correct
on measuring the concept and have established a high correlation
with the concept being measured.

Universitas Indonesia
 86

Table 4.19 Convergent Validity and Discriminant Validity Result with Two-
Stage Approach

Item AVE BA BL CBE PP PQ SEB

AAF Item 0,917 0,613 0,706 0,459 0,445 0,629
ACC Item 0,814 0,903 0,582 0,641 0,409 0,365 0,636
APS Item 0,886 0,633 0,735 0,470 0,523 0,533
BL1 0,673 0,889 0,552 0,352 0,461 0,526
BL2 0,750 0,525 0,869 0,511 0,353 0,517 0,391
BL3 0,543 0,839 0,378 0,290 0,320 0,402
EAC Item 0,646 0,476 0,846 0,409 0,419 0,510
EAF Item 0,699 0,735 0,552 0,855 0,492 0,424 0,606
ECP Item 0,538 0,367 0,806 0,496 0,509 0,411
PP1 0,325 0,279 0,381 0,755 0,368 0,296
PP2 0,398 0,307 0,454 0,804 0,377 0,371
0,588
PP3 0,390 0,307 0,427 0,733 0,312 0,370
PP5 0,399 0,285 0,443 0,775 0,272 0,435
PQ1 0,418 0,453 0,421 0,316 0,846 0,295
PQ2 0,440 0,409 0,413 0,380 0,832 0,282
PQ3 0,682 0,406 0,447 0,443 0,339 0,845 0,317
PQ4 0,423 0,449 0,481 0,375 0,859 0,302
PQ5 0,354 0,310 0,462 0,380 0,744 0,310
SEB1 0,568 0,484 0,518 0,410 0,302 0,886
SEB2 0,598 0,465 0,572 0,455 0,356 0,897
SEB3 0,799 0,588 0,457 0,537 0,463 0,329 0,926
SEB4 0,595 0,459 0,547 0,428 0,305 0,919
SEB5 0,611 0,431 0,567 0,402 0,330 0,840
(Source: Researcher’s Smart-PLS Output)

From the analysis conducted with Smart-PLS using the two-

stage approach on table 4.19, it is also found that all the variables
surpass the minimum value of Average Variance Extracted (AVE)
that should be ≥ 0,50. Brand attachment variable acquire with 0,814
AVE, self-expressive brand (0,799). Furthermore, Brand loyalty
(0,750), Consumer brand engagement (0,699) and perceived quality
(0,682). Lastly perceived personalization acquires AVE value that is
not as high as the other variable but are still above the minimum
value of 0,50 in order to be considered valid.

Universitas Indonesia
 87

Aside from the evaluation of AVE, the outer loading value

from all items that represent its variables, including the dimensions
that is now treated as indicators for the second-order variable (CBE
and BA) exceed cut off value of 0,70 thereby the corresponding
items within the variables have establish convergent validity in
which means that all the instruments used within the variables are
correct on measuring the concept and have established a high
correlation with the concept being measured.

Continuing the above paragraph, the researcher also

evaluates discriminant validity by seeing the value of outer loading
between one variable and its items where it should establish a value
higher than other variable items in order to be valid (cross loading
requirement). From the discriminant validity evaluation through
cross-loading on above table (table 4.19 and 4.20), it is found that
all items (indicator) met the requirement since all the loading value
towards each of its measured construct or dimensions acquire the
highest value when compared to loadings toward other variable.
Therefore, the items used are correct just to measure the intended
latent constructs. From this point on, since there are no validity and
reliability issues within the lower order components (variables and
dimensions from second order constructs), the researchers will only
display the result from the two-stage approach (Ringle et al., 2012).

Table 4.20 Fornell-Larcker Criterion for Discriminant Validity Evaluation

Variable BA BL CBE PP PQ SEB

BA 0,902
BL 0,676 0,866
CBE 0,771 0,562 0,836
PP 0,495 0,385 0,558 0,767
PQ 0,495 0,504 0,536 0,432 0,826
SEB 0,663 0,514 0,615 0,483 0,364 0,894
(Source: Researcher’s Smart-PLS Output)

In addition to that, the assessment of discriminant validity

can also be seen by using Fornell-Larcker criterion on table 4.20

Universitas Indonesia
 88

above. From the result presented on above table, all variables

acquire the largest variances with its own constructs, therefore it can
be considered that the construct is valid. Within this research, all
variables have met the requirement to be considered valid from the
evaluation of discriminant validity through Fornell-Larcker
criterion, it is also found that brand attachment variable acquire the
highest Fornell-Larcker criterion value of 0,902.

Table 4.21 Heterotrait-Monotrait Ratio (HTMT) Evaluation

Variable BA BL CBE PP PQ
BA
BL 0,778
CBE 0,916 0,679
PP 0,596 0,478 0,715
PQ 0,557 0,579 0,649 0,529
SEB 0,728 0,574 0,709 0,566 0,400
(Source: Researcher’s Smart-PLS Output)

Aside from the evaluation of cross loading, Fornell-Larcker

criterion for discriminant validity, HTMT indices evaluation is also
performed. From table 4.21 above, it is found that there is no HTMT
value within each correlation on the model that reach one (1) and all
the HTMT indices of all variable acquire the value of less than one.
Therefore, it can be concluded that the model has a good validity.

4.5 Structural Model Analysis

4.5.1 Collinearity Evaluation

The evaluation of collinearity within the research model is done

toward the structural model or inner model by seeing the value of Inner VIF
(Inner Variance Inflation Factors) Value. From the evaluation of inner VIF
value, it is found that all the relationship between variables met the
requirement to be considered not having collinearity issue where all the
inner VIF value from each variable is below 5,0. Result can be seen on table
4.22 below.

Universitas Indonesia
 89

Table 4.22 Collinearity Evaluation with Inner VIF Value

Variable BA BL CBE PP PQ SEB

BA 2,521 2,463
BL
CBE 2,673 2,463
PP 1,305 1,305
PQ 1,437
SEB 1,305 1,305
(Source: Researcher’s Smart-PLS Output)

4.5.2 Coefficient of Determination (R2) Evaluation

Table 4.23 Evaluation on Coefficient of Determination (R2)

Endogenous Variable R Square R Square Adjusted

BA 0,480 0,477
BL 0,495 0,491
CBE 0,466 0,464
PQ 0,304 0,300
(Source: Researcher’s Smart-PLS Output)

After the evaluation of collinearity, the researcher also evaluates the

structural model through the assessment of coefficient of determination or
R2 shown on table 4.23 above. Coefficient of determination (R2) shows the
prediction accuracy of a construct within the structural model. The use of
PLS-SEM technique on this research is implemented in order maximize the
value of R2 of the latent endogenous variable on the path model. From table
4.24 above, it can be seen and described that the R2 value of brand
attachment shows predictive accuracy of 0,480, brand loyalty by 0,495, then
consumer brand engagement by 0,466 whilst perceived quality with 0,304
R2 value. The value of R2 adjusted are lower than R2 value since R2 adjusted
considers the contribution from exogenous latent variables that is
insignificant within the research model.

Universitas Indonesia
 90

4.5.3 Predictive Relevance (Q2) Evaluation

Table 4.24 Evaluation of Predictive Relevance with Q2 Cross-Validated

Redundancy

Variable SSO SSE Q² (=1-SSE/SSO)

BA 1.158,000 710,764 0,386
BL 1.158,000 742,717 0,359
CBE 1.158,000 793,050 0,315
PP 1544,000 1.544,000
PQ 1.930,000 1.537,582 0,203
SEB 1.930,000 1.930,000
(Source: Researcher’s Smart-PLS Output)

Furthermore, the researcher also decides to analyse the Q2 predictive

relevance value as presented on table 4.24 above. Q2 value is acquired by
using blindfolding procedure through implementing the value of specified
omission distance D. Blindfolding procedure is a technique that uses a
resampling procedure for samples that miss the data value on every “D”
point within the endogenous constructs indicator and predict its parameter
with the remaining point of data. (Chin, 1998; Henseler et al., 2009;
Tenenhaus et al., 2005). The data point that is gone is considered as missing
value when running the PLS-SEM algorithm. Then, the result of the
estimation through blindfolding is used to predict the missing data point.
The differences between the real data point and one that is predicted then
will be used as input to measure Q2 value. The calculation of Q2 on this
research is done by using cross-validated redundancy approach as advised
by Hair et al. (2017), where the D point value that is recommended in Smart-
PLS is within the range of five up to 12 with the default value of 7 (Hair et
al., 2017). The selections of the D distance value also follow a procedure in
which the requirement is that the amount of data when divided with D value
does not produce integer value (Hair et al., 2017). From the table above, it
can be seen that the value of Q2 on all endogenous variable acquire a value
of more than zero (0) with the highest value of brand attachment (0,386),
followed by brand loyalty (0,359), then consumer brand engagement
(0,315), and lastly perceived quality (0,203). The result then indicates that

Universitas Indonesia
 91

there is a predictive relevance within the model and toward the latent
endogenous variables.

4.5.4 f 2 Effect Sizes, q2 Effect Sizes, & SRMR Evaluation

Table 4.25 Evaluation of f2 Effect Sizes

Variable PP SEB CBE BA PQ BL

PP 0,166 0,076 0,003 0,000
SEB 0,291 0,450 -0.004 0,000
CBE 0,085 0,000
BA 0,023 0,241
PQ 0,068
BL
(Source: Researcher’s Smart-PLS Output)

Furthermore, the researcher also perform f2 effect size as presented

on table 4.25 above, the result of f2 effect sizes show the contribution
amount of exogenous construct towards the R2 value of endogenous
construct namely consumer brand engagement (CBE), brand attachment
(BA), perceived quality (PQ) and brand loyalty (BL). The contribution of f2
effect size toward R2 of consumer brand engagement variable shown the
highest f2 effect size from self-expressive brand (0,291) then followed by
perceived personalization (0,166). Then for the R2 contribution of brand
attachment variable the highest contribution as shown by its f2 value is from
self-expressive brand (0,450), followed by perceived personalization
(0,076). For the R2 contribution toward perceived quality variable, the
highest contribution came from consumer brand engagement variable with
an f2 value of 0,085 followed by brand attachment variable with f2 value of
0,023. Lastly, the highest contributor towards brand loyalty R2 is acquired
from brand attachment variable f2 effect size (0,241) and followed by
perceived quality variable (0,068).

As for the analysis of q2 effect sizes it is not shown within the output
calculation of Smart-PLS, therefore the researcher needs to calculate the
value manually. Similar with the relationship between f2 effect sizes with
the value of R2, q2 effect sizes is a measure to assess the relative predictive

Universitas Indonesia
 92

relevance of a predictor construct toward an endogenous construct through

the manual computation with formula below:

𝑄2 included − 𝑄2 excluded
𝑞2 =
1 − 𝑄2 included

Source: (Hair et al., 2017)

For computing q2 effect sizes using above formula from a construct (e.g.
construct A) toward an endogenous reflective latent variable, the researcher
first has to acquire the value of Q2 through blindfolding method in Smart-
PLS by including construct X (𝑄2 included) and then also by removing
construct X to acquire the value of (𝑄2 excluded). The computation within
Smart-PLS blindfolding method should also use the same omission distance
D value. As the relative measure from Q2 predictive relevance, the value of
q2 effect sizes shows the strength of exogenous construct relevance to an
endogenous construct. Therefore, the table below shows the result of q2
effect size computation of each independent variable toward the dependent
variable.

Table 4.26 Evaluation of q2 Effect Sizes

Variable CBE BA PQ BL
PP 0,083 0,049 0,001 0,000
SEB 0,149 0,309 -0,004 0,000
CBE 0,048 0,000
BA 0,013 0,139
PQ 0,041
(Source: Researcher’s Smart-PLS Output)

From the result presented by table 4.26 above, the computation

shows that almost all independent variable acquire the value of more than
zero (0) in which shows the strength of predictive relevance where among
the highest is SEB toward BA (0,309), SEB toward CBE (0,149), BA
toward BL (0,139), PP toward CBE (0,083), and PP toward CBE (0,083).
As an additional information, the model used on this research also acquire
SRMR value of 0,083 in which indicates that the theory used on the research
model has good validity.

Universitas Indonesia
 93

4.5.5 Size and Significance of Path Coefficient (Direct Effect)

Evaluation

After the evaluation of both measurement model (outer model) and

structural model (inner model), researcher conduct the significance test on
each variable relationship represented on the model. The significance test is
done through bootstrapping method in Smart-PLS application with
subsamples of 5,000 as recommended by Hair et al., (2017). The
bootstrapping method also allow the significance test with the level of
significance of 0,05 with one-tailed type since the hypothesis proposed by
the researcher are hypotheses with a presumed positive relationship. In order
to be considered significant, the value of T-Statistics from the effect of
independent variable toward its dependent variable must be ≥ 1,645 to be
considered to have a significant positive influence and ≥ - 1,645 to be
concluded that the result of significance test acquire a significant negative
influence. Therefore, the result of significance test as acquired from path
coefficient between variables or direct effect are as follow:

Table 4.27 Result of Direct Path Coefficient through Bootstrapping

Procedure

Sample
Path Coefficients Standard Deviation t-statistics
Mean
PP → CBE 0,340 0,342 0,048 7,109
PP → BA 0,228 0,230 0,047 4,935
SEB → CBE 0,450 0,450 0,051 8,831
SEB → BA 0,553 0,553 0,047 11,823
CBE → PQ 0,382 0,383 0,083 4,586
CBE → BL 0,016 0,018 0,062 0,249
BA → PQ 0,200 0,201 0,083 2,409
BA → BL 0,554 0,553 0,057 9,713
PQ → BL 0,221 0,221 0,051 4,310
(Source: Researcher’s Smart-PLS Output)

From the bootstrapping procedure results presented on table 4.27

above, the path coefficient indices present the effect relationship between
constructs in which among the effect relationship analysed through its path
coefficient, the effect of brand attachment toward brand loyalty are the

Universitas Indonesia
 94

highest (0,554) then followed by the effect of self-expressive brand toward

brand attachment (0,553), self-expressive brand toward consumer brand
engagement (0,450), consumer brand engagement toward perceived quality
(0,382), perceived personalization toward consumer brand engagement
(0,340), perceived personalization toward brand attachment (0,228),
perceived quality toward brand loyalty (0,221), and lastly the relationship
effect of brand attachment toward perceived quality (0,200). Within this
analysis there is also a path coefficient that does not meet the optimal value
to be considered significant in which is the effect of consumer brand
engagement toward brand loyalty with path coefficient of 0,016 (p = 0,402
> 0,05) that will further be discussed on its significance’s indices.

Regarding the result significance, there are eight (8) significant

effect and one (1) insignificant effect found through the bootstrapping
method. From the analysis it is found that the effect of perceived
personalization and self-expressive brand is significant toward consumer
brand engagement and the effect of perceived personalization and self-
expressive brand is significant toward brand attachment. Furthermore, the
effect of consumer brand engagement and brand attachment is also found
significant to influence perceived quality. Lastly, for brand loyalty, it is
found that the effect of brand attachment and perceived quality is significant
toward brand loyalty while the effect of consumer brand engagement toward
brand loyalty does not influence brand loyalty significantly.

Universitas Indonesia
 95

4.6 Hypotheses Testing Result

Table 4.28 Summary of Hypotheses Testing Result

t-
Hypotheses Research Reference Journal
Hypotheses statistics
Statement Result Result
value
Perceived
personalizati 7,109
on
Hypothesis
positively Hypothesis supported
H1 supported by
influences by data
data
consumer
brand
engagement
Perceived
personalizati
on Hypothesis
Hypothesis supported
H2 positively 4,935 supported by
by data
influences data
brand
attachment
Self-
expressive
brand
Hypothesis
positively Hypothesis supported
H3 8,831 supported by
influences by data
data
consumer
brand
engagement
Self-
expressive
brand Hypothesis
Hypothesis supported
H4 positively 11,823 supported by
by data
influences data
brand
attachment
Consumer
brand
engagement Hypothesis
Hypothesis supported
H5 positively 4,586 supported by
by data
influence data
perceived
quality

Universitas Indonesia
 96

Table 4.28 Summary of Hypothesis Testing Result

(Continued)
t-
Hypotheses Research Reference Journal
Hypotheses statistics
Statement Result Result
value
Consumer
brand
engagement Hypotheses
Hypothesis supported
H6 positively 0,249 not supported
by data
influence by data
brand
loyalty
Brand
attachment
Hypotheses
positively Hypothesis supported
H7 2,409 supported by
influence by data
data
perceived
quality
Brand
attachment
Hypothesis
positively Hypothesis supported
H8 9,713 supported by
influence by data
data
brand
loyalty
Perceived
quality
Hypothesis
positively Hypothesis supported
H9 4,310 supported by
influences by data
data
brand
loyalty
(Source: Researcher’s Output)

H1: Perceived personalization positively influences consumer brand

engagement
Based on the result of hypotheses testing, the relationship proposed in
hypotheses H1 that perceived personalization positively influence consumer brand
engagement are supported by the data. The positive influence from perceived
personalization toward consumer brand engagement is significant shown by its t-
statistics value of 7,109 or the result of hypotheses testing acquire a critical value
that is above of 1,645. The result from this hypotheses testing is the same with
previous research held by Shanahan et al. (2019). That research found that there is
a positive effect from perceived personalization toward consumer brand

Universitas Indonesia
 97

engagement within the context of personalised advertisement of a fashion brand on

Facebook. In addition to that, Dehghani and Turner (2015) on their studies on
Facebook advertisement suggest that advertisement personalised advertisement
promote opportunities to efficiently target audience while also promoting
interactions with the brand through its ads. In similar vein, Instagram advertisement
also work in a similar way which according to an article in Hootsuite written by
Newberry (2020), the rich options of format for the advertisement allows brand to
implement call to action button in which can promote interactions for users to view,
like, comment, and share the advertisement with their friends.

Continuing the discussion above, in line with the result of H1 hypotheses

testing on this research, (Blasco-Arcas et al., 2016) also found that an effective
personalized advertisement can enhance consumer engagement with a brand as it
foster the sense of relationship within the social media environment. The results
from these previous research can also be explained from the effect toward consumer
brand engagement, where personalised advertisement on social media can gave the
consumer a voice, in which enable the development of two-way communication
between the consumer and the brand (Hollebeek et al., 2014).

H2: Perceived personalization positively influences Brand Attachment

Based on the Smart-PLS structural model output analysis on this research,
the relationship between perceived personalization and brand attachment that is
hypothesized by H2 acquire t-statistic value of 4,935 or acquired a critical value
that is above 1,645. Therefore, it can be concluded that hypotheses H2 is accepted
and are supported by the data acquired for the context being investigated. The result
of hypotheses testing is also consistent with the previous research that is conducted
by Shanahan et al. (2019). That research result implies there is a positive influence
from perceived personalised message in advertisement (perceived personalization)
toward consumer’s emotional attachment to a brand (brand attachment). Research
by Shanahan et al. (2019) discussed that in the context of personalised
advertisement on Facebook, marketers can collect consumer’s information that
represents their attachment preferences in which the personalised advertisement can
be customized to match the consumer’s liking based on that preferences. This way

Universitas Indonesia
 98

the advertisement on social media can be used to highlight certain features that
match the preferences of consumers especially those who have prior affinity toward
the brands to grow emotionally along with the brand.

Continuing the discussion, previous research held by Zhang (2015)

highlight certain points that is similar to the finding of this hypotheses test result,
in which a brand or organization can utilize personalized message that is in
accordance to the preference of its consumers to foster the brand relationship with
its consumer. The personalized message within the advertisement then can also be
tailored and positioned in social media so that the personalised advertisement can
make the brand remain within the consumer’s mind as the consumer’s come across
the advertisement. (Whan Park et al., 2010).

H3: Self-Expressive Brand positively influences Consumer brand

Engagement
Based on the Smart-PLS structural model output analysis on this research,
the relationship between perceived personalization and brand attachment that is
hypothesized by H3 acquire t-statistic value of 8,831 in which the critical value
obtained is above 1,645. Therefore, it can be concluded that hypotheses H3 is
accepted and are supported by the data. The result of hypotheses testing is also the
same with previous research that is conducted by Algharabat et al. (2020). That
research found there is direct positive effect of self-expressive brand on consumer
brand engagement in the context of a brand’s social media group page. The research
found that consumers can express their personal values and self-concepts within the
social media through the available platform, that is the brand social media group.
Regarding the result on the context of personalised advertisement on Instagram that
is carried on this research, the personalised advertisement on Instagram works in
similar way where consumers can interact and express themselves by giving
comments, likes, or share the personalised advertisement as it is formed by
Instagram to mimic an organic photo or video posts (Instagram, 2020).

Such activities carried on social media platform were also discussed on

previous research. Where self-expression activities mentioned above toward a

Universitas Indonesia
 99

firm’s generated contents can stimulate further interaction such as discussion,

where consumers can further reflect their inner states that enhance their engagement
(Algharabat et al., 2020). Leckie et al. (2016) also found similar findings in which
a brand attributes can encourage the consumers to express themselves as they find
a congruent attribute of the brand to their self-concept in which stimulates them to
engage with the brands. Therefore, concluding the result of hypotheses H3, when a
personalized advertisement from a brand carried on Instagram are congruent and
reflect the consumer inner states and self, consumers are more likely to conduct a
positive cognitive, emotional, behavioural interactions with the corresponding
brand on Instagram.

H4: Self-Expressive Brand positively influences Brand Attachment

Based on the Smart-PLS structural model output analysis on this research,
the relationship between perceived personalization and brand attachment that is
hypothesized by H4 acquire t-statistic critical value of 11,823 that is above 1,645,
Therefore, it can be concluded that hypotheses H4 is accepted and are supported by
the data. The result of this hypotheses testing is the same with the previous research
findings by Lee and Workman (2015). That research found self-expressive brands
lead to a stronger brand attachment between the customer and the brand. Lee and
Workman (2015) implies that when a brand project a congruence signal that
connects with an individual inner states and self, a brand can remind that individual
about something significant within the individual’s life where he or she may find
attached emotionally to. Within the similar context with Lee and Workman (2015)
research on fashion brand, their research also explains that clothing brand are able
to enhance consumer’s his or her inner self not just by using the brand, but also
through advertising such as product promotions in which support the consumer’s
path to acquire the products of the brand while also remain connected with the
brand. In addition to that, Represented by the brand personalised advertisement
that is considered as self-expressive by the consumer, it can be inferred that when
a brand that is valued as socially visible, and reflective toward the individual’s self
that indicates strong consumer’s self-brand relationship can lead toward stronger
attachment to the brand (Dwayne Ball & Tasaki, 1992).

Universitas Indonesia
 100

Furthermore, Stokburger-Sauer et al (2012) also suggest that brands can

communicate and inform desirable identities to distinguish their brand in a market
with a similar product. Within the context of social media, previous research by
Tran et al. (2020) also implies that when a brand advertise on social media such as
Facebook are recognizable to the consumer and the communicated brand values are
aligned within the consumer’s self-concept, it can promote positive feelings in
which support the creation of attachments by the consumer’s toward the brand. In
similar vein, the result derived from this hypothesis also give new insights in which
that an advertisement on Instagram can be designed and formed to communicate
the brand’s desirable personalities that reflect the consumer’s inner self to foster the
consumer’s brand attachment.

H5: Consumer brand engagement positively influences perceived quality

Based on the Smart-PLS structural model output analysis on this research,
the relationship between perceived personalization and brand attachment that is
hypothesized by H5 acquire t-statistic value of 4,586 in which that the critical value
is above 1,645. Therefore, it can be concluded that hypotheses H5 is accepted and
are supported by the data. The result of this hypotheses testing is also consistent
with the previous research that is conducted by Shanahan et al. (2019). That
research found that when a customer is actively engaging with the corresponding
brand through its personalised advertisement on Facebook (consumer brand
engagement) it will positively impact how the customer perceive the superiority of
the brand over other brand and its quality perception (perceived quality).

The research by Shanahan et al. (2019) also highlight that consumer brand
engagement on social media (Facebook) allows consumers to voice their perceived
quality of a product or service from a brand. In line with the result of hypotheses
H5, and findings by Shanahan et al. (2019) the nature of Instagram as social media
platform studied on this research comply with the result in which that the brand’s
personalised advertisement on Instagram can stimulate and also acts as medium for
engagement activities. The personalised advertisement can provide a way where
consumers can learn more (cognitive processing and activation) about the brand,
interact with the brand through commenting, liking (affection), or sharing the

Universitas Indonesia
 101

brand’s personalised advertisement content thus acquiring more information about

the brand’s quality. In line with the hypotheses result, previous research by
Schivinski & Dabrowski (2015) in the context of brand communication on
Facebook also highlight that there is a direct positive relationship between social
media brand communication and perceived quality in which that social media brand
communication like shared brand’s contents are an important source of information
for other consumer regarding the brand perceived quality.

H6: Consumer brand engagement positively influences brand loyalty

Based on the Smart-PLS structural model output analysis on this research,
the relationship between perceived personalization and brand attachment that is
hypothesized by H6 acquire t-statistic value of 0,249 or having a critical value that
is below 1,645. Therefore, it can be concluded that hypotheses H5 is rejected and
are not supported by the data since the positive influences suggested within the
hypotheses between consumer brand engagement to brand loyalty are found to be
insignificant. The result of hypotheses testing is different than the previous result
conducted by Shanahan et al. (2019).

The result of that research founds that consumer brand engagement is

positively related to brand loyalty of a brand being advertised on Facebook. There
are several plausible explanations that can be given regarding the result of
hypotheses H6 test. First, Previous research by Hollebeek (2011) explained that
the relationship between consumer brand engagement and brand loyalty at some
point can reach an optimal level, especially when the level of engagements between
the brands and the consumer above that goes above optimal points can be less
effective on generating enhanced loyalty for those of highly engaged consumers.
These findings can also explain the overall respondents’ answers that are relatively
similar and high in terms of mean value across dimensions of consumer brand
engagements in which can be inferred that respondents are those that exhibit a
relatively high level of engagements with the brands on Instagram.

Furthermore, with a focus toward the brand personalised advertisement as

the medium of the brand and consumer’s engagement on this research, such result
could also be indicated due to the brand’s personalised advertisement display that

Universitas Indonesia
 102

is subject to the advertisement optimization delivery and automatic placement

toward the particularly high performing audience (Newberry, 2020). Another factor
that may influence this result is also majority of the respondents that spend a
relatively lengthy time that is above one hour per day when using social media,
particularly Instagram and also majority of the respondents have been following the
brand from a range that is above six month period. From the consumer’s perspective
as Instagram users, these conditions may lead to a condition where consumers came
across a repetitive advertisement that they already see or interact with. In which in
this case can render the direct effect of consumer brand engagement toward brand
loyalty is insignificant, meaning that when the consumer’s positive cognitive,
emotional, and behavioural reactions on their interactions with a focal brand are
within its optimal point, the interactions between the consumer’s and the brand does
not significantly enhance the consumer’s level of commitment and also their
propensity to repurchase the brands.

H7: Brand attachment positively influences perceived quality

Based on the Smart-PLS structural model output analysis on this research,
the relationship between brand attachment and perceived quality that is
hypothesized by H7 acquire t-statistic value of 2,409 where it means that the critical
value is above 1,645. Therefore, it can be concluded that hypotheses H7 is accepted
and are supported by the data since brand attachment is found to have a significant
positive influence toward perceived quality. The result of hypotheses testing is also
consistent with the previous research that is conducted by Shanahan et al. (2019).
The result of that research found there is direct positive influence from brand
attachment toward the consumers perception on how superior the brand and its
quality perception (perceived quality) on the context of personalised advertisement
on Facebook. In accordance to the result of hypotheses H7 on this research,
consumers attachment toward a brand are based on their emotion (Whan Park et al.,
2010) in which that these emotionally led attachment help consumer define their
quality perception toward a certain brand based on their subjective definition
(Shanahan et al., 2019). Zeithaml (1988) also describe that perceived quality is
influenced by the consumer’s subjective judgement that is related to their unique
needs, product experiences, consumption situations of the consumers.

Universitas Indonesia
 103

H8: Brand attachment positively influences brand loyalty

Based on the Smart-PLS structural model output analysis on this research,
the relationship between brand attachment and brand loyalty that is hypothesized
by H8 acquire t-statistic value of 9,713 that is above the critical value of 1,645.
Therefore, it can be concluded that hypotheses H8 is accepted and are supported by
the data. The result of hypotheses testing is also consistent with the previous
research that is conducted by (Shanahan et al., 2019). That research implies that
there is a positive direct effect from an emotionally led attachment toward a brand
(brand attachment) to the degree of commitment of consumers to repurchase a
product from a brand (brand loyalty). In accordance with the result of hypotheses
H8 on this research, Lee and Workman (2015) also found similar findings in which
emotionally led attachment influence consumers to develop stronger commitment
and improve their propensity to repurchase the brand’s product in the future.

In regard to above findings, it is found that a positive relationship between

the consumer’s emotional attachment and their commitment to be loyal toward a
brand are rooted from the consumers positive emotional feelings and close
interpersonal relationships between the consumer and the brand. (Chaudhuri &
Holbrook, 2001). Similarly, a previous research also highlights that brand loyalty
is usually greater when the consumer’s experience a positive emotion or affect
toward the brand in which the depth of emotions and feelings of the consumer
toward the brand influence consumer purchase behaviour (Akçura et al., 2004;
Whan Park et al., 2010). In line with the results of hypotheses H8, previous research
also found that as consumers become more attached to a brand, it is most likely that
they continue to repurchase the corresponding brand in the future.(Franz‐Rudolf,
2006).

H9: Perceived quality positively influences brand loyalty

Based on the Smart-PLS structural model output analysis on this research,
the relationship between perceived quality and brand loyalty that is hypothesized
by H9 acquire t-statistic of 4,310 meaning that the critical value obtained is above
1,645. Therefore, it can be concluded that hypotheses H9 is accepted and are
supported by the data. The result of hypotheses testing is also consistent with the

Universitas Indonesia
 104

previous research that is conducted by Shanahan et al. (2019). The corresponding

research shows that there is a significant direct positive relationship between the
perceived quality of a brand to the consumer’s brand loyalty.

In accordance with the result of hypotheses H9 within this research,

Erdoǧmuş & Büdeyri-Turan (2012) on their research of ready-to-wear fashion
brand consumers, found that perceived quality is one important factor that can result
to a deep commitment that causes repetitive purchase toward a brand. Previous
studies by Oliver (2010) also confirms that when a brand has earned a high quality
perception, especially regarding their product superiority, it is believed that the
result will lead to generate a strong sense of brand-directed preference to the focal
brand where consumer’s will develop a stronger commitment and continue to
repurchase the brand through the pre-established quality preference toward the
brand. Based on the findings from H9 hypotheses testing, it can be inferred that
when consumers have a high-quality perception toward a brand it will positively
impact how they will uphold their deep commitment and repurchase the
corresponding brand product in the future

Universitas Indonesia
 CHAPTER V

CONCLUSION

5.1 Conclusion

This research adapts a previous research by Shanahan et al. (2019) that

investigate and analyse the effect of personalization toward consumer brand
engagement and brand attachment that predicts consequence of perceived quality
and brand loyalty of a brand that advertise on Facebook. In addition to that, this
research also conduct modification within the research model by adding one
predictor variable namely self-expressive brand (Algharabat et al., 2020; Leckie et
al., 2016; Lee & Workman, 2015) that depicts the extent of a brand on reflecting
consumer’s individual inner self and image that predicts outcome such as consumer
brand engagement and emotional attachment toward a brand in which is
investigated on the core reference journal and this research. Based on the result of
the acquired data and the researcher’s analysis, the findings are quite similar
although there is difference found when compared to the main reference paper by
Shanahan et al. (2019). Based on the researcher’s analysis, the difference found
within the result are caused by several different reasons starting from the difference
within the geographical domain that this research is carried on (Indonesia) that
causes different consumer behaviour from a different culture and social perspective
of the samples for this research. Therefore, referring to the stated research problem
and research purpose on chapter one, the researcher can answer the research
problem with several summary based on the findings from this research which are:

1) Social media perceived personalization positively influences consumer

brand engagement and brand attachment of a local footwear brand that
advertise on Instagram.
2) Self-expressive brand positively influences consumer brand engagement
and brand attachment of a local footwear brand that advertise on Instagram
3) Consumer brand engagement in the context of personalised advertisement
carried by a local footwear brand on Instagram is found to only positively

105
Universitas Indonesia
 106

influence the perception of quality of the local footwear brand (perceived

quality) but does not influence brand loyalty of the consumer from local
footwear brand that advertise on Instagram. Furthermore, emotionally led brand
attachment found to positively influence both consequence on perception of the
quality and loyalty toward the local footwear brand that advertise on Instagram.

4) Lastly, perceived quality is found to positively influence the consumer’s loyalty

toward the local footwear brand that advertise on Instagram.

Based on the result that are summarized on above paragraph, it can also be inferred
that the purpose of this research, which are to gain insights on the relationship
between perceived quality; self-expressive brand toward consumer brand
engagement; brand attachment, consumer brand engagement; brand attachment
toward perceived quality; brand loyalty and lastly perceived quality toward brand
loyalty of local footwear brand that advertise on Instagram has been achieved.
Therefore, the discussion regarding managerial implications for business
practitioners, recommendation for future research and the research limitations will
be discussed on several following sections.

5.2 Managerial Implication

Referring to the result of this research, the findings as described within the
previous chapters can be used as insights by managements of local footwear brand
within the footwear industries as reference to formulate a relevant strategy and
decisions. Brands within the local footwear industries that are using Instagram as
one of their digital marketing channels to advertise their product could refer to this
research findings to consider several of the findings in consumer behaviour in
regard to their advertisement that can be carried in personalised scheme on
Instagram. Therefore, several of managerial implication that can answer both
business and marketing problem of this research is as follows:

1) The result of this research shows that the utilization of personalised ads
scheme on Instagram to deliver advertisement is important to stimulate
consumer’s engagement toward the brand and positive emotional
attachment toward the brand which can lead to a positive perception of

Universitas Indonesia
 107

quality for the brands. In addition to that, emotionally led attachment toward
the brand also significantly impact the consumer’s loyalty toward the brand.
2) Continuing above point, This research found that perceived personalization
have a significant direct influence toward the consumer’s behavioural,
emotional, and cognitive states which makes them wanting to interact with
the brand as they come across the local footwear brand’s personalised
advertisement on Instagram and also foster their emotional attachment as
the personalised advertisement kept the brand “close” within the consumer’s
self and mind. In detail, based on the descriptive analysis it is found that
indicators from perceived personalization construct, PP1 produce the
highest mean in which that the consumer perceive the advertisement from
the local brand on Instagram is able to give them a purchase
recommendation that is best suited according to their individual preference.
Furthermore, when looking toward the outer loading result on chapter four
on perceived personalization indicators, the finding shows that PP2 and PP5
produce the highest score. Based on these findings, from the consumer’s
perspective, local footwear brand that advertise on Instagram might want to
also accommodate customized selection of products by enabling the
consumers to pre-order the products to their liking in which the promotion
and offerings can also be incorporated in Instagram ads by implementing
redirect button in the advertisement toward the website customization page
which are commonly found in foreign brands website, and can be adopted
by local brands. Furthermore, footwear brands can take advantages of
Instagram advertisement targeting options, specifically the custom audience
feature This way, local footwear brand can choose audience based on their
look alike characteristics or lookalike audience. Look alike audience is
advanced feature in Instagram ads in which allows brands further target the
advertisement audience to people that already have prior affinity toward the
brand such as similar people from their follower base and data.
3) The result of this research also finds that self-expressive brand in the context
of personalised advertisement on Instagram plays an important role for a
brand to stimulate their consumer engagement and on fostering the

Universitas Indonesia
 108

consumers emotional attachment toward the brand. Therefore, brands can

incorporate their brand image and personality that are congruent with the
consumer’s inner self through its advertising content. Through the outer
loading value of self-expressive brand, it is recommended that local
footwear brand which advertise on Instagram to develop an advertisement
contents that reflects the consumer’s personality (SEB2), able to projects
the consumer’s inner self on Instagram (SEB3) and can help express the
consumer’s real self on Instagram (SEB4) through leveraging the capability
to customize Instagram advertisement content. The advertisement can be
presented with variety of format options such as photo, video, and IGTV to
provide a more relatable illustration on the contents, for example, a footwear
brands can illustrate an active, outgoing, outdoorsy image whilst
incorporating a strong and catchy copywriting that captures the current
fashion trend within the advertisement which reflect the brand’s image
attributes that the consumer’s find within a match with their inner self and
encourage self-expression in social media, this way marketers of footwear
brand can display an attractive visual story telling. In addition to that, the
listed footwear brands can also incorporate “re-gram” contents from their
consumers. Within Instagram, many users of Instagram like to post their
experience when using a product from a brand on their Instagram posts or
stories. Thereby, local footwear brands can incorporate re-gram activity by
asking permission from the consumers to uses their posts to be aired as an
Instagram advertisement or “re-grammed”. Since this is somewhat a “user-
generated content”, the contents are seen to be more natural and subtle
which possessed certain degrees of personal values from the actual user that
initially posts their experience on their personal social media profile. By
leveraging a more natural and subtle actual users’ content, it might capture
and relate to others Instagram users with similar self-concept when it is
aired and turned as Instagram ads towards people that are within similar
characteristics. Furthermore, as the advertisement on Instagram is already
integrated with interactions options such as like, comment, and share within
the format, the corresponding advertisement can generate interactions in

Universitas Indonesia
 109

which are based on the consumer’s perception that the advertisement

contents that they see encourage them to express themselves in social media.
4) This research shows consumer brand engagement is found to significantly
influence the quality perception of a local footwear brand. These findings
indicate that, as consumers are actively engaging with the brand through its
personalised advertisement on Instagram, it will positively impact how
consumers perceive the brand’s superiority and quality through information
they acquire from the brand’s personalised advertisement. Based on the
outer loading value from dimensions of consumer brand engagement
constructs, several recommendation can be made such as, footwear brands
or brands in general that advertise on Instagram to conduct research based
on their personalised advertisement performance that captures the most
“likes” to understand what type of formats, and contents that the consumer’s
loved to see and more likely to interact with in which brands can develop
an advertisement contents that implicitly contains their brand’s product
superiority that are less conspicuous and subtle for the consumers to interact
with and thus promote an increase on like and interactions that broaden the
spread of information for the brands quality (affection). Second it is
recommended that brands display advertisement that are interactive and
informative by also considering the implementation on variety of call-to-
action buttons within the advertisement such as learn more, watch more, and
contact us to promote more interactions that result on better information
gained by the consumers (activation). More importantly, it is recommended
for brands to pay attention on their advertisement timeliness or schedules
customization based on the targeted user’s frequency of use on Instagram
when airing the advertisement. The reason is to optimize engagement
outcomes by leveraging it on the optimal user’s online time that result on
better acquired information as the consumers can learn and engage with the
brand’s advertisement that contains the latest information (cognitive
processing). All and all these objectives can also be focused by choosing the
engagement objectives within the initial marketing objective specification

Universitas Indonesia
 110

of Instagram ads to focus the advertisement display on generating more

comments, likes, shares, responses, and broaden the information reach.
5) This research find that brand attachment has significant direct influence
toward the quality perception of local footwear brands and especially
towards consumer’s loyalty toward the brand. These findings indicate that
as consumers are more emotionally close and attached with a brand, their
feelings of affection, connection, and passion toward the brand will
positively impact their perception toward the brand’s quality and
particularly result in deeper commitment toward the brand. Based on the
outer loadings, it is shown that affection dimension and connection
dimension of brand attachment dimension acquired the highest outer
loading value. Based on this, local footwear brand that advertise on
Instagram can incorporate visually inspiring promotional contents that
particularly drive emotions by showing what people what they can achieve
together with the brand by associating the goals with the brand’s product
that can facilitate the accomplishment of that objectives. Therefore, upon
configuring the advertisement local footwear brand can possibly implement
the contents by hiring influencer as a model that are relatable to people’s
goal in which can leave emotional impressions such as passion, personal
connection, and affection which is easier to latch onto the associated
products that the brands try to promote.
6) This research also found that perceived quality significantly influences the
consumer’s deep commitment and willingness to repurchase local footwear
brand. The result indicates that, when the quality of a brand is perceived to
be high in quality by its consumers it can positively impact the consumer’s
commitment toward the brand and their propensity to conduct repurchase
activities through the already pre-established quality perception of the brand
that match their subjective preference. Referring to the outer loading values
that are similar across the construct of perceived quality, several
recommendation can be made for local footwear brands that advertise on
Instagram, such as to consistently projects their brand product’s quality
attributes such as symbolic design, functional product features, quality

Universitas Indonesia
 111

material used which resonates with consumer preference and sets the brand
apart from alternatives which signals uniqueness and competitive advantage
compared to other brands and particularly foreign brands.

5.3 Research Limitations

Regarding the limitation of this research, upon the planning and execution
of this research there are several limitations that the researcher come across. Several
of those limitations are as follows:

1. Respondents within this research are still dominated with several group
from the populations, especially gender and age where it will cause
challenges on generalizing the results when compared to a more evenly
distributed sample group. Even so, referring to the result of this research on
chapter four, this research captures the real phenomenon regarding the
consumer base on four of the brands that targeted certain group within the
market.
2. Second, there are limitations on theory for the variable used on this research
namely perceived personalization that focuses on the context of
personalised advertisement on social media and theory that highlight the
effect of personalization perception toward consumer brand engagement
and brand attachments. The researcher finds difficulties to acquire theories
that highlight the relationship between these variables considering the
investigated phenomenon of social media personalization are still limited in
number of previous researches that investigate the relationship between the
variables.
3. Due to the nature of Instagram that is an image sharing social networking
site that allows the creation of advertisement with similar format of organic
posts such as feed post, stories, carousel, and IGTV, there are possibilities
that respondents refer to Instagram organic posts like photo or video on
answering the questionnaire instead of the real advertisement with a
“sponsored” tag within the contents. Even so, the researcher already tried to
minimize the misinterpretation for the respondents by first providing
definitions of Instagram advertisement within the questionnaire.

Universitas Indonesia
 112

4. Since this research uses cross-sectional data for analysis that are collected
within a specific period, the result of this research should also be treated
carefully in terms of generalization. Especially regarding the current
situation of global pandemic due to Covid-19 most of the respondents
seemingly spent longer using their gadgets and social media due to the
“physical distancing policy” in which the researcher think may influence
the result of this research.

5.4 Recommendation for Future Research

Referring to the limitations of this research, the researcher hopefully can

suggest several points of reference for the development of future research that are
related to the context of this research which is personalised advertisement on social
media. Therefore, several of the recommendation for future research are as follow:

1. Adding a multi-group analysis based on different classifications of groups

such as different brand on different industries, length of time spent on social
media, and others.
2. Considering this research investigate the impacts of social media
personalisation and self-expressive brand toward consumer brand
engagement and brand attachment with outcomes of perceived quality and
brand loyalty on Instagram. The researcher recommend that future research
can be carried out to further investigate the context as represented by the
research model in different context of industries and geographical domain
that may help with the generalization regarding consumer behaviour in
different culture and type of industries.
3. Research model modification and expanding the research model by adding
relevant variable that may predict consumer brand engagement and brand
attachment and influence perceived quality and brand loyalty of a brand that
are not yet implemented to investigate the predicted outcomes and final
consequences of this research.

Universitas Indonesia
 REFERENCES

Aaker, D. (1991). Managing brand equity: Capitalizing on the value of a brand

name. New York, NY: The Free Press.
Aaker, D. (2009, September 6). Beyond functional benefits. Marketing News, p.
23.Aaker, D. (1991). Managing Brand Equity. Free Press: N. Y.

Akçura, M. T., Gönül, F. F., & Petrova, E. (2004). Consumer Learning and Brand
Valuation: An Application on Over-the-Counter Drugs. Marketing Science,
23(1), 156–169. https://doi.org/10.1287/mksc.1030.0028

Algharabat, R., Rana, N. P., Alalwan, A. A., Baabdullah, A., & Gupta, A. (2020).
Investigating the antecedents of customer brand engagement and consumer-
based brand equity in social media. Journal of Retailing and Consumer
Services, 53(March 2019), 101767.
https://doi.org/10.1016/j.jretconser.2019.01.016

Alkhateeb, F. M., Clauson, K. A., Khanfar, N. M., & Latif, D. A. (2008). Legal and
regulatory risk associated with Web 2.0 adoption by pharmaceutical
companies. Journal of Medical Marketing, 8(4), 311–318.

Alt, M., & Griggs, S. (1988). Can a brand be cheeky? Marketing Intelligence &
Planning.

Andriani, D. (2019, April 14). Geliat Sneakers Lokal Tembus Pasar Global:
Lifestyle. Retrieved February 28, 2020, from
https://lifestyle.bisnis.com/read/20190414/220/911616/geliat-sneakers-lokal-
tembus-pasar-global

APJII. (2017). Penetrasi & Profil Perilaku Pengguna Internet Indonesia. Apjii, 51.
www.apjii.or.id

APJII. (2019). Buletin APJII Edisi-40 2019. 6. https://apjii.or.id/survei

113
Universitas Indonesia
 114

Avnet, T., & Higgins, E. T. (2006). How Regulatory Fit Affects Value in Consumer
Choices and Opinions. Journal of Marketing Research, 43(1), 1–10.
https://doi.org/10.1509/jmkr.43.1.1

Ayer, A. J. (1955). What is communication. Studies in Communication, 11.

Babin, B. J., Darden, W. R., & Griffin, M. (1994). Work and/or Fun: Measuring
Hedonic and Utilitarian Shopping Value. Journal of Consumer Research,
20(4), 644–656. https://doi.org/10.1086/209376

Barnlund, D. C. (1963). Toward a meaning-centered philosophy of communication.

ETC: A Review of General Semantics, 454–469.

Becker, J. M., Klein, K., & Wetzels, M. (2012). Hierarchical Latent Variable
Models in PLS-SEM: Guidelines for Using Reflective-Formative Type
Models. Long Range Planning, 45(5–6), 359–394.
https://doi.org/10.1016/j.lrp.2012.10.001

Belk, R. W. (1988). Possessions and the Extended Self. Journal of Consumer

Research, 15(2), 139. https://doi.org/10.1086/209154

Berelson, B., & Steiner, G. A. (1964). Human behavior: An inventory of scientific

findings.

Berthon, P., Pitt, L., & Campbell, C. (2008). Ad lib: When customers create the ad.
California Management Review, 50(4), 6–30.

Berthon, P. R., Pitt, L. F., Plangger, K., & Shapiro, D. (2012). Marketing meets
Web 2.0, social media, and creative consumers: Implications for international
marketing strategy. Business Horizons, 55(3), 261–271.
https://doi.org/10.1016/j.bushor.2012.01.007

Blasco-Arcas, L., Hernandez-Ortega, B. I., & Jimenez-Martinez, J. (2016).

Engagement platforms: The role of emotions in fostering customer
engagement and brand image in interactive media. Journal of Service Theory
and Practice, 26(5), 559–589. https://doi.org/10.1108/JSTP-12-2014-0286

Universitas Indonesia
 115

Blau, P. M. (1964). Justice in social exchange. Sociological Inquiry, 34(2), 193–

206.

Boerman, S. C., Kruikemeier, S., & Zuiderveen Borgesius, F. J. (2017). Online

Behavioral Advertising: A Literature Review and Research Agenda.
Journal of Advertising, 46(3), 363–376.
https://doi.org/10.1080/00913367.2017.1339368

Brodie, R. J., Hollebeek, L. D., Jurić, B., & Ilić, A. (2011). Customer engagement:
Conceptual domain, fundamental propositions, and implications for
research. Journal of Service Research, 14(3), 252–271.
https://doi.org/10.1177/1094670511411703

Budden, C. B., Anthony, J. F., Budden, M. C., & Jones, M. A. (2007). Managing
the evolution of a revolution: Marketing implications of Internet media usage
among college students. College Teaching Methods & Styles Journal (CTMS),
3(3), 5–10.

Bullmore, J. (1984). The brand and its image re-visited. International Journal of
Advertising, 3(3), 235–238.
Carroll, B. A., & Ahuvia, A. C. (2006). Some antecedents and outcomes of brand
love. Marketing Letters, 17(2), 79–89. https://doi.org/10.1007/s11002-006-
4219-2

Caywood, C. L., Schultz, D. E., & Wang, P. (1991). Integrated marketing

communications: A survey of national consumer goods advertisers. AAAA.

Chaffey, D. (2008). E-marketing and Internet marketing definition. DaveChaffey.

Com: Your Guide to Digital Business.

Chaudhuri, A., & Holbrook, M. B. (2001). The chain of effects from brand trust
and brand affect to brand performance: The role of brand loyalty. Journal of
Marketing, 65(2), 81–93. https://doi.org/10.1509/jmkg.65.2.81.18255

Cheah, J. H., Ting, H., Ramayah, T., Memon, M. A., Cham, T. H., & Ciavolino, E.
(2019). A comparison of five reflective–formative estimation approaches:
reconsideration and recommendations for tourism research. In Quality and

Universitas Indonesia
 116

Quantity (Vol. 53, Issue 3). Springer Netherlands.

https://doi.org/10.1007/s11135-018-0821-7

CNN Indonesia. “Melacak Jejak Tren Sneakers.” Gaya Hidup, CNN Indonesia, 12
Nov. 2018, www.cnnindonesia.com/gaya-hidup/20181109191643-277-
345367/melacak-jejak-tren-sneakers.

Court, D., Elzinga, D., Mulder, S., & Vetvik, O. J. (2009). The consumer decision
journey. McKinsey Quarterly, 3(3), 96–107.

Cooper, G. 2015. How to Advertise on Instagram [Guide]. Accessed on 28th April

2020. Retrieved from https://www.socialvantage.com/advertising-on-
instagram-guide/

Dance, F. E. X. (1970). The “Concept” of Communication. Journal of

Communication, 20(2), 201–210. https://doi.org/10.1111/j.1460-
2466.1970.tb00877.x

Das, K., Gryseels, M., Sudhir, P., & Tan, K. T. (2016). Unlocking Indonesia’s
Digital Opportunity. McKinsey & Company, October, 1–28.
https://www.mckinsey.com/~/media/McKinsey/Locations/Asia/Indonesia/Ou
r Insights/Unlocking Indonesias digital
opportunity/Unlocking_Indonesias_digital_opportunity.ashx

de Chernatony, L. (1992). Creating powerful brands; the strategic route to success

in consumer industry. Oxford: Butterworth-Heinemann.

Dehghani, M., & Tumer, M. (2015). A research on effectiveness of Facebook

advertising on enhancing purchase intention of consumers. Computers in
Human Behavior, 49, 597–600.

DelVecchio, D., Henard, D. H., & Freling, T. H. (2006). The effect of sales
promotion on post-promotion brand preference: A meta-analysis. Journal of
Retailing, 82(3), 203–213.

Duckworth, G. (1991). Brands and the role of advertising. Understanding Brands.

London: Kogan Page, 59–81.

Universitas Indonesia
 117

Dwayne Ball, A., & Tasaki, L. H. (1992). The Role and Measurement of
Attachment in Consumer Behavior. Journal of Consumer Psychology, 1(2),
155–172. https://doi.org/10.1016/S1057-7408(08)80055-1

Economist Intelligence Unit. (2018). Preparing for disruption Technological

Readiness Ranking. Eiu, 21. http://pages.eiu.com/rs/753-RIQ-
438/images/Technological_readiness_report.pdf?mkt_tok=eyJpIjoiTWpBeU
5EVTNPRGMxT1RndyIsInQiOiJ4TURqODlRRVwvcGpVellKb2p4Vno1Q
W1JN1wvYkwzY1Z3dVV2U3dLQytWTkhqQmpEXC91ZlpHQlBxOTV5N
09LT3RHZlJGaHFhXC82MW5IbmJDRUQ0UEtEU2dxbjJ6eER5Q

Erdoǧmuş, I., & Büdeyri-Turan, I. (2012). The role of personality congruence,

perceived quality and prestige on ready-to-wear brand loyalty. Journal of
Fashion Marketing and Management, 16(4), 399–417.
https://doi.org/10.1108/13612021211265818

Escalas, J. E., & Bettman, J. R. (2005). Journal of Consumer Research Volume 32

issue 3 2005 [doi 10.1086%2F497549] Escalas, Jennifer Edson; Bettman,
James R. -- Self‐Construal, Reference Gro.pdf. 32(3), 378–389.
https://doi.org/10.1086/653494

Fetscherin, M., & Heinrich, D. (2015). Consumer brand relationships research: A

bibliometric citation meta-analysis. Journal of Business Research, 68(2),
380–390.

Fornell, C., & Bookstein, F. L. (1982). Two Structural Equation Models: LISREL
and PLS Applied to Consumer Exit-Voice Theory. Journal of Marketing
Research, 19(4), 440–452. https://doi.org/10.1177/002224378201900406

Fournier, S. (2013). Consumers and Their Brands Developing. Journal of

Consumer Research, 24(4), 343–353. https://doi.org/10.1086/209515

Franz‐Rudolf, E. (2006). Are brands forever? How brand knowledge and

relationships affect current and future purchases. Journal of Product &amp;
Brand Management, 15(2), 98–105.
https://doi.org/10.1108/10610420610658938

Universitas Indonesia
 118

Frazer, C. F. (1983). Creative strategy: A management perspective. Journal of

Advertising, 12(4), 36–41.

Gardner, B. B., & Levy, S. J. (1955). The product and the brand. Harvard Business
Review, 33(2), 33–39.

Gibbs, M., Meese, J., Arnold, M., Nansen, B., & Carter, M. (2015). #Funeral and
Instagram: death, social media, and platform vernacular. Information
Communication and Society, 18(3), 255–268.
https://doi.org/10.1080/1369118X.2014.987152

Graeff, T. R. (1996). Using promotional messages to manage the effects of brand

and self‐image on brand evaluations. Journal of Consumer Marketing.

Grove. (1998). from the SAGE Social Science Collections . All Rights. Hispanic
Journal of Behavioral Sciences, 9(2), 183–205.
https://doi.org/10.1177/07399863870092005

Hair Jr., J.F., Ringle, C., Sarstedt, M., (2011). PLS-SEM: indeed a silver bullet.
Journal of Marketing Theory & Practice 19 (2), 139–151.

Hair, J.F., Ringle, C.M. and Sarstedt, M. (2013). Partial least squares structural
equation modeling: rigorous applications, better results and higher
acceptance. Long Range 119 Planning, Vol. 46 Nos 1/2, pp. 1-12.

Hair, J.F., Hult, G.T.M., Ringle, C.M. and Sarstedt, M. (2014), A Primer on Partial
Least Squares Structural Equation Modeling, Sage, Thousand Oaks, CA.

Hair, J.F., Sarstedt, M., Pieper, T.M. and Ringle, C.M. (2012a), “The use of partial
least squares structural equation modeling in strategic management research:
a review of past practices and recommendations for future applications”,
Long Range Planning, Vol. 5 Nos 5/6, pp. 320-340.

Hair, J.F., Sarstedt, M., Ringle, C.M. and Mena, J.A. (2012b), “An assessment of
the use of partial least squares structural equation modeling in marketing
research”, Journal of the Academy of Marketing Science, Vol. 40 No. 3, pp.
414-433.

Universitas Indonesia
 119

Hair Jr, J., Hult, G. T., Ringle, C., & Sarstedt, M. (2016). A Primer on Partial Least
Squares Structural Equation Modeling (PLS-SEM) - Joseph F. Hair, Jr., G.
Tomas M. Hult, Christian Ringle, Marko Sarstedt. In Sage.

Hair, Joseph F. Jr., G. Tomas M. Hult, Christian M. Ringle and Marko Sarstedt
Sage, Ketchen, D. J. (2017). A Primer on Partial Least Squares Structural
Equation Modeling. SAGE Publications Vol. 46, pp. 184–185.

Hankinson, G., & Cowking, P. (1995). What do you really mean by a brand?
Journal of Brand Management, 3(1), 43–50.
https://doi.org/10.1057/bm.1995.29

Hazan, C., & Zeifman, D. (1999). Pair bonds as attachments. Handbook of

Attachment: Theory, Research, and Clinical Applications, 336–354.

Henseler, J., Ringle, C.M. and Sinkovics, R.R. (2009), “The use of partial least
squares path modeling in international marketing”, Advances in International
Marketing, Vol. 20, pp. 277-320.

Henseler, J., & Chin, W. W. (2010). A Comparison of Approaches for the Analysis
of Interaction Effects Between Latent Variables Using Partial Least Squares
Path Modeling. Structural Equation Modeling: A Multidisciplinary Journal,
17(1), 82–109. https://doi.org/10.1080/10705510903439003

Henseler, J. and Fassott, G. (2010), “Testing moderating effects in PLS path

models: an illustration of available procedures”, in Esposito, V.V., Chin,
W.W., Henseler, J. and Wang, H. (Eds), Handbook of Partial Least Squares:
Concepts, Methods and Applications, Springer, Berlin, pp. 713-735.

Henseler, J., Ringle, C.M. and Sarstedt, M. (2012), “Using partial least squares path
modeling in advertising research: basic concepts and recent issues”, Handbook
of Research on International Advertising, Edward Elgar Publishing,
Cheltenham.

Henseler, J., & Sarstedt, M. (2013). Goodness-of-fit indices for partial least squares
path modeling. Computational Statistics, 28(2), 565–580.

Hoben, J. B. (1954). English communication at Colgate re-examined. Journal of

Universitas Indonesia
 120

Communication, 4(3), 76–83.

Hollebeek, L. D., Glynn, M. S., & Brodie, R. J. (2014). Consumer brand

engagement in social media: Conceptualization, scale development and
validation. Journal of Interactive Marketing, 28(2), 149–165.
https://doi.org/10.1016/j.intmar.2013.12.002

Hootsuite. (2019). Digital 2019: Indonesia — DataReportal – Global Digital

Insights. 77. https://datareportal.com/reports/digital-2019-
indonesia?rq=Digital 2019%3A Indonesia

Howard, J. A. (1977). Consumer behavior: Application of theory (Vol. 325).

McGraw-Hill New York.

Howard, J. A., & Sheth, J. N. (1969). The theory of buyer behavior.

Hu, L. T., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance
structure analysis: Conventional criteria versus new alternatives. Structural
Equation Modeling, 6(1), 1–55. https://doi.org/10.1080/10705519909540118

IPSOS. (2018). Instagram’s Impact on Indonesian Business.

https://www.ipsos.com/en/instagrams-impact-indonesian-businesses

Jörg, H. (2009). The use of partial least squares path modeling in international
marketing. In R. C. M., R. R. Sinkovics, & P. N. Ghauri (Eds.), New
Challenges to International Marketing (Vol. 20, pp. 277–319). Emerald Group
Publishing Limited. https://doi.org/10.1108/S1474-7979(2009)0000020014

Kang, J., & Park-Poaps, H. (2010). Hedonic and utilitarian shopping motivations
of fashion leadership. Journal of Fashion Marketing and Management, 14(2),
312–328. https://doi.org/10.1108/13612021011046138

Kannan, P. K., & Li, H. “Alice.” (2017). Digital marketing: A framework, review
and research agenda. International Journal of Research in Marketing, 34(1),
22–45. https://doi.org/10.1016/j.ijresmar.2016.11.006

Kaplan, A. M., & Haenlein, M. (2010). Users of the world, unite! The challenges
and opportunities of Social Media. Business Horizons, 53(1), 59–68.
https://doi.org/10.1016/j.bushor.2009.09.003

Universitas Indonesia
 121

Kietzmann, J. H., Hermkens, K., McCarthy, I. P., & Silvestre, B. S. (2011). Social
media? Get serious! Understanding the functional building blocks of social
media. Business Horizons, 54(3), 241–251.
https://doi.org/10.1016/j.bushor.2011.01.005

Kozinets, R. V. (2002). The Field behind the Screen: Using Netnography for
Marketing Research in Online Communities. Journal of Marketing Research,
39(1), 61–72. https://doi.org/10.1509/jmkr.39.1.61.18935

Kramer, T. (2007). The effect of measurement task transparency on preference

construction and evaluations of personalized recommendations. Journal of
Marketing Research, 44(2), 224–233.

Krisdamarjati, Y. A. (2020, February 29). Ventela dan Tantangan Popularitas

Sepatu Lokal di Indonesia.
https://bebas.kompas.id/baca/riset/2020/02/29/ventela-dan-tantangan-
popularitas-sepatu-lokal-di-indonesia/.

Leckie, C., Nyadzayo, M. W., & Johnson, L. W. (2016). Antecedents of consumer

brand engagement and brand loyalty. Journal of Marketing Management,
32(5–6), 558–578. https://doi.org/10.1080/0267257X.2015.1131735

Lee, S. H., & Workman, J. E. (2015). Determinants of brand loyalty: Self-construal,

self-expressive brands, and brand attachment. International Journal of
Fashion Design, Technology and Education, 8(1), 12–20.
https://doi.org/10.1080/17543266.2014.974691

Li, C. (2016). When does web-based personalization really work? The distinction
between actual personalization and perceived personalization. Computers in
Human Behavior, 54, 25–33. https://doi.org/10.1016/j.chb.2015.07.049

Lohmöller, J.-B. (2013). Latent variable path modeling with partial least squares.
Springer Science & Business Media.

Malhotra, N. K. (2010). Marketing research: An applied orientation. Upper Saddle

River, NJ: Pearson/Prentice Hall.

Universitas Indonesia
 122

Maslowska, E., Smit, E. G., & van den Putte, B. (2016). It Is All in the Name: A
Study of Consumers’ Responses to Personalized Communication. Journal of
Interactive Advertising, 16(1), 74–85.
https://doi.org/10.1080/15252019.2016.1161568

McKinsey & Company. (2016). Unlocking Indonesia’s digital opportunity. October

2016. Accessed: January, 2020

Mohammed, R., Fisher, R. J., Jaworski, B. J., & Paddison, G. J. (2004). Internet
marketing. Auflage, New York.

Munaf, T., & Pesik, R. J. (2019). Ekonomi Kreatif outlook 2019. Badan Ekonomi
Kreatif, 23–24.

Muniz Jr., A. M., & O’Guinn, T. C. (2001). Brand Community. Journal of

Consumer Research, 27(4), 412–432. https://doi.org/10.1086/319618

Muñiz Jr, A. M., & Schau, H. J. (2011). How to inspire value-laden collaborative
consumer-generated content. Business Horizons, 54(3), 209–217.

Newberry, C. “How to Advertise on Instagram: A 5-Step Guide to Using Instagram.

Accessed on 28th April 2020. Ads.” Hootsuite Social Media Management, 13
Apr. 2020. Retrieved from blog.hootsuite.com/instagram-ads-guide/.

Nikhashemi, S. R., Valaei, N., & Tarofder, A. K. (2017). Does Brand Personality
and Perceived Product Quality Play a Major Role in Mobile Phone
Consumers’ Switching Behaviour? Global Business Review, 18(3_suppl),
S108–S127. https://doi.org/10.1177/0972150917693155

Noar, S. M., Harrington, N. G., & Aldrich, R. S. (2009). The Role of Message
Tailoring in the Development of Persuasive Health Communication Messages.
Annals of the International Communication Association, 33(1), 73–133.
https://doi.org/10.1080/23808985.2009.11679085

Oliver, R. L. (1999). Whence consumer loyalty? Journal of Marketing,

63(SUPPL.), 33–44. https://doi.org/10.2307/1252099

Park, C. W., Macinnis, D. J., & Priester, J. (2006). *C. Whan Park is Joseph A.

Universitas Indonesia
 123

DeBell Professor of Marketing, Deborah J. MacInnis is Charles and Ramona

I. Hilliard Professor of Marketing, and Joseph Priester is Assistant Professor
of Marketing, all three at Marshall School of Business, University of Southe.
12(2).

Park, T., Shenoy, R., & Salvendy, G. (2008). Effective advertising on mobile
phones: A literature review and presentation of results from 53 case studies.
Behaviour and Information Technology, 27(5), 355–373.
https://doi.org/10.1080/01449290600958882

Pervan, S. J., Bove, L. L., & Johnson, L. W. (2009). Reciprocity as a key

stabilizing norm of interpersonal marketing relationships: Scale development
and validation. Industrial Marketing Management, 38(1), 60–70.

Petty, R. E., & Cacioppo, J. T. (1986). The Elaboration Likelihood Model of

Persuasion BT - Communication and Persuasion: Central and Peripheral
Routes to Attitude Change (R. E. Petty & J. T. Cacioppo (eds.); pp. 1–24).
Springer New York. https://doi.org/10.1007/978-1-4612-4964-1_1

Petty, R. E., Wheeler, S. C., & Bizer, G. Y. (2000). Attitude functions and
persuasion: An elaboration likelihood approach to matched versus mismatched
messages. Why We Evaluate: Functions of Attitudes, 133–162.

Pletikosa Cvijikj, I., & Michahelles, F. (2013). Online engagement factors on

Facebook brand pages. Social Network Analysis and Mining, 3(4), 843–861.
https://doi.org/10.1007/s13278-013-0098-8

Ravi Dhar, & Klaus Wertenbroch. (2000). Consumer Choice Between Hedonic and
Utilitarian Goods. Journal of Marketing Research, XXXVII(February), 60–71.

Ries, A., Trout, J., & Askey, B. (1986). Positioning: The battle for your mind (Vol.
1). McGraw-Hill New York.

Rimer, B. K., & Kreuter, M. W. (2006). Advancing Tailored Health

Communication: A Persuasion and Message Effects Perspective. Journal of
Communication, 56(suppl_1), S184–S201. https://doi.org/10.1111/j.1460-
2466.2006.00289.x

Universitas Indonesia
 124

Ringle, C. M., Sarstedt, M., & Straub, D. W. (2012). A critical look at the use of
PLS-SEM in MIS quarterly. MIS Quarterly: Management Information
Systems, 36(1). https://doi.org/10.2307/41410402

Roberts, M. Lou, & Zahay, D. (2012). Internet marketing: Integrating online and
offline strategies. Cengage Learning.

Ross, N. (1992). A history of direct marketing. Unpublished Paper, Direct

Marketing Association, New York, NY.

Sarstedt, M., Hair, J. F., Cheah, J. H., Becker, J. M., & Ringle, C. M. (2019). How
to specify, estimate, and validate higher-order constructs in PLS-SEM.
Australasian Marketing Journal, 27(3), 197–211.
https://doi.org/10.1016/j.ausmj.2019.05.003

Sashi, C. M. (2012). Customer engagement, buyer-seller relationships, and social

media. Management Decision, 50(2), 253–272.
https://doi.org/10.1108/00251741211203551

Schiffman, L., Bednall, D., O’Cass, A., Paladino, A., & Kanuk, L. (2005).
Consumer behaviour . Frenchs Forest. New South Wales: Pearson Education
Australia.

Schivinski, B., & Dabrowski, D. (2015). The impact of brand communication on

brand equity through Facebook. Journal of Research in Interactive
Marketing, 9(1), 31–53. https://doi.org/10.1108/JRIM-02-2014-0007

Sembada, A. Y., & Koay, K. Y. (2019). How perceived behavioral control affects
trust to purchase in social media stores. Journal of Business Research,
September, 0–1. https://doi.org/10.1016/j.jbusres.2019.09.028

Seok Sohn, Y., Han, J. K., & Lee, S.-H. (2012). Communication strategies for
enhancing perceived fit in the CSR sponsorship context. International Journal
of Advertising, 31(1), 133–146.

Shanahan, T., Tran, T. P., & Taylor, E. C. (2019). Getting to know you: Social
media personalization as a means of enhancing brand loyalty and perceived
quality. Journal of Retailing and Consumer Services, 47(January 2018), 57–

Universitas Indonesia
 125

65. https://doi.org/10.1016/j.jretconser.2018.10.007

Shannon, C. E., & Weaver, W. (1949). The mathematical theory of communication.

Univ. Illinois Press, Urbana, I, 11, 117.

Smith, K. (2007). What is digital marketing?[Web log message]. Hämtad 2014-04-

07 http://digitalmarketing101. blogspot. se/2007/10/what-is ….

Solomon, M. R. (2009). Marketing: Real people, real decisions. Pearson Education.

Sprott, D., Czellar, S., & Spangenberg, E. (2009). The importance of a general
measure of brand engagement on market behavior: Development and
validation of a scale. Journal of Marketing Research, 46(1), 92–104.
https://doi.org/10.1509/jmkr.46.1.92

Srinivasan, S. S., Anderson, R., & Ponnavolu, K. (2002). Customer loyalty in e-

commerce: An exploration of its antecedents and consequences. Journal of
Retailing, 78(1), 41–50. https://doi.org/10.1016/S0022-4359(01)00065-3

Stone, M. (1974). Cross‐validatory choice and assessment of statistical predictions.

Journal of the Royal Statistical Society: Series B (Methodological), 36(2),
111–133.

Strahilevitz, M., & Myers, J. G. (1998). Donations to Charity as Purchase

Incentives: How Well They Work May Depend on What You Are Trying to
Sell. Journal of Consumer Research, 24(4), 434–446.
https://doi.org/10.1086/209519

Sunikka, A., & Bragge, J. (2009). Promotional messages in multichannel banking:

Attractive or annoying? Journal of Financial Services Marketing, 14(3), 245–
263. https://doi.org/10.1057/fsm.2009.22

Tapscott, D., & Williams, A. D. (2007). How Mass Collaboration Changes

Everything–Wikinomics. New York, Penguin.

Tenenhaus, M., Vinzi, V. E., Chatelin, Y. M., & Lauro, C. (2005). PLS path
modeling. Computational Statistics and Data Analysis, 48(1), 159–205.
https://doi.org/10.1016/j.csda.2004.03.005

Universitas Indonesia
 126

Thompson, M., MacInnis, D., & Park, W. (2005). The ties that bind: measuring the
strength of customers attachment to brands. Journal of Consumer Psychology,
15(1), 77–91.

Tiago, M. T. P. M. B., & Veríssimo, J. M. C. (2014). Digital marketing and social

media: Why bother? Business Horizons, 57(6), 703–708.
https://doi.org/10.1016/j.bushor.2014.07.002

Tran, T. P. (2017). Personalized ads on Facebook: An effective marketing tool for

online marketers. Journal of Retailing and Consumer Services, 39, 230–242.

Vesanen, J. (2007). What is personalization? A conceptual framework. European

Journal of Marketing, 41(5–6), 409–418.
https://doi.org/10.1108/03090560710737534

Wee Meng, P. (2020). Deloitte Consumer Insights Dawn of the digital age in
Indonesia. January.

Wetzels, M., Odekerken-Schröder, G., & Oppen, C. Van. (2009). Assessing Using
PLS Path Modeling Hierarchical and Empirical Construct Models :
Guidelines. MIS Quarterly, 33(1), 177–195.

Whan Park, C., Maclnnis, D. J., Priester, J., Eisingerich, A. B., & Lacobucci, D.
(2010). Brand attachment and brand attitude strength: Conceptual and
empirical differentiation of two critical brand equity drivers. Journal of
Marketing, 74(6), 1–17. https://doi.org/10.1509/jmkg.74.6.1

Wymbs, C. (2011). Digital marketing: The time for a new “academic major” has
arrived. Journal of Marketing Education, 33(1), 93–106.
https://doi.org/10.1177/0273475310392544

Yoo, B., & Donthu, N. (2001). Developing and validating a multidimensional

consumer-based brand equity scale. Journal of Business Research, 52(1), 1–
14. https://doi.org/10.1016/S0148-2963(99)00098-3

Zeithaml, V. A. (1988). Consumer Perceptions of Price, Qualitye. Journal of

Marketing, 52(3), 2–22. https://doi.org/10.2307/1251446

Universitas Indonesia
 127

Zhang, K. Z. K., Benyoucef, M., & Zhao, S. J. (2015). Consumer participation and
gender differences on companies’ microblogs: A brand attachment process
perspective. Computers in Human Behavior, 44, 357–368.
https://doi.org/10.1016/J.CHB.2014.11.068

Universitas Indonesia
 Attachment 1 Research Questionnaire

Kuesioner Penelitian

Responden Yth.

Perkenalkan saya Adhiwira Indrabrata, mahasiswa Fakultas Ekonomi dan Bisnis

Universitas Indonesia. Saat ini saya sedang melakukan penelitian mengenai
pengaruh personalized advertising (proses iklan yang dipersonalisasi) terhadap
loyalitas konsumen merek sepatu lokal yang beriklan di Instagram. Jika Anda
merupakan salah satu pengguna Instagram, pengikut akun Instagram, serta
memiliki satu produk dari merek sepatu lokal Brodo, Portee Goods, Geoff Max,
dan NAH Project, saya mengharapkan bantuan Anda untuk mengisi kuesioner
berikut dengan memberikan penilaian secara objektif. Atas bantuan dan partisipasi
Anda, saya mengucapkan terima kasih.

PEMBUKAAN
MENGENAL IKLAN YANG DIPERSONALISASI

Personalization merupakan cara penyiaran iklan suatu merek produk atau jasa yang
telah disesuaikan kepada aktivitas seseorang di Internet (Shanahan et al., 2019).

Pada platform Instagram, terdapat fitur Instagram ads yang digunakan merek untuk
menyiarkan iklannya dalam bentuk Instagram story atau post yang bisa muncul dan
dapat dilihat oleh user Instagram yang memiliki kriteria sesuai dengan target
audience merek tersebut(seperti: status following Anda ke suatu merek, minat Anda
terhadap jenis produk tertentu, search query (kata pencarian) Anda di Instagram,
dan juga merek yang teman-teman Anda ikuti di Instagram (Hootsuite., 2020).

128
Universitas Indonesia
 129

BAGIAN I

SCREENING QUESTION

Berilah tanda silang (x) untuk jawaban yang sesuai dengan Anda

1. Apakah Anda memiliki produk dari merek sepatu lokal berikut ini? (boleh
lebih dari 1 merek)
a. Brodo
b. Portee Goods
c. Geoff Max
d. NAH Project
e. Tidak ada (Terima kasih atas partisipasinya, Anda tidak perlu
melanjutkan mengisi kuesioner ini)

2. Dari merek sepatu tersebut (no.1), sebutkan merek yang paling terakhir
Anda beli
a. Brodo
b. Portee Goods
c. Geoff Max
d. NAH Project

3. Apakah Anda menggunakan Instagram?

a. Ya b. Tidak (terima kasih atas partisipasinya, Anda tidak perlu
melanjutkan mengisi kuesioner ini)

Universitas Indonesia
 130

4. Apakah Anda mengikuti akun Instagram merek yang Anda sebutkan pada
pertanyaan (No.2)?

a. Ya b. Tidak (Terima kasih atas partisipasinya, Anda tidak perlu

melanjutkan mengisi kuesioner ini

5. Apakah Anda berusia 18-34 tahun?

a. Ya b. Tidak (Terimakasih atas partisipasinya, Anda tidak perlu

melanjutkan mengisi kuesioner ini)

6. Apakah Anda berdomisili di Pulau Jawa?

a. Ya b. Tidak (Terima kasih atas partisipasinya, Anda tidak perlu

melanjutkan mengisi kuesioner ini)

Universitas Indonesia
 131

BAGIAN II
PERTANYAAN UTAMA

Berikan jawaban Anda terhadap pernyataan-pernyaataan berikut ini terkait dengan

MEREK SEPATU yang paling terakhir Anda beli (sesuai jawaban Anda pada
pertanyaan 2).
Berilah tanda silang (x) pada kotak yang paling sesuai dengan Anda (dari skala 1-
7), dengan rincian sebagai berikut:
1: sangat tidak setuju
2: tidak setuju
3: agak tidak setuju
4: netral
5: agak setuju
6: setuju
7: sangat Setuju.

Item Pertanyaan 1 2 3 4 5 6 7

Perceived Personalization

Iklan dari merek tersebut di Instagram memberikan

rekomendasi pembelian yang sesuai dengan kebutuhan
saya

Iklan dari merek tersebut di Instagram memungkinkan

saya untuk mendapatkan produk yang dibuat sesuai
untuk saya

Secara keseluruhan, iklan dari merek tersebut yang

saya temui di Instagram disesuaikan dengan situasi
yang saya alami

Universitas Indonesia
 132

Saya merasa menjadi pelanggan yang unik ketika

melihat iklan dari merek tersebut di Instagram

Saya percaya bahwa iklan dari merek tersebut di

Instagram disesuaikan dengan kebutuhan saya

Item Pertanyaan 1 2 3 4 5 6 7

Self-Expressive Brand

Iklan dari merek tersebut di Instagram melambangkan

siapa saya sebenarnya

Iklan dari merek tersebut di Instagram mencerminkan

kepribadian saya

Iklan dari merek tersebut di Instagram merupakan

gambaran dari siapa saya

Iklan dari merek tersebut di Instagram mencerminkan

diri saya yang sebenarnya

Iklan dari merek tersebut di Instagram berkontribusi

terhadap citra diri saya

Consumer Brand Engagement

Cognitive Processing

Melihat iklan dari merek tersebut di Instagram

membuat saya memikirkan mereknya

Saya sering memikirkan merek tersebut ketika saya

melihat iklannya di Instagram

Saya tertarik untuk mempelajari lebih banyak

mengenai merek tersebut ketika saya melihat iklannya
di Instagram

Affection

Universitas Indonesia
 133

Saya merasa sangat positif ketika saya melihat iklan

dari merek tersebut di Instagram

Saya menjadi senang ketika saya melihat iklan dari

merek tersebut di Instagram

Saya merasa baik ketika saya melihat iklan dari merek

tersebut di Instagram

Saya bangga ketika saya melihat iklan dari merek

tersebut di Instagram

Item Pertanyaan 1 2 3 4 5 6 7

Activation

Saya lebih banyak menghabiskan waktu melihat iklan

dari merek tersebut dibandingkan jenis iklan lainnya di
Instagram

Kapanpun saya melihat iklan merek tersebut di

Instagram, saya biasanya membuka iklannya untuk
melihat lebih lanjut

Merek tersebut adalah salah satu merek yang biasanya

saya coba cari setelah melihat iklannya di Instagram

Brand Attachment

Affection

Saya merasakan kasih sayang terhadap merek tersebut

ketika saya melihat iklannya di instagram

Saya merasa sangat mengenal merek tersebut ketika

saya melihat iklannya di instagram

Universitas Indonesia
 134

Saya merasa mencintai merek tersebut ketika saya

melihat iklannya di instagram

Saya merasa damai dengan merek tersebut ketika

melihat iklannya di Instagram

Connection

Saya merasa memiliki hubungan yang sangat pribadi

dengan merek tersebut ketika meilhat iklannya di
instagram

Saya merasa dekat dengan merek tersebut ketika

melihat iklannya di instagram

Saya merasa terikat dengan merek tersebut ketika

melihat iklannya di instagram

Passion

Saya bersemangat dengan merek tersebut ketika saya

melihat iklannya di Instagram

Saya merasa sangat senang terhadap merek tersebut

ketika melihat iklannya di Instagram

Saya merasa terpikat dengan merek tersebut ketika

saya meilhat iklannya di instagram

Universitas Indonesia
 135

Item Pertanyaan 1 2 3 4 5 6 7

Perceived Quality

Merek tersebut memiliki kualitas tinggi

Merek tersebut kemungkinan memiliki kualitas yang

sangat tinggi

Merek tersebut kemungkinan memiliki nilai fungsional

yang sangat tinggi

Merek tersebut sangat mungkin untuk diandalkan

Merek tersebut harusnya memiliki kualitas yang sangat

baik

Brand Loyalty

Saya menilai diri saya setia terhadap merek tersebut

Merek tersebut akan menjadi pilihan pertama saya di

kategori merek sepatu lokal

Saya tidak akan membeli merek lain apabila merek

tersebut masih tersedia di toko

BAGIAN III

DATA RESPONDEN

Berilah tanda silang (x) untuk jawaban yang sesuai dengan Anda

1. Jenis Kelamin
a. Pria
b. Wanita

Universitas Indonesia
 136

2. Status:
a. Tidak menikah
b. Menikah
c. Pernah menikah

3. Usia (diisi dalam angka) .... tahun

4. Pendidikan terakhir atau yang sedang dijalani saat ini

a. SD
b. SMP
c. SMA/Sederajat
d. Diploma
e. S1
f. S2
g. S3
h. Lainnya (....)

5. Pekerjaan
a. Pelajar/Mahasiswa
b. Pengusaha
c. Pegawai Negeri
d. Pegawai Swasta/Pegawai BUMN
e. Lainnya (........)

6. Domisili Anda (Isi dengan Kota, Provinsi)

7. Jumlah pengeluaran per bulan

a. < Rp 1.500.000
b. Rp 1.500.000 – Rp 3.000.000
c. Rp 3.000.000 – Rp 5.000.000
d. Rp 5.000.000 – Rp 10.000.000
e. > Rp 10.000.000

Universitas Indonesia
 137

8. Sudah berapa lama Anda mengikuti Akun Instagram brand (Brodo/Geoff

Max/NAH Project/Portee Goods)?
a. < 6 bulan
b. 6 – 12 bulan
c. 13 – 24 bulan
d. > 24 bulan

9. Akun sosial media yang Anda miliki (boleh pilih lebih dari 1)
a. Instagram
b. Facebook
c. Twitter
d. Linkedin
e. Messenger App: WhatsApp/Line/WeChat

10. Waktu rata-rata per hari Anda dalam menggunakan social media
a. Kurang dari 1 jam
b. 1 – 3 jam
c. 3 – 5 jam
d. 5 – 7 jam
e. Lebih dari 7 jam

Universitas Indonesia
 Attachment 2 SMART-PLS Result

Repeated Indicator ApproachOuter model: Calculate → PLS Algorithm →

Loading & Obtain Latent Variable Score

138
Universitas Indonesia
 139

Two Stage Approach

Outer model: Calculate → PLS Algorithm → Assessment of Loading and R2
Inner model: Calculate → Bootstrapping → Assessment of Path Coefficient (T-
value)

Universitas Indonesia
 140

Base Data

Setting
Data file Settings
Data file MAIN TEST ADHI WITH LVS [386 records]
Missing value marker none
Data Setup Settings
Algorithm to handle missing data None
Weighting Vector -
PLS Algorithm Settings
Data metric Mean 0, Var 1
Initial Weights 1.0
Max. number of iterations 300
Stop criterion 7
Use Lohmoeller settings? No
Weighting scheme Path
Bootstrapping Settings
Complexity Complete Bootstrapping
Bias-Corrected and Accelerated (BCa)
Confidence interval method
Bootstrap
Parallel processing Yes
Samples 5000
Significance level 0.05
Test type One Tailed
Construct Outer Weighting Mode
Settings
BA Automatic
BL Automatic
CBE Automatic
PP Automatic
PQ Automatic
SEB Automatic

Universitas Indonesia
 141

Model Selection Criteria

AICu HQc
AIC AICc (Corrected BIC HQ
(Unbiased (Corrected
(Akaike's Akaikes (Bayesian (Hannan
Akaikes Hannan-
Information Information Information Quinn
Information Quinn
Criterion) Criterion) Criteria) Criterion)
Criterion Criterion)
BA -247.067 -244.055 141.038 -235.200 -242.361 -242.220
BL -256.747 -252.727 131.410 -240.924 -250.472 -250.247
CBE -237.410 -234.398 150.695 -225.542 -232.703 -232.563
PQ -134.883 -131.871 253.222 -123.015 -130.177 -130.036

Universitas Indonesia
 Attachment 3 IBM SPSS 23 PRE-TEST RESULT

RELIABILITY
/VARIABLES= ALL VARIABLES
/SCALE('ALL VARIABLES') ALL
/MODEL=ALPHA
/STATISTICS=DESCRIPTIVE SCALE
/SUMMARY=TOTAL.

Case Processing Summary

N %
Cases Valid 30 100.0
a
Excluded 0 .0
Total 30 100.0
a. Listwise deletion based on all variables in the procedure.

Reliability Statistics
Cronbach's Alpha N of Items
.673 5

Item-Total Statistics
Scale Mean if Scale Variance if Corrected Item-Total Cronbach's Alpha if
Item Deleted Item Deleted Correlation Item Deleted
PP1 23.167 8.420 .349 .653
PP2 23.300 7.666 .554 .580
PP3 23.333 6.782 .484 .594
PP4 23.533 7.430 .262 .721
PP5 23.067 7.030 .588 .552

RELIABILITY
/VARIABLES=SEB1 SEB2 SEB3 SEB4 SEB5

Reliability Statistics
Cronbach's Alpha N of Items
.916 5

142
Universitas Indonesia
 143

Item-Total Statistics
Cronbach's
Scale Mean if Scale Variance if Corrected Item- Alpha if Item
Item Deleted Item Deleted Total Correlation Deleted
SEB1 20.400 22.593 .851 .883
SEB2 20.500 24.121 .740 .907
SEB3 20.467 23.292 .844 .886
SEB4 20.700 23.597 .761 .903
SEB5 19.933 25.306 .736 .907

RELIABILITY
/VARIABLES=ECP1 ECP2 ECP3
.
Reliability Statistics
Cronbach's Alpha N of Items
.831 3

Item-Total Statistics
Scale Mean if Scale Variance if Corrected Item-Total Cronbach's Alpha if
Item Deleted Item Deleted Correlation Item Deleted
ECP1 11.967 3.551 .714 .746
ECP2 12.033 2.723 .796 .662
ECP3 11.867 4.189 .599 .853

RELIABILITY
/VARIABLES=EAF1 EAF2 EAF3 EAF4

Reliability Statistics
Cronbach's Alpha N of Items
.852 4

Item-Total Statistics
Scale Mean if Scale Variance if Corrected Item-Total Cronbach's Alpha if
Item Deleted Item Deleted Correlation Item Deleted
EAF1 16.033 8.102 .715 .805
EAF2 16.067 7.030 .850 .742
EAF3 16.067 8.064 .690 .814
EAF4 16.333 8.023 .549 .880

Universitas Indonesia
 144

RELIABILITY
/VARIABLES=EAC1 EAC2 EAC3
.
Reliability Statistics
Cronbach's Alpha N of Items
.797 3

Item-Total Statistics
Scale Mean if Scale Variance if Corrected Item-Total Cronbach's Alpha if
Item Deleted Item Deleted Correlation Item Deleted
EAC1. 10.733 5.375 .528 .875
EAC2 10.400 5.834 .646 .719
EAC3 10.267 5.720 .799 .586

RELIABILITY
/VARIABLES=AAF1 AAF2 AAF3 AAF4
Reliability Statistics
Cronbach's Alpha N of Items
.895 4

Item-Total Statistics
Scale Mean if Scale Variance if Corrected Item-Total Cronbach's Alpha if
Item Deleted Item Deleted Correlation Item Deleted
AAF1 15.500 11.224 .777 .862
AAF2 14.767 13.220 .647 .907
AAF3 15.400 10.662 .849 .834
AAF4 15.633 10.930 .810 .849

RELIABILITY
/VARIABLES=ACC1 ACC2 ACC3

Reliability Statistics
Cronbach's Alpha N of Items
.941 3

Universitas Indonesia
 145

Item-Total Statistics
Cronbach's
Scale Mean if Scale Variance if Corrected Item- Alpha if Item
Item Deleted Item Deleted Total Correlation Deleted
ACC1 9.967 9.068 .871 .919
ACC2 9.800 8.924 .865 .923
ACC3 10.033 8.171 .898 .898

RELIABILITY
/VARIABLES=APS1 APS2 APS3

Reliability Statistics
Cronbach's Alpha N of Items
.947 3

Item-Total Statistics
Cronbach's
Scale Mean if Scale Variance if Corrected Item- Alpha if Item
Item Deleted Item Deleted Total Correlation Deleted
APS1
11.033 6.654 .879 .934

APS2 11.167 5.316 .926 .903

APS3 11.067 6.478 .885 .928

RELIABILITY
/VARIABLES=PQ1 PQ2 PQ3 PQ4 PQ5

Reliability Statistics
Cronbach's Alpha N of Items
.920 5

Item-Total Statistics
Scale Mean if Scale Variance if Corrected Item-Total Cronbach's Alpha if
Item Deleted Item Deleted Correlation Item Deleted
PQ1 23.367 12.516 .843 .891
PQ2 23.600 14.110 .604 .937
PQ3 23.533 13.361 .818 .898
PQ4 23.533 12.051 .889 .881
PQ5 23.300 12.424 .823 .895

Universitas Indonesia
 146

RELIABILITY
/VARIABLES=BL1 BL2 BL3

Reliability Statistics
Cronbach's Alpha N of Items
.897 3

Item-Total Statistics
Scale Mean if Scale Variance if Corrected Item-Total Cronbach's Alpha if
Item Deleted Item Deleted Correlation Item Deleted
BL1 10.467 10.257 .803 .847
BL2 10.000 10.690 .768 .877
BL3 10.467 9.775 .820 .833

FACTOR
/ALL VARIABLES
/MISSING LISTWISE
/ANALYSIS PP1 PP2 PP3 PP4 PP5
/PRINT INITIAL KMO AIC EXTRACTION
/CRITERIA FACTORS(1) ITERATE(25)
/EXTRACTION PC
/ROTATION NOROTATE
/SAVE REG(ALL)
/METHOD=CORRELATION.

FACTOR
/VARIABLES PP1 PP2 PP3 PP4 PP5

KMO and Bartlett's Test

Kaiser-Meyer-Olkin Measure of Sampling Adequacy. .507
Bartlett's Test of Sphericity Approx. Chi-Square 45.914
df 10
Sig. .000

Universitas Indonesia
 147

Anti-image Matrices
PP1 PP2 PP3 PP4 PP5
Anti-image Covariance PP1 .420 -.183 -.274 .240 -.092
PP2 -.183 .514 .001 -.013 -.213
PP3 -.274 .001 .576 -.247 .084
PP4 .240 -.013 -.247 .510 -.259
PP5 -.092 -.213 .084 -.259 .494
Anti-image Correlation PP1 .465a -.394 -.557 .518 -.201
PP2 -.394 .691a .002 -.025 -.424
PP3 -.557 .002 .483a -.456 .157
PP4 .518 -.025 -.456 .335a -.517
PP5 -.201 -.424 .157 -.517 .574a
a. Measures of Sampling Adequacy(MSA)

Communalities
Initial Extraction
PP1 1.000 .474
PP2 1.000 .664
PP3 1.000 .469
PP4 1.000 .165
PP5 1.000 .569

Total Variance Explained

Initial Eigenvalues Extraction Sums of Squared Loadings
Component Total % of Variance Cumulative % Total % of Variance Cumulative %
1 2.340 46.807 46.807 2.340 46.807 46.807
2 1.313 26.260 73.068
3 .798 15.967 89.035
4 .330 6.590 95.626
5 .219 4.374 100.000

Universitas Indonesia
 148

Component Matrixa
Component
1
PP1 .688
PP2 .815
PP3 .685
PP4 .406
PP5 .754

FACTOR
/VARIABLES PP1 PP2 PP3 PP5
KMO and Bartlett's Test
Kaiser-Meyer-Olkin Measure of Sampling Adequacy. .631
Bartlett's Test of Sphericity Approx. Chi-Square 28.450
df 6
Sig. .000

Anti-image Matrices
PP1 PP2 PP3 PP5
Anti-image Covariance PP1 .574 -.242 -.272 .056
PP2 -.242 .514 -.007 -.300
PP3 -.272 -.007 .727 -.072
PP5 .056 -.300 -.072 .674
Anti-image Correlation PP1 .623a -.446 -.421 .091
PP2 -.446 .613a -.011 -.510
PP3 -.421 -.011 .696a -.103
PP5 .091 -.510 -.103 .617a
a. Measures of Sampling Adequacy(MSA)

Communalities
Initial Extraction
PP1 1.000 .628
PP2 1.000 .698
PP3 1.000 .461
PP5 1.000 .464
Extraction Method: Principal Component Analysis.

Universitas Indonesia
 149

Total Variance Explained

Initial Eigenvalues Extraction Sums of Squared Loadings
Component Total % of Variance Cumulative % Total % of Variance Cumulative %
1 2.250 56.256 56.256 2.250 56.256 56.256
2 .885 22.117 78.373
3 .547 13.675 92.048
4 .318 7.952 100.000
Extraction Method: Principal Component Analysis.

Component Matrixa
Component
1
PP1 .792
PP2 .835
PP3 .679
PP5 .681

RELIABILITY
/VARIABLES=PP1 PP2 PP3 PP5
/SCALE('ALL VARIABLES') ALL
/MODEL=ALPHA
/STATISTICS=DESCRIPTIVE SCALE
/SUMMARY=TOTAL.
Case Processing Summary
N %
Cases Valid 30 100.0
Excludeda 0 .0
Total 30 100.0
a. Listwise deletion based on all variables in the
procedure.

Universitas Indonesia
 150

Reliability Statistics
Cronbach's Alpha N of Items
.721 4

Item-Total Statistics
Scale Mean if Scale Variance if Corrected Item- Cronbach's Alpha
Item Deleted Item Deleted Total Correlation if Item Deleted
PP1 17.600 4.662 .583 .622
PP2 17.733 4.616 .626 .601
PP3 17.767 4.185 .448 .714
PP5 17.500 4.810 .432 .704

FACTOR
/VARIABLES SEB1 SEB2 SEB3 SEB4 SEB5
KMO and Bartlett's Test
Kaiser-Meyer-Olkin Measure of Sampling Adequacy. .880
Bartlett's Test of Sphericity Approx. Chi-Square 94.346
df 10
Sig. .000

Anti-image Matrices
SEB1 SEB2 SEB3 SEB4 SEB5
Anti-image Covariance SEB1 .265 -.146 -.076 -.088 -.060
SEB2 -.146 .399 -.070 .001 -.011
SEB3 -.076 -.070 .286 -.102 -.111
SEB4 -.088 .001 -.102 .398 -.071
SEB5 -.060 -.011 -.111 -.071 .439
a
Anti-image Correlation SEB1 .849 -.451 -.277 -.270 -.175
SEB2 -.451 .875a -.207 .002 -.026
a
SEB3 -.277 -.207 .872 -.304 -.315
SEB4 -.270 .002 -.304 .903a -.171
SEB5 -.175 -.026 -.315 -.171 .915a
a. Measures of Sampling Adequacy(MSA)

Universitas Indonesia
 151

Communalities
Initial Extraction
SEB1 1.000 .829
SEB2 1.000 .693
SEB3 1.000 .821
SEB4 1.000 .721
SEB5 1.000 .688
.

Total Variance Explained

Initial Eigenvalues Extraction Sums of Squared Loadings
Component Total % of Variance Cumulative % Total % of Variance Cumulative %
1 3.753 75.062 75.062 3.753 75.062 75.062
2 .458 9.151 84.213
3 .359 7.176 91.389
4 .229 4.586 95.975
5 .201 4.025 100.000

Component Matrixa
Component
1
SEB1 .911
SEB2 .833
SEB3 .906
SEB4 .849
SEB5 .830

FACTOR
/VARIABLES ECP1 ECP2 ECP3
KMO and Bartlett's Test
Kaiser-Meyer-Olkin Measure of Sampling Adequacy. .661
Bartlett's Test of Sphericity Approx. Chi-Square 35.935
df 3
Sig. .000

Anti-image Matrices

Universitas Indonesia
 152

ECP1 ECP2 ECP3

Anti-image Covariance ECP1 .429 -.256 -.032
ECP2 -.256 .354 -.198
ECP3 -.032 -.198 .618
Anti-image Correlation ECP1 .653a -.656 -.062
a
ECP2 -.656 .609 -.423
ECP3 -.062 -.423 .774a
a. Measures of Sampling Adequacy (MSA)

Communalities
Initial Extraction
ECP1 1.000 .763
ECP2 1.000 .848
ECP3 1.000 .641
Extraction Method: Principal Component Analysis.

Total Variance Explained

Initial Eigenvalues Extraction Sums of Squared Loadings
Component Total % of Variance Cumulative % Total % of Variance Cumulative %
1 2.251 75.029 75.029 2.251 75.029 75.029
2 .523 17.423 92.452
3 .226 7.548 100.000
Extraction Method: Principal Component Analysis.

Component Matrixa
Component
1
ECP1 .873
ECP2 .921
ECP3 .800

Universitas Indonesia
 153

FACTOR
/VARIABLES EAF1 EAF2 EAF3 EAF4
KMO and Bartlett's Test
Kaiser-Meyer-Olkin Measure of Sampling Adequacy. .747
Bartlett's Test of Sphericity Approx. Chi-Square 58.953
df 6
Sig. .000

Anti-image Matrices
EAF1 EAF2 EAF3 EAF4
Anti-image Covariance EAF1 .440 -.153 -.008 -.081
EAF2 -.153 .234 -.190 -.120
EAF3 -.008 -.190 .371 .040
EAF4 -.081 -.120 .040 .661
Anti-image Correlation EAF1 .821a -.478 -.019 -.149
a
EAF2 -.478 .670 -.645 -.306
EAF3 -.019 -.645 .730a .081
EAF4 -.149 -.306 .081 .857a
a. Measures of Sampling Adequacy(MSA)

Communalities
Initial Extraction
EAF1 1.000 .724
EAF2 1.000 .875
EAF3 1.000 .711
EAF4 1.000 .510
Extraction Method: Principal Component Analysis.

Total Variance Explained

Initial Eigenvalues Extraction Sums of Squared Loadings
Component Total % of Variance Cumulative % Total % of Variance Cumulative %
1 2.820 70.488 70.488 2.820 70.488 70.488
2 .625 15.637 86.125
3 .396 9.894 96.019
4 .159 3.981 100.000
Extraction Method: Principal Component Analysis.

Universitas Indonesia
 154

Component Matrixa
Component
1
EAF1 .851
EAF2 .935
EAF3 .843
EAF4 .714

FACTOR
/VARIABLES EAC1 EAC2 EAC3

KMO and Bartlett's Test

Kaiser-Meyer-Olkin Measure of Sampling Adequacy. .605
Bartlett's Test of Sphericity Approx. Chi-Square 37.160
df 3
Sig. .000

Anti-image Matrices
EAC1 EAC2 EAC3
Anti-image Covariance EAC1 .656 .036 -.203
EAC2 .036 .387 -.252
EAC3 -.203 -.252 .310
a
Anti-image Correlation EAC1 .714 .071 -.451
EAC2 .071 .596a -.728
EAC3 -.451 -.728 .565a
a. Measures of Sampling Adequacy(MSA)

Communalities
Initial Extraction
EAC1 1.000 .574
EAC2 1.000 .759
EAC3 1.000 .871
Extraction Method: Principal Component Analysis.

Universitas Indonesia
 155

Total Variance Explained

Initial Eigenvalues Extraction Sums of Squared Loadings
Component Total % of Variance Cumulative % Total % of Variance Cumulative %
1 2.204 73.464 73.464 2.204 73.464 73.464
2 .605 20.171 93.635
3 .191 6.365 100.000
Extraction Method: Principal Component Analysis.

Component Matrixa
Component
1
EAC1 .757
EAC2 .871
EAC3 .933

FACTOR
/VARIABLES AAF1 AAF2 AAF3 AAF4

KMO and Bartlett's Test

Kaiser-Meyer-Olkin Measure of Sampling Adequacy. .829
Bartlett's Test of Sphericity Approx. Chi-Square 68.416
df 6
Sig. .000

Anti-image Matrices
AAF1 AAF2 AAF3 AAF4
Anti-image Covariance AAF1 .363 -.024 -.140 -.097
AAF2 -.024 .570 -.103 -.084
AAF3 -.140 -.103 .271 -.133
AAF4 -.097 -.084 -.133 .330
a
Anti-image Correlation AAF1 .836 -.053 -.445 -.280
AAF2 -.053 .907a -.262 -.193
AAF3 -.445 -.262 .778a -.444
AAF4 -.280 -.193 -.444 .829a

Universitas Indonesia
 156

Communalities
Initial Extraction
AAF1 1.000 .771
AAF2 1.000 .615
AAF3 1.000 .851
AAF4 1.000 .809

Total Variance Explained

Initial Eigenvalues Extraction Sums of Squared Loadings
Component Total % of Variance Cumulative % Total % of Variance Cumulative %
1 3.046 76.143 76.143 3.046 76.143 76.143
2 .492 12.294 88.437
3 .268 6.693 95.130
4 .195 4.870 100.000

Component Matrixa
Component
1
AAF1 .878
AAF2 .784
AAF3 .922
AAF4 .899

FACTOR
/VARIABLES ACC1 ACC2 ACC3

KMO and Bartlett's Test

Kaiser-Meyer-Olkin Measure of Sampling Adequacy. .765
Bartlett's Test of Sphericity Approx. Chi-Square 74.400
df 3
Sig. .000

Universitas Indonesia
 157

Anti-image Matrices
ACC1 ACC2 ACC3
Anti-image Covariance ACC1 .235 -.075 -.117
ACC2 -.075 .249 -.111
ACC3 -.117 -.111 .193
Anti-image Correlation ACC1 .780a -.309 -.548
a
ACC2 -.309 .797 -.508
ACC3 -.548 -.508 .724a

Communalities
Initial Extraction
ACC1 1.000 .889
ACC2 1.000 .882
ACC3 1.000 .914

Total Variance Explained

Initial Eigenvalues Extraction Sums of Squared Loadings
Component Total % of Variance Cumulative % Total % of Variance Cumulative %
1 2.685 89.490 89.490 2.685 89.490 89.490
2 .185 6.180 95.670
3 .130 4.330 100.000

Component Matrixa
Component
1
ACC1 .943
ACC2 .939
ACC3 .956

FACTOR
/VARIABLES APS1 APS2 APS3

KMO and Bartlett's Test

Kaiser-Meyer-Olkin Measure of Sampling Adequacy. .754
Bartlett's Test of Sphericity Approx. Chi-Square 83.720
df 3
Sig. .000

Universitas Indonesia
 158

Anti-image Matrices
APS1 APS2 APS3
Anti-image Covariance APS1 .216 -.101 -.040
APS2 -.101 .142 -.103
APS3 -.040 -.103 .205
Anti-image Correlation APS1 .799a -.574 -.191
a
APS2 -.574 .692 -.604
APS3 -.191 -.604 .785a

Communalities
Initial Extraction
APS1 1.000 .893
APS2 1.000 .937
APS3 1.000 .898

Total Variance Explained

Initial Eigenvalues Extraction Sums of Squared Loadings
Component Total % of Variance Cumulative % Total % of Variance Cumulative %
1 2.728 90.935 90.935 2.728 90.935 90.935
2 .177 5.896 96.831
3 .095 3.169 100.000

Component Matrixa
Component
1
APS1 .945
APS2 .968
APS3 .947

Universitas Indonesia
 159

FACTOR
/VARIABLES PQ1 PQ2 PQ3 PQ4 PQ5

KMO and Bartlett's Test

Kaiser-Meyer-Olkin Measure of Sampling Adequacy. .846
Bartlett's Test of Sphericity Approx. Chi-Square 110.800
df 10
Sig. .000

Anti-image Matrices
PQ1 PQ2 PQ3 PQ4 PQ5
Anti-image Covariance PQ1 .282 -.132 -.064 -.018 -.099
PQ2 -.132 .543 .045 -.111 .061
PQ3 -.064 .045 .264 -.106 -.042
PQ4 -.018 -.111 -.106 .186 -.091
PQ5 -.099 .061 -.042 -.091 .250
Anti-image Correlation PQ1 .873a -.338 -.233 -.079 -.374
PQ2 -.338 .817a .118 -.350 .166
PQ3 -.233 .118 .866a -.480 -.162
PQ4 -.079 -.350 -.480 .815a -.424
PQ5 -.374 .166 -.162 -.424 .851a

Communalities
Initial Extraction
PQ1 1.000 .818
PQ2 1.000 .513
PQ3 1.000 .797
PQ4 1.000 .878
PQ5 1.000 .806

Universitas Indonesia
 160

Total Variance Explained

Initial Eigenvalues Extraction Sums of Squared Loadings
% of Cumulative % of Cumulative
Component Total Variance % Total Variance %
1 3.812 76.249 76.249 3.812 76.249 76.249
2 .600 11.991 88.240
3 .254 5.084 93.324
4 .206 4.127 97.451
5 .127 2.549 100.000

Component Matrixa
Component
1
PQ1 .905
PQ2 .716
PQ3 .893
PQ4 .937
PQ5 .898

FACTOR
/VARIABLES BL1 BL2 BL3

KMO and Bartlett's Test

Kaiser-Meyer-Olkin Measure of Sampling Adequacy. .746
Bartlett's Test of Sphericity Approx. Chi-Square 49.932
df 3
Sig. .000

Anti-image Matrices
BL1 BL2 BL3
Anti-image Covariance BL1 .347 -.123 -.182
BL2 -.123 .410 -.148
BL3 -.182 -.148 .324
Anti-image Correlation BL1 .737a -.327 -.542
BL2 -.327 .794a -.407
BL3 -.542 -.407 .715a
a. Measures of Sampling Adequacy(MSA)

Universitas Indonesia
 161

Communalities
Initial Extraction
BL1 1.000 .836
BL2 1.000 .800
BL3 1.000 .852
Extraction Method: Principal Component Analysis.

Total Variance Explained

Initial Eigenvalues Extraction Sums of Squared Loadings
Component Total % of Variance Cumulative % Total % of Variance Cumulative %
1 2.488 82.924 82.924 2.488 82.924 82.924
2 .297 9.886 92.810
3 .216 7.190 100.000
Extraction Method: Principal Component Analysis.

Component Matrixa
Component
1
BL1 .914
BL2 .895
BL3 .923

Universitas Indonesia