Journal of the Academy of Marketing Science

https://doi.org/10.1007/s11747-023-00933-7

ORIGINAL EMPIRICAL RESEARCH

How product complexity affects consumer adoption of new products:

The role of feature heterogeneity and interrelatedness
Andreas Fürst1,3 · Nina Pecornik1 · Wayne D. Hoyer2

Received: 20 February 2020 / Accepted: 9 March 2023

© The Author(s) 2023

Abstract
Recent technological advancements allow companies to incorporate increasingly heterogeneous and interrelated features into
their products, which heightens the products’ complexity. In four experimental studies conducted with two product categories,
this article reveals similarities and differences in terms of how the heterogeneity and interrelatedness of product features influ-
ence consumer attitudes (i.e., expected product usability and capability) and, in turn, purchase intentions. Moreover, it shows that
both neglected dimensions of product complexity affect the corresponding influence of the number of product features but do so
in considerably different ways. The findings suggest that companies can foster consumer adoption by deemphasizing a product’s
feature heterogeneity, thereby avoiding low expected product usability, and by emphasizing its feature interrelatedness, thereby
promoting high expected product capability. This article provides insights into how companies can manage the complexity of
products during both product design (i.e., before market launch) and product advertising and selling (i.e., after market launch).

Keywords Product features · Product complexity · Product configuration · Product design · New product development ·
Consumer adoption · Feature fatigue

To enhance product capability (i.e., ability to perform desired heterogeneous and interrelated product features that further
functions) and thus respond to rising consumer needs and com- increase products’ structural complexity (Gubbi et al., 2013;
petitive pressure, companies have made products such as home Kannan & Li, 2017). For example, today’s smart home sys-
control systems, communication devices, and cars increas- tems offer control of not only housing technology (e.g., blinds)
ingly complex by adding more product features (Bettencourt but also entertainment (e.g., televisions) and household (e.g.,
& Bettencourt, 2011; Fürst & Staritz, 2022). In their quest coffee machines) technologies, and their features can be inter-
to further improve product capability, companies have taken related, such that blinds can be connected with the television
advantage of recent advancements in manufacturing, electron- and the coffee machine. Yet whether these emerging additional
ics, and information technology to implement increasingly levers for increasing a product’s complexity actually enhance
consumers’ perceived product capability or whether they pre-
dominantly reduce consumers’ perceived product usability (i.e.,
Stefan Wuyts served as Area Editor for this article.
ease of learning and using) remains unclear. Because these con-
* Andreas Fürst sumer perceptions are important drivers of consumer adoption
[email protected] of new products, corresponding knowledge is highly manageri-
Nina Pecornik ally relevant (Nysveen et al., 2005; Sääksjärvi & Samiee, 2011).
[email protected] Despite the importance of knowledge about consumer per-
Wayne D. Hoyer ceptions of complex products, previous research on product
[email protected] complexity primarily takes a company perspective by analyz-
1 ing effects of product architecture on internal R&D or manu-
Friedrich-Alexander University Erlangen-Nürnberg, Lange
Gasse 20, 90403 Nuremberg, Germany facturing processes and costs (e.g., Lau et al., 2011; Song
2 et al., 2015) and how best to handle the related complexity,
McCombs School of Business, University of Texas at Austin,
1 University Station, B6700, 2110 Speedway Austin, such as through modular product architectures (e.g., Bonvoisin
TX 78712 Austin, USA et al., 2016; Campagnolo & Camuffo, 2010). The few excep-
3
Business School, University of Eastern Finland (UEF), tions adopting a consumer perspective are limited to a unidi-
Kuopio, Finland mensional conceptualization of product complexity through

13
Vol.:(0123456789)
 Journal of the Academy of Marketing Science

the number of product features (Goodman & Irmak, 2013; We show that not only the number of product features but
Sela & Berger, 2012; Thompson & Norton, 2011; Thompson also the heterogeneity and interrelatedness of these features
et al., 2005). For example, they focus on the impact of fea- influence consumer attitudes and intentions, thereby reveal-
ture number on consumers’ perceived product usefulness and ing that these novel dimensions have discrete and somewhat
choice (Sela & Berger, 2012); perceived product capability, different effects on consumer product adoption. Third, this
usability, and utility (Thompson et al., 2005); perceived social research provides a detailed understanding of the interplay
utility and choice (Thompson & Norton, 2011); or perceived between the various dimensions of product complexity. It
satisfaction and choice (Goodman & Irmak, 2013). These provides evidence that both feature heterogeneity and feature
studies provide important insights into the impact of the num- interrelatedness moderate the downstream effects of number
ber of product features on consumer attitudes and downstream of features but that they do so in fundamentally different
variables. However, the literature is silent on the correspond- ways. We thereby extend prior knowledge by showing that
ing impact of other dimensions that also constitute a product’s a product with the same number of product features can lead
complexity, such as the heterogeneity and interrelatedness of to significantly different consumer attitudes and intentions.
product features. This research gap is lamentable because it Overall, this article explains why marketing researchers’
impedes insights into whether these neglected dimensions of focus should shift from examining the impact of the pure
product complexity favorably or unfavorably affect consumer number of product features on consumer product adoption to
attitudes and downstream variables, as well as whether they also assessing the corresponding impact of the heterogene-
strengthen or weaken the corresponding impact of the number ity and interrelatedness of these features and the interplay
of features. Insights into these issues could help managers between these dimensions of product complexity.
optimize new product design, advertising, and selling, thus
fostering consumer product adoption.
To address this important research gap, this article exam- Development of conceptual framework
ines the effects and interplay of multiple dimensions of a
product’s structural complexity on consumer attitudes and Dimensions of product complexity
intentions, thus representing the first investigation of con-
sumer perceptions of product complexity beyond pure fea- To develop our multidimensional conceptualization of prod-
ture number. For this purpose, we draw on systems theory uct complexity, we draw on systems theory (Von Bertalanffy,
to identify two additional dimensions of product complex- 1968), which research has previously applied to describe,
ity: the heterogeneity of product features (i.e., the extent of for example, the complexity of organizations and their envi-
functional dissimilarity of features) and the interrelatedness ronment (Achrol & Stern, 1988; Duncan, 1972; Thompson,
of product features (i.e., the extent of functional connec- 1967) and multi-channel systems (Fürst & Scholl, 2022).
tivity of features). Using a state-of-practice market analy- This theory suggests that every system (e.g., a product)
sis of existing products for two product categories (smart consists of structural elements (e.g., features) that can be
home systems and smartphones) and four experiments, we characterized according to three criteria – that is, “according
examine whether and how these dimensions affect consumer to their number” but also “according to their species” and
attitudes (expected product capability and usability) and, in “according to the relations of elements” (Von Bertalanffy,
turn, purchase intention, both separately and in conjunction 1968, p. 54). Because a product offers certain functions to
with the number of product features. perform, the species of its elements could, for example, refer
This article extends the literature in several ways. First, to the functional dissimilarity (i.e., heterogeneity) of features
it shows that, analogous to other functional areas such as in type or nature and the relations of its elements, for exam-
R&D and manufacturing, marketing must consider the ple, to the functional connectivity (i.e., interrelatedness) of
downstream effects not only of the number of product fea- features. Thus, systems theory indicates that not only the
tures but also of other dimensions of product complexity, number of features but also the heterogeneity and interre-
particularly the heterogeneity and interrelatedness of prod- latedness of these features determine a product’s structural
uct features. In addition to providing marketing researchers complexity. Therefore, a product with a specific number of
with more explicit levers for new product design, advertis- features that are highly heterogeneous and highly interre-
ing, and selling, these findings contribute to a more holistic lated is characterized by greater structural complexity than
approach to product complexity management that considers a product with the same number of features that are less het-
both the impact of multiple complexity dimensions on inter- erogeneous and less interrelated. Consequently, according to
nal processes and costs and the impact of these dimensions systems theory, prior studies’ singular focus on the number
on consumer product adoption. Second, this article offers a of product features (e.g., Sela & Berger, 2012; Thompson
thorough understanding of the impact of the various dimen- et al., 2005) does not fully capture the construct of product
sions of product complexity on consumer product adoption. complexity.

13
Journal of the Academy of Marketing Science

The heterogeneity of product features reflects the dissimi- Consumer attitudes

larity of structural elements inherent in a product to perform
specific functions (Duncan, 1972; Lee & Chu, 2021; Thomp- To select the consumer attitudes for our framework, we draw
son, 1967). Studies on the addition of a new feature to an on a benefit–cost logic (Johnson & Payne, 1985; Payne, 1982),
existing product indicate that a feature can be characterized which suggests that consumers assess a product depending
as more or less similar to the base product in terms of the on their benefit- and cost-related needs (De Angelis & Car-
benefits and value it offers (Gattol et al., 2016; Lee & Chu, penter, 2009). According to this logic, the beliefs resulting
2021). Therefore, in the context of our research, the heteroge- from product assessment refer to benefits, such as consumers’
neity of product features can be described by the overall dis- expected gain from the product, and costs, such as consumers’
similarity of features in terms of the purpose they serve. For expected effort, time, and energy for learning about and using
example, in a smart home system, the features of “floor heat- the product (Mukherjee & Hoyer, 2001). Thus, consistent with
ing,” “fridge,” and “television” are very dissimilar and, thus, the selection of similar pairs of constructs by previous research
highly heterogeneous, as they refer to the control of housing, on consumer perceptions of product complexity (Thompson
household, and entertainment technologies, respectively. By et al., 2005) and technology acceptance (Blut et al., 2016), our
contrast, “floor heating,” “ventilation,” and “air conditioning” framework focuses on product capability (a consumer benefit
all refer to housing technology, which makes them consider- from the product) and product usability (a consumer cost of
ably less dissimilar and, thus, less heterogeneous. learning about and using the product). Because our focus is
The interrelatedness of product features captures the con- on products new to the consumer, these constructs relate to
nectivity of structural elements inherent in a product to per- expectations rather than actual experiences. Expected prod-
form specific functions (Duncan, 1972; Pinochet et al., 2018; uct capability refers to consumers’ beliefs about the product’s
Thompson, 1967). Similar to components of product offerings ability to perform desired tasks (Thompson & Norton, 2011;
that can be linked together to enable them to communicate Thompson et al., 2005). Expected product usability reflects
and work with each other (Chang et al., 2014; Pinochet et al., consumers’ beliefs about the ease of learning and using the
2018), we assume that the features of a product can be con- product (Davis, 1989; Thompson et al., 2005). Typically, con-
nected, which allows them to exchange data, and thus their sumers view a product’s capability and usability as compet-
use can be combined. The functional connectivity of a prod- ing needs when forming their purchase decisions (Blut et al.,
uct’s totality of features refers to the proportion of connections 2016; Thompson et al., 2005). For example, Davis (1989, p.
between features. Based on the concept of network density 320) argue that “even if potential users believe that a given
(Gupta et al., 2019; Scott, 1988), it is determined by the actual application is useful, they may, at the same time, believe that
quantity of connections between features compared with the the system is too hard to use and that the performance benefits
maximum quantity of connections between features. For of usage are outweighed by the effort of using the application.”
example, in a smart home system, the features of “television,” Thus, their choice represents the result of a “cognitive trade-off
“multiroom audio,” and “blinds” could all be connected, such between the effort required … (i.e., ease of use) and the quality
that if “television” is used, “multiroom audio” is automatically (i.e., usefulness) of the resulting [choice] decision” (Kim et al.,
activated (or deactivated) and “blinds” go down, representing 2007, p. 115). To ensure new product adoption, examining
a high interrelatedness of features. Conversely, these features whether feature heterogeneity and interrelatedness enhance
could also lack any corresponding connections because of, consumers’ expected product capability or whether they pre-
for example, missing technical interfaces or compatibilities, dominantly reduce consumers’ perceived product usability is
which would indicate low interrelatedness of features. therefore highly managerially relevant (Nysveen et al., 2005;
Previous research has not investigated the heterogeneity Sääksjärvi & Samiee, 2011).
and interrelatedness of product features, which is unfortu-
nate because examining them in addition to and together Consumer intentions
with the number of product features would allow for a more
comprehensive conceptualization of product complexity and Our framework includes product purchase intention as the
thus provide a fuller picture of levers for product design, primary downstream variable. This construct refers to the
advertising, and selling. In this context, it is important to extent to which consumers are willing to buy a product
note that all three dimensions are characteristic of complex (Dodds et al., 1991).
products, as they derive from systems theory, which deals
with the structural aspects of complexity, and they are dis- Overview of framework
tinct because a product can, for example, have features that
are high in heterogeneity but either high or low in interre- Figure 1 shows our framework, including the hypothesized
latedness. Herein, we empirically verify the distinctiveness effects. Our focal hypotheses relate to the main and moderat-
of these dimensions. ing effects of the two novel dimensions of product complexity

13
 Journal of the Academy of Marketing Science

Fig. 1  Conceptual framework. a

Alternatively: Expected product Dimensions of Consumer Consumer
utility or product purchase Product Complexity Attitudes Intentions
behavior

Number of
Product Features

,b
5a
H3a,b

/H
a,b
H4
Heterogeneity of Expected
Product Features H1a / H2a Product Capability H6a / H6b
Product
Purchase Intention a
Interrelatedness of H1b / H2b Expected
Product Features Product Usability

(i.e., heterogeneity and interrelatedness of product features). for consumers. In line with this reasoning, previous research
In terms of main effects, we predict that both dimensions suggests that dissimilar features may enact more different roles
increase expected product capability and decrease expected in a product than similar ones and thus are able to deliver more
product usability, both of which are assumed to promote prod- value (Gibbert & Mazursky, 2009; Kuehnl et al., 2017; Wilk-
uct purchase intention. In Study 1a (smart home context) and, enfeld & Ward, 2001). Therefore, consumers are likely to infer
for replication purposes, Study 1b (smartphone context), we a broad range of functions from highly heterogeneous features
concentrate on the direct main effects of these dimensions on and, thus, high ability of the product to perform desired tasks,
consumer attitudes (H1a/b, H2a/b) while holding constant the which enhances their expected product capability.
number of product features. In terms of moderating effects, However, high heterogeneity of product features also means
we predict that both heterogeneity and interrelatedness of that consumers are less likely to assign the features to the same
product features will increase the positive impact of number group and more likely to consider them separate entities, which
of product features on expected product capability and the requires significant time and energy to engage with the product
negative impact of number of product features on expected (Mukherjee & Hoyer, 2001; Shugan, 1980). For example, a
product usability. Study 2a (smart home context) and, for rep- smart home system that allows control of not only housing
lication purposes, Study 2b (smartphone context) focus on technology but also household technology and entertainment
testing these moderating effects (H4a/b, H5a/b) and related technology offers very different types of functions that may
main effects (H3a/b), thereby manipulating not only hetero- entail high expected costs for customers to learn and use the
geneity and interrelatedness but also the number of product product. In support of this argumentation, previous research
features. In addition, these studies validate Studies 1a’s and suggests that consumers have more difficulty categorizing
1b’s findings related to the direct main effects of the two novel and making sense of dissimilar product features than similar
dimensions (H1a/b, H2a/b) and examine their corresponding ones (Gibbert & Mazursky, 2009; Wilkenfeld & Ward, 2001).
downstream effects on consumer purchase intentions (H6a/b). Therefore, highly heterogeneous features tend to complicate
processing, which makes consumers anticipate considerable
costs of learning and using a product, thereby decreasing their
Hypotheses development expected product usability. Thus, we predict the following:

Effects of heterogeneity of product features H1 The heterogeneity of product features (a) increases
expected product capability and (b) decreases expected
High heterogeneity of product features indicates to consumers product usability.
that the product can perform a wide variety of functions. For
example, a smart home system that controls not only supply Effects of interrelatedness of product features
functions and related housing technology but also housework
functions and related household technology, as well as enjoy- High interrelatedness of product features enables the
ment functions and related entertainment technology, offers a exchange of data between features and, thus, their combined
broad range of functions and, thus, increased expected benefits use, which allows consumers to take advantage of functions

13
Journal of the Academy of Marketing Science

that would not be feasible otherwise. For example, a smart Therefore, consumers are likely to have stronger beliefs that
home system in which the features of “television,” “mul- a feature-rich product is able to perform desired tasks better
tiroom audio,” and “blinds” are connected enables con- than a feature-poor product, resulting in a higher expected
sumers using “television” to turn on (or off) “multiroom product capability.
audio” automatically and have “blinds” go down, leading However, a high number of features and the resulting mul-
to enhanced functions and, thus, greater expected benefits. titude of functions indicates to consumers that the product
Consistent with this reasoning, previous research suggested will have a large number of decision and usage options for
that products that are connected with each other in a network which they must invest considerable effort, time, and energy
are able to carry out additional tasks that are more com- to understand and make full use of the product (Bettman
prehensive and sophisticated, thus providing added value et al., 1991; Iyengar & Lepper, 2000; Meyer et al., 2008).
(Chang et al., 2014; Hoffman & Novak, 2018; Raff et al., This argumentation is consistent with prior research showing
2020). Consequently, consumers are likely to anticipate that a high quantity of functions is associated with a large
enhanced functions from highly interrelated features, which number of decision and usage options that decrease con-
fosters their belief about the product’s capability. sumers’ ability to understand and use the product (Ji et al.,
However, consumers may also perceive high interrelat- 2006; Preece et al., 2002), thereby causing them to infer
edness of product features as offering an almost incompre- high learning costs and develop fears of erroneous product
hensible multitude of possible connections between features use (Goodman & Irmak, 2013; Meyer et al., 2008). Thus, a
they could activate when setting up the product or apply feature-rich product likely overwhelms consumers, which
when using the product, which likely requires considerable can lead them to assume considerable difficulty of learn-
time and energy to understand the product and make use of ing and using the product and thus decrease their expected
its full potential (Mukherjee & Hoyer, 2001; Shugan, 1980). product usability. Therefore, building on Thompson et al.
For example, a smart home system with 10 highly interre- (2005), we predict the following:
lated features could easily offer 15–20 connections between
features to consider, resulting in significant expected costs H3 The number of product features (a) increases expected prod-
for customers to learn and use the product. Previous research uct capability and (b) decreases expected product usability.
indicates that “as the number of [connections] … increases,
the memory structure for [the product] … becomes richer, Previously, we argued that the number of features drives
but also more complex” (Krishnan, 1996, p. 392), and the consumers’ perceived quantity of functions that a product
more connections between features to take into account, the provides, which increases expected product capability and
higher are consumers’ inferences about costs of learning and reduces expected product usability (Bertini et al., 2009;
using the product (Buescher et al., 2009; Meyer et al., 2008). Meyer et al., 2008). Subsequently, we argue that the number
Consequently, highly interrelated features likely complicate of features drives consumers’ perceived quantity of func-
processing, which causes consumers to anticipate significant tions to varying degrees depending on the heterogeneity and
costs of learning and using the product, thereby reducing interrelatedness of these features, which ultimately affects
their expected product usability. Thus: the strength of the impact of feature number on expected
product capability and usability.
H2 The interrelatedness of product features (a) increases In terms of the moderating effect of feature heterogeneity,
expected product capability and (b) decreases expected prior research suggests that dissimilar features tend to have
product usability. different functional roles in a product whereas similar features
tend to compete for the same functional roles (Estes, 2003;
Effects of number of product features Gibbert & Mazursky, 2009; Wilkenfeld & Ward, 2001). Thus,
when increasing the number of dissimilar features by a certain
Although the main effects of feature number on consumer amount, this increase tends to result in less overlap in func-
attitudes are not the focus of our study (for a previous study tions of the product than when increasing the number of simi-
on feature number, see Thompson et al., 2005), we present a lar features by the same amount, leading to a stronger growth
brief reasoning for corresponding hypotheses, which serves in consumers’ perceived quantity of functions. This stronger
as a baseline for the moderating effects of our two dimen- growth in the perceived quantity of functions is, in turn, likely
sions of interest on these main effects. A high number of to lead to a stronger increase in consumers’ expectation of
features signals to consumers that the product can perform a the product’s ability to perform desired tasks and their deci-
multitude of functions. In support of this reasoning, previous sion and usage effort and, thus, time and energy to understand
research indicates that consumers typically infer the quantity and make use of the product (De Angelis & Carpenter, 2009;
of functions and the related benefits from the features of a Goodman & Irmak, 2013). Thus, when feature heterogeneity
product (Bertini et al., 2009; Nowlis & Simonson, 1996). is high (vs. low), increasing the number of features tends to

13
 Journal of the Academy of Marketing Science

disproportionately enhance expected product capability and differ significantly in their complexity. This differentiation
reduce expected product usability. is in line with suppliers that classify products of these types
In terms of the moderating effect of feature interrelat- as either basic or complex feature systems. We carefully
edness, previous studies indicate that the more interrelated scrutinized product manuals, created a list of features for
the features of a product, the more connections they have, each product, and then analyzed the features in terms of their
which allows them to better interact with each other and number, heterogeneity, and interrelatedness.
work together, resulting in more functions (Estes, 2003;
Ng & Wakenshaw, 2017; Pinochet et al., 2018). Therefore, Stimuli for number of product features
increasing the number of highly interrelated features by a
certain amount tends to create more additional connections Drawing on prior studies’ operationalizations (Thompson
in the product than increasing the number of less interrelated & Norton, 2011; Thompson et al., 2005), our analyses of
features by the same amount, resulting in a stronger growth the manuals, and a pretest, we determined the number of
in consumers’ perceived quantity of functions. The stronger features to be 5 for the “low number of features” condition,
growth in the perceived quantity of functions is, in turn, 10 for the “medium number of features” condition, and 15
likely to result in a stronger increase in consumers’ expec- for the “high number of features” condition.
tation of the product’s ability to perform desired tasks and
their costs of learning and using the product (De Angelis & Stimuli for heterogeneity of product features
Carpenter, 2009; Goodman & Irmak, 2013). Consequently,
when feature interrelatedness is high (vs. low), increasing To develop the “low heterogeneity of features” and “high
the number of features tends to disproportionately increase heterogeneity of features” conditions, we performed sev-
expected product capability and decrease expected product eral steps. First, we searched for an appropriate criterion
usability. Overall, we predict the following: to determine product feature heterogeneity. Our analyses
of the manuals showed that products differ especially in
H4 When the heterogeneity of product features is high rather the extent to which their features belong to the same func-
than low, the number of product features (a) increases tional category or to different functional categories. In the
expected product capability more strongly and (b) case of smart home systems, the features were primarily
decreases expected product usability more strongly. related to the control of housing technology, though some
were also related to the control of household and entertain-
H5 When the interrelatedness of product features is high ment technology. For smartphones, the features mainly
rather than low, the number of product features (a) belonged to communication, though some also belonged
increases expected product capability more strongly and to organization and entertainment. Second, we designed
(b) decreases expected product usability more strongly. the “low heterogeneity of features” condition to consist of
features of the same functional category serving a similar
Effects of consumer attitudes purpose (i.e., control of housing technology in the smart
home context and communication in the smartphone con-
Extant research indicates that consumers consider the costs text), while the “high heterogeneity of features” condition
and benefits associated with a product when forming their included features from three functional categories serv-
intentions (Johnson & Payne, 1985; Payne, 1982). Therefore, ing different purposes (i.e., control of housing technology,
and supported by prior research showing that anticipated household technology, and entertainment technology in
product usability and capability enhance expected utility the smart home context and communication, organization,
(Thompson et al., 2005) and intention to use (Nysveen et al., and entertainment in the smartphone context). Third, we
2005), we predict the following: drew on Blau’s (1977) index of heterogeneity to verify the
extent of heterogeneity for both conditions (see part A1
H6 Product purchase intention is increased by (a) expected in Appendix 1).
product capability and (b) expected product usability.
Stimuli for interrelatedness of product features

Development of stimuli To determine the specific quantity of connections between

features for “low interrelatedness of features” and “high
To ensure realistic descriptions in our scenarios, we con- interrelatedness of features,” we used a multistep approach
ducted an extensive state-of-practice market analysis. Our that draws on the concept of network density (Gupta et al.,
sample consisted of smart home systems and smartphones 2019; Scott, 1988). First, we calculated the maximum
that reflect the range of products available on the market and quantity of connections between features cmax for the three

13
Journal of the Academy of Marketing Science

conditions of number of product features (see part A2 in female, M age = 44.80 years; SP: n = 195; 50% female,
Appendix 1). Second, we aimed to find recommendations Mage = 44.90 years) from a representative access panel of
from the literature for determining the proportion of the a professional market research institute (Respondi) to a 2
maximum quantity of connections pci that should represent (heterogeneity of features: low vs. high) × 2 (interrelatedness
a “low interrelatedness of features” and “high interrelated- of features: low vs. high) between-subjects design. Partici-
ness of features,” respectively. Yet work on network analysis pants received points from the panel provider as incentives.
emphasizes the absence of any universal threshold values or Drawing on our stimulus development and a successful
rule of thumb, because “the assessment of the actual density pretest of our manipulations (see Appendix 2, part B1), we
[i.e., interrelatedness of elements] … must take account of manipulated all independent variables (i.e., the two novel
the size of the network [i.e., number of elements]” (Scott, dimensions of product complexity) and held the number of
1988, p. 115). Specifically, in a network with a large number features constant on a medium level. The latter was neces-
of elements, a high proportion of connections between ele- sary to analyze the mere effects of heterogeneity and inter-
ments is typically less feasible or useful than in a network relatedness of features.
with a small number of elements (Scott, 1988). Therefore,
studies on network analysis recommend that for a larger Procedure and measures We conducted a computer-based
number of elements, it is best to rely on lower values as an experiment. First, participants answered questions related to
indicator of high density than for a smaller number of ele- demographics (e.g., age, gender, educational level). Second,
ments, such that even “a density of 0.29 may, under certain we introduced the session as a study on purchase decisions
circumstances, indicate a very high level of cohesion [i.e., and asked participants to imagine that they were interested
connectivity] in the network” (Scott, 1988, p. 115). Moreo- in purchasing a new smart home system (in Study 1a) or
ver, even with an objectively low density of, for example, smartphone (in Study 1b). Subsequently, as shown in parts
0.10, a network with a larger number of elements may be C1, C3, and C4 of the Supplemental Material, we presented
perceived as highly cohesive because of the relatively large the stimuli including an initial description that explained the
absolute quantity of connections. Thus, in the context of our underlying logic of the stimuli. To manipulate the product
study, the literature recommends that products with a larger complexity dimensions, we did not mention their labels (het-
number of features should generally apply lower proportion erogeneity and interrelatedness) but primarily relied on visual
values of the maximum quantity of connections pci as an elements (shapes, colors, and lines with arrows) combined
indicator of high and low interrelatedness than products with with a verbal listing of the features and their categories. Spe-
a smaller number of features (Bandyopadhyay et al., 2010). cifically, to manipulate “heterogeneity,” the features of each
Third, following these recommendations and supported by category (e.g., in the smart home context: housing technol-
our analysis of product manuals and a pretest, we discounted ogy features, household features, entertainment features) had
the proportion value pci for the “high number of features” a unique shape and color and were grouped together in the
condition by the factor 2 compared with the “low number form of a list on the upper part of the respective stimulus.
of features” condition (see part A2 in Appendix 1). Finally, To manipulate “interrelatedness,” the corresponding features
we primarily drew on our analysis of product manuals and were linked by a line with double-sided arrows. Depending
a short pretest (n = 22; 50% female, Mage = 31.91 years) to on the scenario, the presented product varied in heteroge-
determine which specific features to present in our scenarios neity and interrelatedness of features, while the number of
as functionally connected. features (medium level) remained constant. Third, we asked
questions related to the manipulation check, psychographics
(consumer expertise and, in Study 1a, trust), and consumer
Studies 1a and 1b: Main effects of novel attitudes (expected product capability and usability). Part B2
dimensions of product complexity in Appendix 2 provides an overview of the scales and their
reliability; part B3 shows that corresponding checks provided
sufficient support for our manipulations.
Method

Results
Goals, design, and participants To examine the impact
of heterogeneity and interrelatedness of product features We performed a 2 × 2 multivariate analysis of covariance
on consumer attitudes (H1a/b, H2a/b), we conducted (MANCOVA) with heterogeneity and interrelatedness of
Study 1a in a smart home context and, for replication pur- features as independent variables and expected product capa-
poses, Study 1b in a smartphone context. In both stud- bility and usability as dependent variables. As controls, we
ies, we randomly assigned consumers (SH: n = 240; 49% included demographics (age, gender, and educational level)

13
 Journal of the Academy of Marketing Science

and, following Thompson et al. (2005), consumer expertise. Discussion

We found no support for H1a that heterogeneity of features
increases expected product capability (SH: Mlow hete = 4.48, Studies 1a and 1b showed that both feature heterogeneity and
SD = 1.60 vs. Mhigh hete = 4.15, SD = 1.78; F(1, 231) = 2.67, interrelatedness have a discrete impact on consumer attitudes.
p = 0.10; SP: Mlow hete = 4.38, SD = 1.51 vs. Mhigh hete = 4.20, In this context, we found similarities and differences in how
SD = 1.34; F(1, 186) = 1.05, p = 0.31). However, the data these novel dimensions of product complexity affect expected
showed full support for the predicted negative effect of het- product capability and usability. While both dimensions
erogeneity of features on expected product usability (SH: decrease expected product usability, only feature interrelated-
Mlow hete = 4.48, SD = 1.50 vs. Mhigh hete = 4.03, SD = 1.77; ness increases expected product capability. A possible explana-
F(1, 231) = 6.09, p < 0.05; SP: Mlow hete = 5.71, SD = 1.21 vs. tion for the finding that feature heterogeneity does not increase
Mhigh hete = 5.25, SD = 1.18; F(1, 186) = 8.13, p < 0.05), con- expected product capability is that consumers lack trust that
firming H1b. products offered to them with a high dissimilarity of features
In addition, the results provided full support that interrelat- will actually be able to provide the proposed capability. To test
edness of features increases expected product capability (SH: this assumption, we estimated a mediation model based on
Mlow inter = 4.08, SD = 1.63 vs. Mhigh inter = 4.55, SD = 1.71; 10,000 bootstrap samples (Hayes, 2018, PROCESS Model 4).
F(1, 231) = 4.81, p < 0.05; SP: Mlow inter = 3.87, SD = 1.42 In support of our assumption, the results confirmed that feature
vs. Mhigh inter = 4.72, SD = 1.21; F(1, 186) = 20.01, p < 0.01). heterogeneity decreases consumer trust and, in turn, expected
Moreover, we found that interrelatedness of features decreases product capability (indirect effect: b =  − 0.39, SE (boot) = 0.16,
expected product usability (SH: Mlow inter = 4.51, SD = 1.45 95% CI = [− 0.71; − 0.08]), which mitigates any positive effects
vs. Mhigh inter = 4.00, SD = 1.82; F(1, 231) = 6.80, p < 0.05; of feature heterogeneity on expected product capability.
SP: Mlow inter = 5.66, SD = 1.06 vs. Mhigh inter = 5.30, SD = 1.36; Overall, studies 1a and 1b highlight the importance of
F(1, 186) = 4.48, p < 0.05). Thus, the data confirmed H2a and extending the previously unidimensional conceptualization of
H2b. For both product categories, the interaction effects of product complexity to a multidimensional view that includes
heterogeneity and interrelatedness on expected capability and not only number of product features but also their heteroge-
expected usability, respectively, were nonsignificant. Figure 2 neity and interrelatedness. Studies 2a and 2b draw on this
provides an overview of the effects of the two novel dimen- three-dimensional conceptualization of product complexity to
sions of product complexity on consumer attitudes. develop a detailed understanding of the impact and interplay
of the two novel dimensions and the well-known dimension.

Fig. 2  Main effects of novel

dimensions of product complex- (H1a) (H1b)
p < .05
Expected Product Capability

Expected Product Usability

ity. hete = heterogeneity of

features; inter = interrelatedness p = .10 p = .31
p < .05 5.71
5.25
of features; number of features
was held constant 4.48 4.38 4.48
4.15 4.20
4.03

hete = Home
Smart hete = hete = hete =
Smartphone hete = Home
Smart hete = hete = hete =
Smartphone
low high low high low high low high
Smart Home Smartphone Smart Home Smartphone

(H2b)
(H2a)
p < .05
Expected Product Capability

Expected Product Usability

p < .05 p < .01 p < .05 5.66

5.30
4.72
4.55 4.51
4.08 4.00
3.87

inter = Home
Smart inter = inter = inter =
Smartphone inter = Home
Smart inter = inter = inter =
Smartphone
low high low high low high low high
Smart Home Smartphone Smart Home Smartphone

13
Journal of the Academy of Marketing Science

Studies 2a and 2b: Moderating effects SD = 1.51; F(1, 437) = 0.58, p = 0.45). Also consistent with
of novel dimensions of product complexity Studies 1a and 1b, we found ample evidence that feature het-
erogeneity decreases product usability (SH: Mlow hete = 4.45,
Method SD = 1.65 vs. Mhigh hete = 4.17, SD = 1.63; F(1, 427) = 4.51,
p < 0.05; SP: Mlow hete = 5.42, SD = 1.28 vs. Mhigh hete = 5.17,
Goals, design, and participants First, Studies 2a (smart home SD = 1.48; F(1, 437) = 5.37, p < 0.05), in support of H1b.
context) and 2b (smartphone context) aimed to validate the In addition, the data provided full support that feature inter-
findings of Studies 1a and 1b related to the direct main effects relatedness increases product capability (SH: Mlow inter = 4.06,
of heterogeneity and interrelatedness of features on consumer SD = 1.58 vs. Mhigh inter = 4.63, SD = 1.57; F(1, 427) = 16.28,
attitudes (H1a/b, H2a/b) and to analyze related downstream p < 0.01; SP: Mlow inter = 3.66, SD = 1.52 vs. Mhigh inter = 4.59,
effects on consumer intentions (H6a/b). Second, and more SD = 1.35; F(1, 437) = 50.18, p < 0.01). Moreover, we found
important, we performed these studies to investigate the mod- that feature interrelatedness decreases product usability (SH:
erating effects of heterogeneity and interrelatedness of fea- Mlow inter = 4.53, SD = 1.60 vs. Mhigh inter = 4.09, SD = 1.67;
tures (H4a/b, H5a/b) on the corresponding impact of number F(1, 427) = 10.69, p < 0.01; SP: Mlow inter = 5.43, SD = 1.32 vs.
of features (H3a/b). From a representative access panel of Mhigh inter = 5.15, SD = 1.42; F(1, 437) = 7.00, p < 0.05). Thus,
professional market research institute (Respondi), we obtained consistent with Studies 1a and 1b, H2a and H2b were confirmed.
participants (SH: n = 439; 52% female, Mage = 44.59 years; As predicted, we also found ample evidence that feature
SP: n = 449; 53% female, Mage = 45.60 years) and randomly number increases product capability (SH: Mlow num = 4.12,
assigned them to a 2 (number of features: low vs. high) × 2 SD = 1.65 vs. Mhigh num = 4.57, SD = 1.51; F(1, 427) = 10.10,
(heterogeneity of features: low vs. high) × 2 (interrelated- p < 0.01; SP: Mlow num = 3.76, SD = 1.53 vs. Mhigh num = 4.48,
ness of features: low vs. high) between-subjects design. As in SD = 1.42; F(1, 437) = 31.17, p < 0.01). Moreover, the data
Studies 1a and 1b, participants received points from the panel show that feature number decreases product usability (SH:
provider as incentives. Relying on our stimulus development Mlow num = 4.77, SD = 1.62 vs. Mhigh num = 3.84, SD = 1.55;
and a successful pretest of our manipulations (see Appendix F(1, 427) = 48.28, p < 0.01; SP: Mlow num = 5.74, SD = 1.07
3, part C1), we manipulated all independent variables (i.e., the vs. Mhigh num = 4.84, SD = 1.50; F(1, 437) = 69.63, p < 0.01).
three dimensions of product complexity). Thus, H3a and H3b were supported.
Moreover, we tested for the hypothesized moderating
Procedure and measures We measured some demograph- effects. Figures 3 and 4 provide a graphical overview of our
ics (e.g., age, gender, educational level) and then used the results related to these moderating effects. For product capa-
same experimental procedure as in Studies 1a and 1b. After bility, the results did not provide evidence for H4a, which
presenting the stimuli (see parts C2, C5, and C6 of the predicted an interaction between feature number and hetero-
Supplemental Material), we asked questions related to the geneity (SH: F(1, 427) = 0.35, p = 0.56; SP: F(1, 437) = 0.54,
manipulation check, psychographics (consumer expertise p = 0.46). By contrast, we found evidence for H4b, which
and, in Study 2a, trust), consumer attitudes (expected prod- proposed an interaction effect of feature number and inter-
uct capability and usability), and product purchase intention. relatedness on product capability (SH: F(1, 427) = 5.47,
In Appendix 3, part C2, we show details on the scales and p < 0.05; SP: F(1, 437) = 5.01, p < 0.05). Specifically, as
their reliability. Moreover, part C3 provides an overview of predicted, feature number increases product capability
our manipulation checks, which provided sufficient support more strongly when feature interrelatedness is high (SH:
for the appropriateness of our stimuli. Mlow num = 4.24, SD = 1.72 vs. Mhigh num = 5.02, SD = 1.28;
F(1, 427) = 14.97, p < 0.01; SP: Mlow num = 4.08, SD = 1.40 vs.
Mhigh num = 5.09, SD = 1.11; F(1, 437) = 30.04, p < 0.01) rather
Results than low (SH: Mlow num = 4.00, SD = 1.58 vs. Mhigh num = 4.12,
SD = 1.58; F(1, 427) = 0.36, p = 0.55; SP: Mlow num = 3.44,
Test of hypotheses We conducted a 2 × 2 × 2 MANCOVA with SD = 1.57 vs. Mhigh num = 3.87, SD = 1.43; F(1, 437) = 5.62,
number, heterogeneity, and interrelatedness of features as inde- p < 0.05). In addition, we found an interaction between het-
pendent variables and age, gender, educational level, and con- erogeneity and interrelatedness in the smartphone context
sumer expertise as controls on expected product capability and (SP: F(1, 437) = 7.34, p < 0.05), but not in the smart home
usability. For the hypothesized main effects and consistent with context (SH: F(1, 427) = 0.07, p = 0.79), and no evidence of
Studies 1a and 1b, we found no support for H1a that feature a three-way-interaction (SH: F(1, 427) = 1.79, p = 0.18; SP:
heterogeneity increases product capability (SH: Mlow hete = 4.48, F(1, 437) = 0.55, p = 0.46).
SD = 1.62 vs. Mhigh hete = 4.21, SD = 1.56; F(1, 427) = 3.66, For product usability, we found a significant interac-
p = 0.06; SP: Mlow hete = 4.07, SD = 1.53 vs. Mhigh hete = 4.17, tion between feature number and heterogeneity (SH: F(1,

13
 Journal of the Academy of Marketing Science

Fig. 3  Moderating effects of (H4a)

novel dimensions of product
complexity on expected product Smart Home Smartphone
capability. hete = heterogeneity hete = low hete = high hete = low hete = high

Expected Product Capability

Expected Product Capability

of features; inter = interrelated-
p < .05 p < .01 p < .01
ness of features; num = number p = .07
of features 4.75 4.48 4.49
4.22 4.40
4.03 3.86
3.66

num = num =
Low Heterogeneity numHeterogeneity
High = num = num = num =
Low Heterogeneity numHeterogeneity
High = num =
low high low high low high low high

(H4b)
Smart Home Smartphone
inter = low inter = high inter = low inter = high
Expected Product Capability

Expected Product Capability

p < .01 p < .01
p = .55
5.02 p < .05 5.09

4.00 4.12 4.24 3.87 4.08

3.44

Low Interrelatedness
num = num = High
numInterrelatedness
= num = Low
numInterrelatedness
= num = High
numInterrelatedness
= num =
low high low high low high low high

Fig. 4  Moderating effects of

(H5a)
novel dimensions of product Smart Home Smartphone
complexity on expected product
hete = low hete = high hete = low hete = high
usability. hete = heterogeneity of
p < .01 p < .01
features; inter = interrelatedness
Expected Product Usability
Expected Product Usability

of features: num = number of p < .01 p < .01 5.74 5.74

features 4.78 4.77
5.09
4.59
4.13
3.56

Low
numHeterogeneity
= num = High
numHeterogeneity
= num = Low Heterogeneity
num = num = High
numHeterogeneity
= num =
low high low high low high low high

(H5b)
Smart Home Smartphone
inter = low inter = high inter = low inter = high
p < .01 p < .01
Expected Product Usability
Expected Product Usability

p < .01
5.77 5.71
p < .01
5.14 5.10
4.41 4.59
3.92 3.77

Low
numInterrelatedness
= num = High
num Interrelatedness
= num = Low
numInterrelatedness
= num = High
numInterrelatedness
= num =
low high low high low high low high

13
Journal of the Academy of Marketing Science

427) = 4.36, p < 0.05; SP: F(1, 437) = 5.18, p < 0.05). Specifi- negative total effect (SH: b =  − 0.20; 95% CI = [− 0.40
cally, as predicted in H5a, feature number decreases product to − 0.01]; SP: b = 0.02; 95% CI = [− 0.21 to 0.24]). Moreover,
usability more strongly when feature heterogeneity is high consistent with our results related to H2a and H2b, we found
(SH: Mlow num = 4.77, SD = 1.57 vs. Mhigh num = 3.56, SD = 1.42; a positive indirect effect of interrelatedness through capabil-
F(1, 427) = 40.47, p < 0.01; SP: Mlow num = 5.74, SD = 1.02 vs. ity (SH: b = 0.36; 95% CI = [0.19 to 0.55]; SP: b = 0.79; 95%
Mhigh num = 4.59, SD = 1.54; F(1, 437) = 53.86, p < 0.01) rather CI = [0.57 to 1.02]) and a negative indirect effect through
than low (SH: Mlow num = 4.78, SD = 1.68 vs. Mhigh num = 4.13, usability (SH: b =  − 0.05; 95% CI = [− 0.10 to − 0.01]; SP:
SD = 1.58; F(1, 427) = 11.82, p < 0.01; SP: Mlow num = 5.74, b =  − 0.07; 95% CI = [− 0.15 to − 0.02]), accompanied by a
SD = 1.11 vs. Mhigh num = 5.09, SD = 1.38; F(1, 437) = 19.01, positive total effect (SH: b = 0.31; 95% CI = [0.12 to 0.52];
p < 0.01). Moreover, we found support for H5b, proposing SP: b = 0.71; 95% CI = [0.49 to 0.95]). Finally, consistent
an interaction effect of feature number and interrelatedness with the results related to H3a and H3b, we found a posi-
on product usability (SH: F(1, 427) = 4.64, p < 0.05; SP: F(1, tive indirect effect of feature number through capability
437) = 4.47, p < 0.05). As expected, in the smartphone con- (SH: b = 0.29; 95% CI = [0.11 to 0.47]; SP: b = 0.60; 95%
text, feature number decreases product usability more strongly CI = [0.38 to 0.83]) and a negative indirect effect through
when feature interrelatedness is high (SP: Mlow num = 5.71, usability (SH: b =  − 0.11; 95% CI = [− 0.20 to − 0.03]; SP:
SD = 0.92 vs. Mhigh num = 4.59, SD = 1.54; F(1, 437) = 53.68, b = –0.22; 95% CI = [− 0.35 to − 0.11]), together with a par-
p < 0.01) rather than low (SP: Mlow num = 5.77, SD = 1.17 vs. tially significant, positive total effect (SH: b = 0.18; 95%
Mhigh num = 5.10, SD = 1.38; F(1, 437) = 19.57, p < 0.01). By CI = [− 0.04 to 0.39]; SP: b = 0.38; 95% CI = [0.13 to 0.64]).
contrast, in the smart home context, feature number decreases To take into account the nature of the proposed inter-
product usability less strongly when interrelatedness is actions of feature heterogeneity and interrelatedness with
high (SH: Mlow num = 4.41, SD = 1.74 vs. Mhigh num = 3.77, feature number, we estimated a moderated mediation model
SD = 1.54; F(1, 427) = 11.29, p < 0.01) rather than low (SH: (Hayes, 2018, PROCESS Model 7), with feature number as
Mlow num = 5.14, SD = 1.41 vs. Mhigh num = 3.92, SD = 1.56; F(1, the independent variable, feature heterogeneity and inter-
427) = 42.10, p < 0.01). Thus, overall, we found partial sup- relatedness as moderators, expected product capability and
port for H5b. In addition, we found that neither the interaction usability as mediators, and purchase intention as the depend-
between feature heterogeneity and interrelatedness (SH: F(1, ent variable. Again, we controlled for potential effects of
427) = 0.29, p = 0.59; SP: F(1, 437) = 0.41, p = 0.53) nor the age, gender, educational level, and consumer expertise. Con-
three-way interaction was significant (SH: F(1, 427) = 0.32, sistent with the results related to H4a and H5a, feature heter-
p = 0.57; SP: F(1, 437) = 1.52, p = 0.22). ogeneity did not affect the mediation through capability (SH:
Finally, we performed a regression analysis to test the hypoth- index of moderated mediation =  − 0.14; 95% CI = [− 0.52
esized effects of expected product capability and usability on to 0.23]; SP: index of moderated mediation =  − 0.18; 95%
product purchase intention, again controlling for demographics CI = [− 0.62 to 0.24]), but it enhanced the negative effect of
(age, gender, and educational level) and consumer expertise. Our feature number on purchase intention through usability (SH:
findings indicated that product purchase intention is increased index of moderated mediation =  − 0.22; 95% CI = [− 0.42
by both product capability (SH: β = 0.56, p < 0.01; SP: β = 0.65, to − 0.02]; SP: index of moderated mediation =  − 0.14; 95%
p < 0.01) and usability (SH: β = 0.11, p < 0.05; SP: β = 0.18, CI = [− 0.30 to − 0.02]). In addition, consistent with our
p < 0.01), providing support for H6a and H6b. results related to H4b, feature interrelatedness enhanced
the positive effect of feature number on purchase inten-
Test of mediation on purchase intention We estimated a tion through capability (SH: index of moderated media-
mediation model based on 10,000 bootstrap samples (Hayes, tion = 0.45; 95% CI = [0.08 to 0.80]; SP: index of moder-
2018, PROCESS Model 4) with each of the three dimensions ated mediation = 0.44; 95% CI = [0.00 to 0.88]). Moreover,
of product complexity as independent variables, expected consistent with our results related to H5b, feature interrelat-
product capability and usability as mediators, and purchase edness reduced the negative effect of feature number on pur-
intention as the dependent variable. Moreover, we controlled chase intention through usability in the smart home context
for potential effects of age, gender, educational level, and (SH: index of moderated mediation = 0.20; 95% CI = [0.03
consumer expertise. Consistent with the results related to to 0.38]), but it increased the latter effect in the smartphone
H1a and H1b, we found no indirect effect of feature hetero- context (SP: index of moderated mediation =  − 0.11; 95%
geneity through capability (SH: b =  − 0.17; 95% CI = [− 0.35 CI = [− 0.25 to − 0.01]).
to 0.01]; SP: b = 0.08; 95% CI = [− 0.13 to 0.30]), but we
did observe a negative indirect effect through usability (SH: Test of mediation on other downstream variables To validate
b =  − 0.03; 95% CI = [− 0.08 to − 0.00]; SP: b =  − 0.07; 95% our previous findings, we reestimated the mediation and mod-
CI = [− 0.14 to − 0.01]), as well as a partially significant, erated mediation models with other downstream variables.

13
 Journal of the Academy of Marketing Science

First, following Thompson et al. (2005), we used product dimensions of complexity. It makes three key contributions
utility as the dependent variable and found the same pattern to research on new product design, advertising, and selling.
of effects as with product purchase intention (see Appendix 3,
part C4). Second, we aimed to verify whether feature hetero- More comprehensive understanding of product complexity
geneity and interrelatedness combined with feature number management By considering two additional dimensions
also have a downstream effect on product purchase behavior. of product complexity, our research enables marketing
Therefore, at the end of the smartphone study, we entered researchers to more comprehensively analyze the design
participants in a lottery and asked them the following on a of products and its impact on consumer perceptions and to
7-point scale (1 = “another smartphone,” 7 = “the described derive more differentiated insights into the optimal advertis-
smartphone”): “If you are among the winners of our lottery, ing and selling of products. Moreover, it complements previ-
for which smartphone may we send you a discount voucher ous research on complex products in an R&D and manufac-
for purchase?” We assumed that the more participants tended turing context (e.g., Ethiraj & Levinthal, 2004; Song et al.,
toward “the described smartphone,” the more favorable their 2015) by revealing that the heterogeneity and interrelated-
purchase behavior related to this smartphone would be. The ness of features have not only internal downstream effects
results showed the same pattern of effects as with product on processes and costs but also external effects on consumer
purchase intention (see Appendix 3, part C4). perceptions. In doing so, it contributes to a more holistic
approach to product complexity management that considers
both internal and external downstream effects of multiple
Discussion complexity dimensions.

Studies 2a and 2b replicated the findings of Studies 1a and Thorough understanding of the impact of the various dimen‑
1b by showing that feature heterogeneity and interrelatedness sions of product complexity By providing evidence that fea-
have discrete and somewhat different impacts on consumer ture heterogeneity and interrelatedness influence consumer
attitudes. Specifically, the results confirmed that both dimen- attitudes and intentions, we extend research on consumer
sions decrease expected product usability whereas only inter- perceptions of complex products, which so far has focused
relatedness increases expected product capability. We again on the impact of feature number (Goodman & Irmak, 2013;
tested whether consumers’ lack of trust in products offered to Sela & Berger, 2012; Thompson & Norton, 2011; Thomp-
them with a high dissimilarity of features could help explain son et al., 2005). Recent advancements in manufacturing,
our finding that feature heterogeneity does not increase electronics, and information technology and the resulting
expected product capability. The results of a mediation model increase in companies’ ability and tendency to rely on highly
based on 10,000 bootstrap samples (Hayes, 2018, PROCESS heterogeneous and interrelated product features make this
Model 4) again provide support for this explanation by show- theoretical insight even more important. We also show that
ing that heterogeneity decreases consumers’ product trust the two additional dimensions have both similarities and
and, in turn, expected product capability (indirect effect: differences in how they affect consumer attitudes and inten-
b =  − 0.20, SE (boot) = 0.10, 95% CI = [− 0.39; − 0.01]). More tions. In terms of similarities, both dimensions decrease
important, studies 2a and 2b revealed that not only feature expected product usability and, in turn, consumer purchase
number (see Thompson et al., 2005) but also feature hetero- intention. In terms of differences, feature interrelatedness
geneity and interrelatedness have downstream effects on con- enhances expected product capability and, in turn, consumer
sumer intentions and that these novel dimensions both also purchase intention, while feature heterogeneity does not,
affect the corresponding impact of feature number. Table 1 potentially because of a lack of consumer trust in products
summarizes the findings of the studies. It shows largely con- with a high dissimilarity of features. In a pre-usage context,
sistent findings across all studies and some noteworthy pat- consumers are typically unable and at least partly unwilling
terns of effects, which we subsequently discuss in more detail. to test a broad range of different product features, such that
doubts about the product’s ability to perform all these func-
tions are likely to prevail. Because Thompson et al. (2005)
General discussion examine the impact of feature number on the same depend-
ent variables, a comparison of our findings with theirs is
Theoretical contributions particularly necessary. We found that in terms of expected
product usability, both feature interrelatedness and hetero-
Previous research on consumer perceptions of complex geneity showed similar effects to those of feature number.
products has mostly ignored dimensions other than num- By contrast, in terms of expected product capability, feature
ber of features. This research represents the first investiga- interrelatedness had a similar effect, but feature heterogene-
tion of how consumers perceive products across multiple ity did not. Thus, in contrast with the quantity of features,

13
Journal of the Academy of Marketing Science

Table 1  Summary of results of hypotheses testing across studies

Hypotheses Studies Studies
1a (SH)/1b (SP) 2a (SH)/2b (SP)
(2 × 2 design) (2 × 2 × 2 design)

H1: The heterogeneity of product features …

(a) increases expected product capability. a) – / – a) – / –
(b) decreases expected product usability. b) ✓ / ✓ b) ✓ / ✓
H2: The interrelatedness of product features …
(a) increases expected product capability. a) ✓ / ✓ a) ✓ / ✓
(b) decreases expected product usability. b) ✓ / ✓ b) ✓ / ✓
H3: The number of product features …
(a) increases expected product capability. a) ✓ / ✓
(b) decreases expected product usability. b) ✓ / ✓
H4: When the heterogeneity of product features is high rather than low, the number of product features …
(a) increases expected product capability more strongly. a) – / –
(b) decreases expected product usability more strongly. b) ✓ / ✓
H5: When the interrelatedness of product features is high rather than low, the number of product features …
(a) increases expected product capability more strongly. a) ✓ / ✓
(b) decreases expected product usability more strongly. b) – / ✓
H6: Product purchase intention in increased by …
(a) expected product capability. a) ✓ / ✓
(b) expected product usability. b) ✓ / ✓

✓ confirmed; − not confirmed; SH smart home, SP smartphone

consumers do not seem to use the dissimilarity of features as offerings is likely to lead to adverse consequences on the
an indicator of a product’s ability to perform desired tasks. customer side whereas linking together the individual ele-
Moreover, the external downstream effects of feature ments of the bundle can create additional functions that help
heterogeneity and interrelatedness on consumer percep- solve customer problems.
tions tend to differ in valence from the internal down-
stream effects of these dimensions on internal processes Detailed understanding of the interplay between the various
and costs. For example, studies on product modularization dimensions of product complexity By revealing that both
argue that complex products should consist of modules as feature heterogeneity and interrelatedness moderate the
subsystems that highly differ in their functional properties external downstream effects of feature number, our research
and show minimal interdependence, which allows develop- shows that depending on these dimensions, the same feature
ment and production of a large variety of products at lower number can lead to considerably different consumer reac-
costs and change of the product architecture without loss of tions. While feature heterogeneity solely strengthens the
functionality or performance (e.g., Bonvoisin et al., 2016; usability-decreasing effect of feature number, feature inter-
Campagnolo & Camuffo, 2010). Therefore, while the het- relatedness primarily strengthens the capability-increasing
erogeneity of a product’s subsystems tends to have positive effect of feature number. In this context, a particularly nota-
internal downstream effects and interrelatedness negative ble result relates to the mixed moderating impact of feature
ones, our research indicates an opposite pattern for external interrelatedness on the negative effect of feature number on
downstream effects and, thus, a potential trade-off between expected product usability: in the smartphone context, the
internal and external consequences of these dimensions of usability-decreasing effect of feature number is especially
product complexity. strong when feature interrelatedness is high; by contrast, in
In addition, our findings contribute to research on solu- the smart home context, this effect is especially strong when
tion offerings that bundle products and services, or both, to feature interrelatedness is low. In a smartphone context, to
solve customer problems (e.g., Kindström & Kowalkowski, establish a connection consumers must typically select and
2009; Nordin & Kowalkowski, 2010; Windahl & Lakemond, link each interrelated feature with other interrelated features
2010). Studies in this field frequently refer to the high com- manually (e.g., “camera” followed by “SMS” to send a pic-
plexity of these offerings (Kreye, 2019; Zou et al., 2018). ture). Thus, when the features of a smartphone are highly
Our findings suggest that broadening the range of these interrelated, an increasing number of features and, thus,

13
 Journal of the Academy of Marketing Science

connections between features require more additional deci- customer side. Thus, in this context, managers may need
sion and usage effort than when the features are less inter- to trade off between internal and external consequences
related. By contrast, in a smart home context, consumers are of these product complexity dimensions, for example, by
typically able to create predefined processes (so-called auto- relying on standardized interfaces between the structural
mations) that enable them to regularly execute connected elements, which allows mitigation of negative effects on
features in an automatic way. Thus, when the features of a internal processes and costs while ensuring high connectiv-
smart home are highly interrelated, consumers are able to ity of these elements and, thus, enhanced functionality for
include a large share of features into predefined processes consumers.
that allow them to automatically execute a series of con- Moreover, to gain the maximum from a given number
nected features when using the product (e.g., “television” of features managers should simultaneously consider the
in combination with “multiroom audio” and “blinds” going heterogeneity and interrelatedness of these features when
down). Therefore, when the interrelatedness of smart home designing, advertising, and selling products. For example,
features is high, an increasing number of features and, thus, by limiting actual or perceived feature heterogeneity and
connections between features require less additional deci- interrelatedness, they may lower the risk that consumers will
sion and usage effort than when the interrelatedness of smart expect a feature-rich product to be difficult to learn and use.
home features is low and, thus, few features can be inte- Moreover, by fostering actual or perceived feature interrelat-
grated into predefined processes and many features must be edness, they can help exploit the potential of a feature-rich
manually executed. Consequently, the direction of the mod- new product to foster consumer beliefs about the product’s
erating impact of feature interrelatedness seems to depend capability. Conversely, managers should consider a product’s
on a product’s degree of automation. When the degree of number of features when deciding on the heterogeneity and
automation is rather low, feature interrelatedness tends to interrelatedness of these features. We show that when the
increase the negative effect of feature number on expected product is equipped with only a few features, managers can
product usability, whereas when the degree of automation afford to rely on high feature heterogeneity. By contrast, par-
is rather high, it tends to decrease the corresponding effect. ticularly for a feature-rich product, managers should ensure
that features are interrelated.
Practical implications
Promoting functional connectivity of features Rather than
This research offers recommendations on how best to man- simply adding new features, and thus continuing the ongo-
age the complexity of products across different stages of ing trend toward feature creep, managers could emphasize
the innovation process, thereby maximizing the probability the interrelatedness of features. Specifically, when designing
of new product success rather than failure and elimination products, we advise to consider the functional connectiv-
(Cooper, 1979; Prigge et al., 2018). It provides guidance ity of features an important criterion for the selection of
on product design (i.e., how to configure the various com- features. These connections should not only be technically
plexity dimensions) before market launch, and after market possible but also be easily buildable and capable of provid-
launch, it offers advice on product advertising and selling ing actual value for consumers. When creating manuals for
(i.e., which messaging about these dimensions to send to products with high feature interrelatedness, it is important
target groups). to include clear instructions on which features can be linked,
how this can be done, and what benefits result from these
Adopting a broader perspective on product complexity man‑ connections. Moreover, when advertising and selling prod-
agement Managers need to understand that consumers care ucts, managers should highlight the functional connectiv-
not only about a product’s pure number of features but also ity of features and demonstrate the ease of use and associ-
about their functional dissimilarity and interdependence. ated benefits for consumers. For example, in commercials,
Thus, these characteristics are relevant for both R&D and print media, brochures, and websites, the interrelatedness
manufacturing and marketing. Therefore, when designing of features could be illustrated by connecting lines between
products managers should care not only about consequences feature-related symbols. In this context, as well as in con-
for internal processes and costs but also about consequences sultative and sales meetings, selected interrelated features
for consumer perceptions. For modular product architec- and their ease of use and resulting benefits could also be
tures, this advice may be particularly important because, highlighted and demonstrated to consumers.
in this context, modules as key structural elements should
highly differ in their functional properties and not be inter- Downplaying functional dissimilarity of features. Instead
dependent from other modules to optimize the consequences of constantly broadening the functional range of features
of product design. By contrast, our findings show that this to offer an “all-in-one” solution for every purpose, manag-
product design would lead to adverse consequences on the ers should rather deemphasize the dissimilarity of features.

13
Journal of the Academy of Marketing Science

Specifically, they could frame a product as offering comple- transferring our conceptualization to other product catego-
mentary features for a similar overall purpose. For example, ries. In this context, examining differences between products
rather than highlighting that a smart home system controls with a high versus low degree of automation would be par-
both housing and lifestyle technology, they could empha- ticularly fruitful. Fifth, we focused on consumers. Further
size the product’s overarching purpose of smart home con- studies could extend this investigation to business customers.
trol, which makes the features appear more homogeneous. Sixth, to examine the impact of our findings on actual prod-
Moreover, particularly for products with high feature hetero- uct purchase, future research could run a field experiment or
geneity, managers may more strongly try to build trust, for analyze secondary data.
example, by providing credible information about the prod-
uct’s capability, such as consumer reviews about subjective
user experiences or third-party reviews about the results of Appendix 1
objective product tests.
Details on development of stimuli
Solving the capability–usability trade‑off through user‑
friendly design Finally, though not directly following from Stimuli for heterogeneity of product features
the findings of our research as the previous recommenda-
tions did, we recommend that rather than purely relying on To validate the operationalization of the heterogeneity of
weighing increased product capability with decreased product features, we drew on Blau’s (1977) index of heterogeneity
usability, or vice versa, managers could also try to relieve the (B) using the formula B = 1 − ∑ pi2, with pi as the proportion
tension between these competing consumer needs. To this of features belonging to the functional category i. This index
end, a viable approach could be to develop more user-friendly describes the probability that two randomly selected fea-
product designs, such as through user surfaces that are more tures belong to different functional categories. Its minimum
intuitive and better handling and decision support based on value Bmin is 0, and its maximum value Bmax can be calcu-
artificial intelligence. lated with the formula Bmax = (x – 1)/x, where x refers to the
quantity of functional categories. Thus, when features can
belong to three functional categories, such as in our research,
Limitations and avenues for further research the value range of this index is between 0 and 0.67. For
our “low heterogeneity of features” condition, we obtained
This article has several limitations that offer avenues for B = 0 for the smart home and smartphone contexts, and for
further research. First, our study investigated the impact our “high heterogeneity of features” condition, we found
of various dimensions of product complexity on consumer B values between 0.64 and 0.67 for both product contexts
attitudes and intentions before they actually use the prod- (see parts A1 and B1 of the Supplemental Material). These
uct. Future studies might examine consumer reactions to results provide evidence for the appropriateness of our low/
product complexity during or after product usage to provide high classification in this dimension of product complexity.
further nuance to our findings. Second, literature on infor-
mation load suggests an inverted U-shaped effect of product Stimuli for interrelatedness of product features
information on consumer attitudes and intentions (Eppler
& Mengis, 2004; Roetzel, 2019). However, our theoretical To calculate the maximum quantity of connections between
reasoning for the hypotheses draws on the implicit assump- features cmax for the three conditions of number of prod-
tion that consumers already have a fair amount of informa- uct features, we applied the binomial coefficient “n choose
tion and thus focuses on the negative slope section from 2” and the related formula cmax = n(n − 1)/2 (Goetgheluck,
this literature. Future research could examine whether our 1987; Wasserman & Faust, 1994), where n is the number of
findings hold in situations when consumers have only a lim- features. Using this logic leads to 5(5 − 1)/2 = 10 maximum
ited amount of information. Alternatively, because our data connections for the “low number of features” condition,
do not allow us to test for quadratic effects, future research 10(10 − 9)/2 = 45 maximum connections for the “medium
could also investigate the prevalence of an inverted U-shaped number of features” condition, and 15(15 − 14)/2 = 105
effect. Third, our study controlled for potential effects of maximum connections for the “high number of features”
demographics and psychographics. Further research could condition.
investigate these and other consumer characteristics, such To discount the proportion value pci for the “high number
as brand knowledge, product category knowledge, and of features” condition by the factor 2 compared with the “low
product involvement as potential moderators. Fourth, we number of features” condition, we relied for “low number of
concentrated on smart home systems and smartphones as features” on pci = 0.10 for “low interrelatedness of features,”
product categories. Additional research would benefit from thus using 1 connection (i.e., c = pci * cmax = 0.10 * 10 = 1),

13
 Journal of the Academy of Marketing Science

and pci = 0.50 for “high interrelatedness of features,” thus of heterogeneity nor the interaction was significant. Thus,
using 5 connections (i.e., c = pci * cmax = 0.50 * 10 = 5). By all manipulations worked as intended. In addition, we ana-
contrast, for “high number of features,” we applied pci = 0.05 lyzed whether participants perceived the products as equal
for “low interrelatedness of features,” resulting in 5 connec- in terms of number of features and found no significant
tions (i.e., c = pci * cmax = 0.05 * 105 = 5), and pci = 0.25 for differences among the four scenarios (SH: F(1, 71) = 2.64,
“high interrelatedness of features,” leading to 26 connec- p = 0.06; SP: F(1, 73) = 1.88, p = 0.14).
tions (i.e., c = pci * cmax = 0.25 * 105 = 26). Similarly, for a We also pretested whether the two dimensions empiri-
medium number of features, we used the average between cally differ from each other. For this purpose, we presented
the proportion value pci of the “low number of features” con- participants with two products with the same number of
dition and the “high number of features” condition. Thus, we features. The products differed only in the heterogeneity
relied on pci = 0.075 for “low interrelatedness of features,” and interrelatedness of features, such that one was low in
obtaining 3 connections (i.e., c = pci * cmax = 0.075 * 45 = 3), heterogeneity but high in interrelatedness and the other was
and pci = 0.375 for “high interrelatedness of features,” result- high in heterogeneity but low in interrelatedness. We asked
ing in 17 connections (i.e., c = pci * cmax = 0.375 * 45 = 17) participants to rate the systems/smartphones on a 7-point
(see parts A2 and B2 of the Supplemental Material). scale (1 = “The smart home systems/smartphones are identi-
cal,” 7 = “The smart home systems/smartphones are different
from each other”) and found that they rated the two products
Appendix 2 as very different (SH: M = 6.57; SP: M = 6.40).

Details on measures of studies 1a and 1b

Details on studies 1a and 1b
We measured consumer expertise with five items (e.g., “How
Details on pretest for studies 1a and 1b familiar are you with [smart home systems; smartphones]?”
“not familiar at all/extremely familiar”; SH: α = 0.91; SP:
To ensure the suitability of our manipulations and rule α = 0.90; Thompson et al., 2005), consumer trust with three
out confounding effects, we carried out comprehen- items (e.g., “I expect this [smart home system; smartphone] to
sive pretests of stimuli with consumers in the smart deliver on its promise”; SH: α = 0.91; Li et al., 2008), expected
home (SH) and smartphone (SP) contexts (SH: n = 75; product capability with three items (e.g., “This [smart home
48% female, Mage = 47.69 years; SP: n = 77; 47% female, system; smartphone] will perform well”; SH: α = 0.97; SP:
Mage = 45.87 years). We used two 2 × 2 analyses of vari- α = 0.95), and expected product usability with eight items (e.g.,
ance (ANOVAs) with heterogeneity and interrelatedness of “Learning to use this [smart home system; smartphone] will be
features as independent variables and the two manipulation easy for me”; SH: α = 0.97; SP: α = 0.96). Part D of the Sup-
check questions as dependent variables, respectively, while plemental Material shows the complete items for each scale.
holding the number of features constant on a medium level.
We applied 7-point scales for all questions. As expected, Details on manipulation checks for studies 1a and 1b
for the manipulation check question for the heterogeneity
of features (“This [smart home system; smartphone] con- Similar to our pretests of manipulations, for both Studies
sists of features of the same category vs. features of dif- 1a and 1b, we used a 2 × 2 ANOVA with heterogeneity and
ferent categories”) as the dependent variable, the analysis interrelatedness of features as independent variables and the
showed only a significant effect for heterogeneity of fea- two manipulation check questions as dependent variables,
tures (SH: Mlow hete = 4.94, SD = 1.89 vs. Mhigh hete = 5.87, respectively. For the manipulation check question related to
SD = 1.32; F(1, 71) = 5.86, p < 0.05; SP: Mlow hete = 3.64, heterogeneity of features, the results showed only a signifi-
SD = 1.67 vs. Mhigh hete = 5.06, SD = 1.64; F(1, 73) = 13.88, cant effect of heterogeneity (SH: Mlow hete = 4.54, SD = 2.01
p < 0.01), but neither the effect of interrelatedness nor the vs. Mhigh hete = 5.65, SD = 1.66; F(1, 236) = 21.50, p < 0.01;
interaction was significant. Similarly, for the manipulation SP: Mlow hete = 3.71, SD = 1.92 vs. Mhigh hete = 5.02, SD = 1.29;
check question for the interrelatedness of features as the F(1, 191) = 28.78, p < 0.01). By contrast, the effect of inter-
dependent variable (“This [smart home system; smartphone] relatedness and the interaction did not reach significance. For
shows a low degree of functional connectivity vs. a high the manipulation check question related to interrelatedness
degree of functional connectivity”), only interrelatedness of features, only interrelatedness had a significant effect in
of features showed a significant effect (SH: Mlow inter = 3.77, the smart home context (SH: Mlow inter = 3.42, SD = 1.83 vs.
SD = 1.94 vs. Mhigh inter = 5.93, SD = 1.09; F(1, 71) = 36.62, Mhigh inter = 5.72, SD = 1.51; F(1, 236) = 111.93, p < 0.01);
p < 0.01; SP: Mlow inter = 3.23, SD = 1.73 vs. Mhigh inter = 5.71, neither the effect of heterogeneity nor the interaction was sig-
SD = 1.20; F(1, 73) = 51.13, p < 0.01), but neither the effect nificant. In the smartphone context, we observed a significant

13
Journal of the Academy of Marketing Science

main effect of not only interrelatedness (SP: Mlow inter = 3.01, reached significance. Similarly, for interrelatedness of
SD = 1.58 vs. Mhigh inter = 5.17, SD = 1.28; F(1, 191) = 108.86, product features we used the same manipulation check
p < 0.01) but also heterogeneity (F(1, 191) = 6.76, p < 0.05), question as in Studies 1a and 1b. The results showed a sig-
while the interaction was not significant (F(1, 191) = 0.14, nificant effect for interrelatedness of product features (SH:
p = 0.71). Given the significant effect of heterogeneity, we Mlow inter = 3.32, SD = 2.03 vs. Mhigh inter = 5.03, SD = 1.81;
followed the common advice (Perdue & Summers, 1986) and F(1, 104) = 22.11, p < 0.01; SP: Mlow inter = 3.33, SD = 1.84
practice (Bellezza et al., 2014; Sipilä et al., 2021) to compare vs. Mhigh inter = 4.94, SD = 1.67; F(1, 65) = 15.84, p < 0.01),
effect sizes of all significant effects. Perdue and Summers while no other effects were significant. Thus, all manipula-
(1986, p. 323) recommended that “when in the analysis of tions worked as intended.
the manipulation check for A the effects sizes for B [and AB]
… are much smaller than that for A, their statistical signifi- Details on measures of studies 2a and 2b
cance probably should not be of great concern”. In our study,
we found that for the manipulation check question related to We used the same scales for measuring consumer exper-
interrelatedness of features, the effect size of interrelatedness tise (SH: α = 0.90; SP: α = 0.90), consumer trust (SH:
(η2 = 0.36) was 12 times higher than that of heterogeneity α = 0.91), expected product capability (SH: α = 0.96; SP:
(η2 = 0.03), providing sufficient support for our manipula- α = 0.96), and expected product usability (SH: α = 0.97;
tion. Therefore, and following the suggestion of Sawyer et al., SP: α = 0.96). In addition, we measured product purchase
(1995, p. 592) that manipulation and confounding checks add intention with three items from Dodds et al. (1991) on
less informational value “when … the independent variables a 7-point scale (1 = “strongly disagree,” 7 = “strongly
are isomorphic with their operationalization,” such as in our agree”) (e.g., “My willingness to buy this [smart home
study, we relied on the corresponding stimuli. system; smartphone] is very high”; SH: α = 0.98; SP:
α = 0.98). In part D of the Supplemental Material, we show
the complete items for each scale.
Appendix 3
Details on manipulation checks for studies 2a and 2b
Details on studies 2a and 2b
As in our pretests of manipulations, for both Studies 2a and
Details on pretest for studies 2a and 2b 2b, we used a 2 × 2 × 2 ANOVA with number, heterogene-
ity, and interrelatedness of features as independent variables
As in Studies 1a and 1b, we conducted comprehensive and the three manipulation check questions as the respec-
pretests of our stimuli with consumers in both smart tive dependent variables. For the manipulation check ques-
home (SH) and smartphone (SP) contexts (SH: n = 112; tion related to number of features, we observed a significant
51% female, Mage = 43.72 years; SP: n = 73; 40% female, effect of not only number (SH: Mlow num = 3.33, SD = 1.96
­M age = 45.05 years). We used two 2 × 2 × 2 ANOVAs vs. Mhigh num = 6.03, SD = 1.50; F(1, 431) = 276.67, p < 0.01;
with number, heterogeneity, and interrelatedness of fea- SP: Mlow num = 3.08, SD = 1.74 vs. Mhigh num = 5.50, SD = 1.54;
tures as independent variables and the three manipula- F(1, 441) = 254.87, p < 0.01) but also interrelatedness (SH:
tion check questions as dependent variables. Again, we F(1, 431) = 23.77, p < 0.01; SP: F(1, 441) = 24.04, p < 0.01).
applied 7-point scales for all questions. As expected, Given the significant effect of interrelatedness, we followed
for the manipulation check question with the number of common recommendations (Perdue & Summers, 1986)
product features (“This [smart home system; smartphone] and practices (Bellezza et al., 2014; Sipilä et al., 2021) to
has a low number of features vs. a high number of fea- compare effect sizes. The size of the focal effect of number
tures”) as the dependent variable, the results showed only (SH: η2 = 0.39; SP: η2 = 0.37) was significantly higher than
a significant effect for number of product features (SH: the effect of interrelatedness (SH: η2 = 0.05; SP: η2 = 0.05),
Mlow num = 3.86, SD = 1.84 vs. Mhigh num = 5.99, SD = 1.32; providing sufficient support for our manipulation. For the
F(1, 104) = 46.00, p < 0.01; SP: Mlow num = 3.42, SD = 1.82 manipulation check question related to heterogeneity of fea-
vs. Mhigh num = 5.36, SD = 1.46; F(1, 65) = 25.27, p < 0.01), tures, the results showed a strong, significant effect of heter-
while no other effects were significant. For heterogeneity ogeneity (SH: Mlow hete = 3.70, SD = 2.26 vs. Mhigh hete = 5.73,
of product features, we used the same manipulation check SD = 1.57; F(1, 431) = 130.28, p < 0.01, η 2 = 0.23; SP:
question as in Studies 1a and 1b and found support for a Mlow hete = 3.82, SD = 1.85 vs. Mhigh hete = 5.00, SD = 1.71;
successful manipulation (SH: Mlow hete = 3.75, SD = 1.63 F(1, 441) = 51.01, p < 0.01, η2 = 0.10). We also found a
vs. Mhigh hete = 5.18, SD = 1.55; F(1, 104) = 22.19, p < 0.01; weak, significant effect of number (SH: F(1, 431) = 22.38,
SP: M low hete = 3.40, SD = 1.99 vs. M high hete = 4.61, p < 0.01, η2 = 0.05; SP: F(1, 441) = 22.58, p < 0.01, η2 = 0.05)
SD = 1.59; F(1, 65) = 7.28, p < 0.05), while no other effects and, at least for one product context, of interrelatedness

13
 Journal of the Academy of Marketing Science

(SH: F(1, 431) = 10.40, p < 0.01, η2 = 0.02). The size of (SH: b = 0.36; 95% CI = [0.13 to 0.59]; SP: b = 0.58; 95%
the focal effect of heterogeneity was significantly higher CI = [0.37 to 0.80]) and usability (SH: b =  − 0.17; 95%
in all cases, providing sufficient support for our manipula- CI = [− 0.26 to − 0.09]; SP: b =  − 0.17; 95% CI = [− 0.26
tion (Bellezza et al., 2014; Perdue & Summers, 1986; Sipilä to − 0.09]), confirming our mediation findings related to
et al., 2021). Finally, for the manipulation check question purchase intention as the dependent variable.
related to interrelatedness, we found a strong, significant
effect of interrelatedness (SH: Mlow inter = 3.05, SD = 1.87 Test of moderated mediation on expected product util‑
vs. Mhigh inter = 5.52, SD = 1.72; F(1, 431) = 232.92, p < 0.01, ity We also ran a moderated mediation model (Hayes, 2018,
η2 = 0.35; SP: Mlow inter = 3.08, SD = 1.79 vs. Mhigh inter = 5.15, PROCESS Model 7) with feature number as the independ-
SD = 1.48; F(1, 441) = 195.85, p < 0.01, η2 = 0.31); the ent variable, feature heterogeneity and interrelatedness as
results also showed a weak, significant effect of number (SH: moderators, expected capability and usability as mediators,
F(1, 431) = 49.60, p < 0.01, η2 = 0.10; SP: F(1, 441) = 54.47, and expected product utility as the dependent variable. For
p < 0.01, η2 = 0.11). Again, we compared effect sizes and feature heterogeneity, we found a significant moderated
found that the size of the focal effect of interrelatedness was mediation not through capability (SH: index of moderated
significantly higher in all cases, providing sufficient support mediation =  − 0.17; 95% CI = [− 0.65 to 0.30]; SP: index of
for our manipulation (Bellezza et al., 2014; Perdue & Sum- moderated mediation =  − 0.17; 95% CI = [− 0.59 to 0.23])
mers, 1986; Sipilä et al., 2021). Therefore, and again consid- but through usability (SH: index of moderated media-
ering the suggestion of Sawyer et al. (1995) that manipula- tion =  − 0.28; 95% CI = [− 0.54 to − 0.03]; SP: index of mod-
tion and confounding checks have less informational value erated mediation =  − 0.11; 95% CI = [− 0.24 to − 0.02]). For
when the independent variables of an experiment are iso- feature interrelatedness, our results indicate a significant mod-
morphic with their operationalization, such as in our study, erated mediation through capability (SH: index of moderated
we drew on the corresponding stimuli. mediation = 0.57; 95% CI = [0.10 to 1.04]; SP: index of mod-
erated mediation = 0.43; 95% CI = [0.00 to 0.85]) and usability
Details on results (tests of mediation on other downstream (SH: index of moderated mediation = 0.27; 95% CI = [0.04
variables) to 0.52]; SP: index of moderated mediation =  − 0.08; 95%
CI = [− 0.18 to − 0.00]). In summary, these findings are fully
Test of mediation on expected product utility For valida- consistent with our mediated moderation findings related to
tion purposes, we estimated a mediation model based on purchase intention as the dependent variable.
10,000 bootstrap samples (Hayes, 2018, PROCESS Model
4) with each of the three dimensions of product complexity Test of mediation on product purchase behavior We con-
as independent variables, expected capability and usability ducted mediation analysis based on 10,000 bootstrap samples
as mediators, and expected product utility as the depend- (Hayes, 2018, PROCESS Model 4) with each of the three
ent variable. We measured expected product utility on a dimensions of product complexity as independent variables,
7-point semantic differential scale with the same items that expected capability and expected usability as mediators,
Thompson et al. (2005) use (“Please evaluate this [smart and product purchase behavior as the dependent variable.
home system; smartphone] by the following characteristics: Consistent with our other findings, we found a significant
bad vs. good; unlikeable vs. likeable; not useful vs. useful; indirect effect of heterogeneity of features not through capa-
low quality vs. high quality; undesirable vs. desirable; unfa- bility (b = 0.05; 95% CI = [− 0.08 to 0.19]) but through usa-
vorable vs. favorable”; SH: α = 0.97; SP: α = 0.97). In line bility (b =  − 0.06; 95% CI = [− 0.12 to − 0.01]). Furthermore,
with the results related to purchase intention as the depend- in line with our previous findings, we found a significant
ent variable, we found a significant indirect effect of fea- indirect effect of feature interrelatedness through capability
ture heterogeneity not through capability (SH: b =  − 0.21; (b = 0.48; 95% CI = [0.32 to 0.66]) and usability (b =  − 0.06;
95% CI = [− 0.43 to 0.01]; SP: b = 0.08; 95% CI = [− 0.12 95% CI = [− 0.14 to − 0.01]) as well as of feature number
to 0.29]) but through usability (SH: b =  − 0.05; 95% through capability (b = 0.36; 95% CI = [0.22 to 0.54]) and
CI = [− 0.11 to − 0.00]; SP: b =  − 0.05; 95% CI = [− 0.11 usability (b =  − 0.17; 95% CI = [− 0.31 to − 0.05]).
to − 0.01]). Moreover, also in line with the results related
to purchase intention as the dependent variable, we found Test of moderated mediation on product purchase behav‑
a significant indirect effect of feature interrelatedness ior We estimated a moderated mediation model (Hayes,
through capability (SH: b = 0.46; 95% CI = [0.24 to 0.69]; 2018, PROCESS Model 7) with feature number as the inde-
SP: b = 0.76; 95% CI = [0.56 to 0.98]) and usability (SH: pendent variable, feature heterogeneity and interrelatedness
b =  − 0.08; 95% CI = [− 0.14 to − 0.03]; SP: b = –0.06; 95% as moderators, expected capability and usability as mediators,
CI = [− 0.11 to − 0.01]). In addition, the results indicated sig- and product purchase behavior as the dependent variable.
nificant indirect effects of feature number through capability For feature heterogeneity, we found a significant moderated

13
Journal of the Academy of Marketing Science

mediation not through capability (index of moderated media- Bonvoisin, J., Halstenberg, F., Buchert, T., & Stark, R. (2016). A sys-
tion =  − 0.11; 95% CI = [− 0.38 to 0.15]) but through usabil- tematic literature review on modular product design. Journal of
Engineering Design, 27(7), 488–514.
ity (index of moderated mediation =  − 0.11; 95% CI = [− 0.25 Buescher, M., Slack, R., Rouncefield, M., Procter, R., Hartswood, M.,
to − 0.01]). For feature interrelatedness, our results show a & Voss, A. (2009). Configuring user-designer relations. Springer.
significant moderated mediation through capability (index Campagnolo, D., & Camuffo, A. (2010). The concept of modularity in
of moderated mediation = 0.27; 95% CI = [0.00 to 0.56]) management studies: A literature review. International Journal of
Management Reviews, 12(3), 259–283.
and usability (index of moderated mediation =  − 0.08; 95% Chang, Y., Dong, X., & Sun, W. (2014). Influence of characteristics
CI = [− 0.21 to − 0.00]). Overall, these findings are fully in of the Internet of Things on consumer purchase intention. Social
line with our mediated moderation findings related to pur- Behavior and Personality, 42(2), 321–330.
chase intention as the dependent variable. Cooper, R. G. (1979). The dimensions of industrial new product suc-
cess and failure. Journal of Marketing, 43(3), 93–103.
Davis, F. D. (1989). Perceived usefulness, perceived ease of use, and
user acceptance of information technology. MIS Quarterly, 13(3),
Supplementary Information The online version contains supplemen- 319–340.
tary material available at https://d​ oi.o​ rg/1​ 0.1​ 007/s​ 11747-0​ 23-0​ 0933-7. De Angelis, M., & Carpenter, G. S. (2009). The effect of adding fea-
tures on product attractiveness: The role of product perceived con-
Funding Open Access funding enabled and organized by Projekt gruity. In A. L. McGill & S. Shavitt (Eds.), Advances in consumer
DEAL. research (Vol. 36, pp. 651–652), Duluth, MN: Association for
Consumer Research.
Declarations Dodds, W. B., Monroe, K. B., & Grewal, D. (1991). Effects of price,
brand, and store information on buyers’ product evaluations. Jour-
Conflict of interest The authors declare that they have no conflict of nal of Marketing Research, 28(3), 307–319.
interest. Duncan, R. B. (1972). Characteristics of organizational environments
and perceived environmental uncertainty. Administrative Science
Open Access This article is licensed under a Creative Commons Attri- Quarterly, 17(3), 313–327.
bution 4.0 International License, which permits use, sharing, adapta- Eppler, M. J., & Mengis, J. (2004). The concept of information overload:
tion, distribution and reproduction in any medium or format, as long A review of literature from organization science, accounting, market-
as you give appropriate credit to the original author(s) and the source, ing, MIS, and related disciplines. The Information Society, 20(5),
provide a link to the Creative Commons licence, and indicate if changes 325–344.
were made. The images or other third party material in this article are Estes, Z. (2003). A tale of two similarities: Comparison and integra-
included in the article's Creative Commons licence, unless indicated tion in conceptual combination. Cognitive Science, 27, 911–921.
otherwise in a credit line to the material. If material is not included in Ethiraj, S. K., & Levinthal, D. (2004). Modularity and innovation in
the article's Creative Commons licence and your intended use is not complex systems. Management Science, 50(2), 159–173.
permitted by statutory regulation or exceeds the permitted use, you will Fürst, A., & Scholl, M. (2022). Multi-channel management and
need to obtain permission directly from the copyright holder. To view a design: An analysis of their impact on multi-channel conflict
copy of this licence, visit http://​creat​iveco​mmons.​org/​licen​ses/​by/4.​0/. and success. Marketing ZFP, 44(3), 24–43.
Fürst, A., & Staritz, M. (2022). Creating superior value in the eyes
of the customer: An analysis of the two value drivers and value
paths. Marketing ZFP, 44(3), 3–23.
References Gattol, V., Sääksjärvi, M., Gill, T., & Schoormans, J. (2016). Feature
fit – The role of congruence and complementarity when adding
Achrol, R. S., & Stern, L. W. (1988). Environmental determinants of versus deleting features from products. European Journal of
decision-making uncertainty in marketing channels. Journal of Innovation Management, 19(4), 589–607.
Marketing Research, 25(1), 36–50. Gibbert, M., & Mazursky, D. (2009). How successful would a phone-
Bandyopadhyay, S., Rao, A. R., & Sinha, B. (2010). Models for social pillow be. Journal of Consumer Psychology, 19(4), 652–660.
networks with statistical applications. SAGE. Goetgheluck, P. (1987). Computing binomial coefficients. American
Bellezza, S., Gino, F., & Keinan, A. (2014). The red sneakers effect: Mathematical Monthly, 94(4), 360–365.
Inferring status and competence from signals of nonconformity. Goodman, J. K., & Irmak, C. (2013). Having versus consuming: Fail-
Journal of Consumer Research, 41(1), 35–54. ure to estimate usage frequency makes consumers prefer multi-
Bertini, M., Ofek, E., & Ariely, D. (2009). The impact of add-on fea- feature products. Journal of Marketing Research, 50(1), 44–54.
tures on consumer product evaluations. Journal of Consumer Gubbi, J., Buyya, R., Marusic, S., & Palaniswami, M. (2013). Internet
Research, 36(1), 17–28. of Things (IoT): A vision, architectural elements, and future direc-
Bettencourt, L. A., & Bettencourt, S. L. (2011). Innovating on the tions. Future Generation Computer Systems, 29(7), 1645–1660.
cheap. Harvard Business Review, 89(6), 88–94. Gupta, A., Kumar, A., Grewal, R., & Lilien, G. L. (2019). Within-seller
Bettman, J., Johnson, E., & Payne, J. (1991). Consumer decision mak- and buyer–seller network structures and key account profitability.
ing. In T. Robertson & H. Kassarjian (Eds.), Handbook of con- Journal of Marketing, 83(1), 108–132.
sumer behavior (pp. 50–84). Prentice Hall. Hayes, A. F. (2018). Introduction to mediation, moderation, and con-
Blau, P. M. (1977). Inequality and heterogeneity: A primitive theory of ditional process analysis: A regression-based approach (2d ed.).
social structure. Free Press. The Guilford Press.
Blut, M., Wang, Ch., & Schoefer, K. (2016). Factors influencing Hoffman, D. L., & Novak, T. P. (2018). Consumer and object
the acceptance of self-service technologies. Journal of Service experience in the Internet of Things: An assemblage theory
Research., 19(4), 396–416. approach. Journal of Consumer Research, 44(6), 1178–1204.

13
 Journal of the Academy of Marketing Science

Iyengar, S. S., & Lepper, M. R. (2000). When choice is demotivating: Pinochet, L. H. C., Lopes, E. L., Srulzon, C. H. F., & Onusic, L. M.
Can one desire too much of a good thing? Journal of Personality (2018). The influence of the attributes of “Internet of Things”
and Social Psychology, 79(6), 995–1006. products on functional and emotional experiences of purchase
Ji, Y. G., Park, J., Lee, C., & Yun, M. (2006). A usability checklist intention. Innovation & Management Review, 15(3), 303–320.
for the usability evaluation of mobile phone user interface. Inter- Preece, J., Rogers, Y., & Sharp, H. (2002). Interaction design. Wiley.
national Journal Human-Computer Interaction, 20(3), 207–231. Prigge, J., Homburg, Ch., & Fürst, A. (2018). Addressing a product
Johnson, E. J., & Payne, J. W. (1985). Effort and accuracy in choice. management’s orphan: How to externally implement product
Management Science, 31(4), 395–414. eliminations in a B2B Setting. Industrial Marketing Management,
Kannan, P. K., & Li, H. (2017). Digital marketing: A framework, 68(1), 56–73.
review and research agenda. International Journal of Research Raff, S., Wentzel, D., & Obwegeser, N. (2020). Smart products: Con-
in Marketing, 34(1), 22–45. ceptual review, synthesis, and research directions. Journal of
Kim, H.-W., Chan, H. C., & Gupta, S. (2007). Value-based adoption Product Innovation Management, 37(5), 379–404.
of mobile internet: An empirical investigation. Decision Support Roetzel, P. G. (2019). Information overload in the information age: A
Systems, 43(1), 111–126. review of the literature from business administration, business
Kindström, D., & Kowalkowski, C. (2009). Development of industrial psychology, and related disciplines with a bibliometric approach
service offerings: A process framework. Journal of Service Man- and framework development. Business Research, 12(2), 479–522.
agement, 20(2), 156–172. Sääksjärvi, M., & Samiee, S. (2011). Assessing multifunctional innova-
Kreye, M. E. (2019). Does a more complex service offering increase tion adoption via an integrative model. Journal of the Academy of
uncertainty in operations? International Journal of Operations & Marketing Science, 39(5), 717–735.
Production Management, 39(1), 75–93. Sawyer, A. G., Lynch, J. G., & Brinberg, D. L. (1995). A Bayesian
Krishnan, H. S. (1996). Characteristics of memory associations: A analysis of the information value of manipulation and confound-
consumer-based brand equity perspective. International Journal ing checks in theory tests. Journal of Consumer Research, 21(4),
of Research in Marketing, 13(4), 389–405. 581–595.
Kuehnl, Ch., Fürst, A., Homburg, Ch., & Staritz, M. (2017). Toward a Scott, J. (1988). Social network analysis. Sociology, 22(1), 109–127.
differentiated understanding of the value-creation chain. British Sela, A., & Berger, J. (2012). How attribute quantity influences option
Journal of Management, 28(3), 444–463. choice. Journal of Marketing Research, 49(6), 942–953.
Lau, A. K. W., Yam, R. C. M., & Tang, E. (2011). The impact of Shugan, S. M. (1980). The cost of thinking. Journal of Consumer
product modularity on new product performance: Mediation by Research, 7(2), 99–111.
product innovativeness. Journal of Product Innovation Manage- Sipilä, J., Alavi, S., Edinger-Schons, L. M., Dörfer, S., & Schmitz, Ch.
ment, 28(2), 270–284. (2021). Corporate social responsibility in luxury contexts: Poten-
Lee, J., & Chu, W. (2021). The effect of adding focal-goal similar tial pitfalls and how to overcome them. Journal of the Academy of
versus dissimilar attributes on convergence product purchase deci- Marketing Science, 49(2), 280–303.
sion: The role of relational and item-specific elaboration style. Song, W., Wu, Z., Li, X., & Xu, Z. (2015). Modularizing product
Journal of Consumer Behaviour, 21(2), 1–14. extension services: An approach based on modified service blue-
Li, F., Kashyap, R., Zhou, N., & Yang, Z. (2008). Brand trust as a print and fuzzy graph. Computers & Industrial Engineering, 85,
second-order factor: An alternative measurement model. Interna- 186–195.
tional Journal of Market Research, 50(6), 817–839. Thompson, D. V., Hamilton, R. W., & Rust, R. T. (2005). Feature fatigue:
Meyer, R. J., Zhao, S., & Han, J. K. (2008). Biases in valuation vs. When product capabilities become too much of a good thing. Jour-
usage of innovative product features. Marketing Science, 27(6), nal of Marketing Research, 42(4), 431–442.
1083–1096. Thompson, D. V., & Norton, M. I. (2011). The social utility of feature
Mukherjee, A., & Hoyer, W. D. (2001). The effect of novel attributes on creep. Journal of Marketing Research, 48(3), 555–565.
product evaluation. Journal of Consumer Research, 28(3), 462–472. Thompson, J. D. (1967). Organizations in action. McGraw-Hill.
Ng, I. C. L., & Wakenshaw, S. Y. L. (2017). The Internet-of-Things: Von Bertalanffy, L. (1968). General system theory: Foundations, devel-
Review and research directions. International Journal of Research opment, applications. George Braziller.
in Marketing, 34(1), 3–21. Wasserman, S., & Faust, K. (1994). Social network analysis: Methods
Nordin, F., & Kowalkowski, C. (2010). Solutions offerings: A critical and applications. Cambridge University Press.
review and reconceptualisation. Journal of Service Management, Wilkenfeld, M. J., & Ward, T. B. (2001). Similarity and emergence
21(4), 441–459. in conceptual combination. Journal of Memory and Language,
Nowlis, S. M., & Simonson, I. (1996). The effect of new product fea- 45(1), 21–38.
tures on brand choice. Journal of Marketing Research, 33(1), Windahl, C., & Lakemond, N. (2010). Integrated solutions from a
36–46. service-centered perspective: Applicability and limitations in the
Nysveen, H., Pedersen, P. E., & Thorbjørnsen, H. (2005). Intentions to capital goods industry. Industrial Marketing Management, 39(8),
use mobile services: Antecedents and cross-service comparisons. 1278–1290.
Journal of the Academy of Marketing Science, 33(3), 330–346. Zou, W., Brax, S.A., & Rajala, R. (2018). Complexity and its dimensions
Payne, J. W. (1982). Contingent decision behavior. Psychological Bul- in the servitization literature: a systematic review. Spring Servitization
letin, 92(2), 382–402. Conference, Copenhagen 14–16 May.
Perdue, B. C., & Summers, J. O. (1986). Checking the success of
manipulations in marketing experiments. Journal of Marketing Publisher's Note Springer Nature remains neutral with regard to
Research, 23(4), 317–326. jurisdictional claims in published maps and institutional affiliations.

13