Optimizing service at a Quick Service Restaurant

(BOX8) through data insights.

Submitted by:
Aman | BE/10224/16
Biswayanjit | BE/10598/16
Ishit | BE/10600/16
 ABOUT THE COMPANY

Poncho Hospitality Pvt Ltd also known as Box8 was founded in 2012 by two IIT graduates- Anshul Gupta and Amit Raj. The company
started as a small outlet in a corporate cafeteria. The main idea of the company was to serve Desi meals to people in a convenient,
easy to carry box.

3 pillars of Box8 which provides a delightful end to end experience to the customers every single time are:

● Desi Taste
● Technology
● Logistics

Box8’s full stack approach gives it a complete control over all aspects of food delivery from procurement to last mile delivery.
The most important forte of Box8 is its superfast delivery. The QSR delivers piping hot meals in under 38 minutes with no delivery
charges.
 ABOUT THE COMPANY

Box8 currently has 3 brands:

● Box8
● MOJO Pizza
● Mealful wraps

With around 120 outlets across 5 cities in India- Delhi, Gurgaon, Mumbai, Pune, Bangalore the company currently serves
around 60000 meals per day, putting the 3 brands together.
The QSR has recently launched a new category named ‘Insta Fresh’. Under this category the company delivers essential
items to its customers using the same supply chain network.
The company is also planning to launch 2 new brands in the coming days named- Itminan Biryani and NH1 Bowls.
 INTRODUCTION TO THE PROBLEM

The QSR sector in India is expected to grow at a CAGR of more than 18% between 2021-2025 due to rapid expansion of food
delivery services, expansion in young & working population as well as increasing urbanisation.

QSRs are the future of the restaurant industry because of the following reasons-

● Growing youth population who are most likely to spend quite a sum of their earning on good food
● Increasing per capita income
● Internet penetration in tier 2 and tier 3 cities

But at the same time, there are a number of challenges that a QSR faces primarily because of the fast and efficient service expected
by the customers. These challenges can vary from problems being as basic as providing the fastest last mile delivery compared to
competitors to being as complex as optimizing manpower and asset performance. QSRs also focus on identifying their potential
customers and ensure that there is a constant growth (expansion being the key to the business model).
 INTRODUCTION TO THE PROBLEM

Problems faced by a QSR-

● Manpower balancing and their performance optimisation

● Asset performance optimisation
● Ensuring fastest last mile delivery as compared to competitors
● Maintaining proper quality of food
● Driving customer loyalty
● Bringing down wait times / delay in delivering orders
● Maintaining margins
● Averting food / inventory wastage
● To ensure FIFO (First In First out) is strictly followed at the outlet for any product inventory
● Health and hygiene consciousness of employees (especially during COVID-19)
● Effective management of the supply chain
● Reduce the obsolescence cost
 LITERATURE SURVEY

Koranne and Borgave et al. (2016) This paper provides us with a study about the trends in quick service restaurant. The
authors have focussed on the fact that QSRs are the future of the food industry keeping in mind the increasing
urbanisation, rise in the disposal income in the hands of Indian population, growth of youth population in the working
sector who are ready to spend a good amount of their income on food as well as the internet penetration in almost all the
areas across the country. This paper also suggests that irrespective of the demographics of the market, there is a growing
preference of eating out or ordering food from QSRs. It has also been observed that with the fast changing world, the
family and working patterns have also changed and people actually have got less time for cooking at home.

Sakshi Kuchhal et al. (2015) This paper focuses on the organized food service markets- QSR. It gives us an idea about
the drivers of this sector which are rising disposable income and expanding middle class, changing demographics, health
and hygiene consciousness, localisation of menus. At the same time it highlights the constraints in this sector too. For
example: Lack of skilled manpower, FDI policy, Numerous licenses, price rise etc.
 LITERATURE SURVEY

Elgendy and Elragal et al. (2014) This research paper gives us an idea about the importance of data analytics in running
a business successfully and effectively. Data analytics has gained lot of interest because of its unprecedented
opportunities and benefits. We are living in an era where high velocity data are being produced daily. Within the data,
intrinsic patterns and details are present which needs to extracted and utilised. Data Analytics plays an important role in
decision making. With the support of data, we can actually set up new processes or re-engineer the existing ones. This
paper has suggested that data analytics is of tremendous significance in this era of data overflow and can help
businesses with unforeseen benefits.
 OBJECTIVES OF THE PROJECT

Through this project we intend to do the following:

● To ensure smooth functioning of the outlets along with focus on improvement of the overall performance.
● Control late orders in order to ensure customer satisfaction and enhancement in reputation of the company.
● Manpower and asset performance optimization are other major objectives.
● To ensure that SOPs are strictly followed at the outlets.
● To boost sales by identifying the potential market areas.
● Process adherence by the employees present at the outlet.
 Certain important processes:

● Late Order Analysis

● Roster management
● Optimization of the service areas

METHODOLOGY ● Manpower Balancing

● Inventory management
● Scanning Analysis
● Managing the front end operations efficiently
 LATE ORDER ANALYSIS

Orders placed by the customers get delayed due to different reasons. Delay in delivering orders affects the customer
experience in a negative way which further diminishes the reputation/ image of the company in terms of its service. One of the
most interesting fortes about Box8 is that it promises to deliver orders in under 38 minutes. In a highly competitive food sector
market, providing the fastest last mile delivery of piping hot meals is what gives a QSR some advantage over its competitors.

Therefore, to ensure that the end-end experience of a customer is delightful, late orders need to be avoided. In case there are
some, we need to analyze why were the orders delayed and suggest the operations team necessary changes which may help
us in avoiding late orders from the next time.

As we cannot expect every process to proceed ideally, some buffer is given to the outlets. For example, there may be
instances of heavy rainfall, bike issue of DMs, some cases when customer is not reachable, power cut at the outlet etc. which
are for sure to affect the service and would lead to late orders. Therefore, each outlet is given a buffer of 15% which means
that we can allow 15% of the total orders placed to get delayed. Anything above that has to be analyzed.

Due to various restrictions imposed during COVID it is difficult to ensure delivery within 38 minutes. But still we are operating
our outlets with a controlled late order percentage.
 3 major time spans are involved from receiving an order to delivering it to the
customer. These are:

· Make Time – The time that is required to prepare the order. (receiving time to
prepared time)
· Wait Time – After the order is prepared, based on the availability of Delivery
Boys the order is dispatched. So this is the time span from when the order is
LATE ORDER marked prepared to when it is dispatched on being assigned on a delivery

ANALYSIS boy’s name.

· Transit time of DMs – This is the time taken by the delivery boy to collect the
prepared order from the outlet and deliver it to the concerned customer.
(dispatched to delivered)

Spike in any of these may result to late orders.

 LATE ORDER ANALYSIS
This is a sample late order mail that was sent to one of our outlets on 12th June.
 LATE ORDER ANALYSIS

To reach to the conclusion about why the orders were delayed at the
Sheet 1
outlet, we had to write some SQL queries, import the data to a google
sheet and then figure out what went wrong.

Sheet 1- It gives an overall idea about how many (number and %

of late orders), when (session wise late order i.e. lunch, dinner
and snacks) and why orders were delayed (Make time, wait time
or transit time).

Sheet 2- Gives us an idea about the check in and check out time
of the employees at the outlet. It also displays the orders
delivered by each delivery boy.

Sheet 3- The DM count graph gives us an idea about whether or

not DMs (Delivery Master) were present at the outlet in a specific
hour.
 LATE ORDER ANALYSIS
Sheet 2: Sheet 3:
 ROSTER MANAGEMENT

Rosters play a very important role in an organisation. Manpower management or manpower performance optimisation is
totally dependent on the rosters we set.

Roster basically refers to the list or plan showing turns of duty or leave of individuals or groups in an organisation.

By setting up proper rosters, we need to ensure that maximum employees are present at the outlets during both the rush
hours of the day.

So what are these rush hours?

Box8 outlets operate from 11am to 1am. Orders are received almost in every hour but there are some specific hours in
which order flow is very high. These are 12pm-2pm during lunch session known as lunch rush hour and 7pm-10pm in the
dinner session known as dinner rush hour. Rosters are set such that maximum manpower is made available during both
these rush hours in order to ensure an efficient service from the outlet end.
 ROSTER MANAGEMENT

Most of the employees working at our outlets are from Bihar, Bengal, UP and NE states. After the lockdown was imposed a
major share of these employees returned back to their homes leading to an overall manpower shortage. With an overall
negative manpower it was very difficult for us to handle the outlets. That’s when we came up with the idea of break shift.

Break shift is basically a concept where the employee reports at the outlet, manages the morning and lunch rush hour, then
goes to the staff room situated nearby and comes back after minimum of 3 hours of rest (maximum 4 hours). He/she can
again work in the dinner rush and can even be present at the outlet till late in the night. For example in shifts like
11am-4pm and 7pm-1am.

The outlet has the least rush between 4pm and 7pm. It can even operate with minimum manpower present during this time
period which is why giving break shift helped us in utilising our manpower much more efficiently.

But we had to also think of a new incentive structure for employees on break shift.
 ROSTER MANAGEMENT

Here we have given break shift to some employees in

order to ensure that maximum manpower is made
available in both the rush hours. Week off or half days
have been given to most of the employees keeping in
mind that they also need to rest at least one day a
week to perform with maximum efficiency.

Besides, giving week offs / half days is also essential

from a company point of view. We need to pay the
employees some extra amount (paid offs) in addition
to their salary if they work more than 26 days a month.

So roster management is an important part to optimize the

service in which we take care of company’s margin in
addition to ensuring efficient functioning of the outlets. Attached here is the roster for one of our outlets
 OPTIMIZATION OF SERVICE AREAS

Each outlet serves certain specific areas which is generally in a radius of 5-6 km from the outlets. This is so because we need to ensure that
the delivery boys are able to reach the customer at the earliest, once orders are prepared. But, there are certain outlets which serve even up
to a radius of 10-11 kms. This is purely based on sales an outlet receives from neighbouring areas, keeping in mind that the service also
needs to be in control.

One of our major achievements during this project was that we introduced the concept of time based active outlet service areas.

One of our outlets in Gurgaon named DLF Cybercity was facing high late order percentage problem for quite a long time. When analysed properly it was
observed there’s an area on the other side of NH from where orders were received in good numbers regularly which were delayed due to heavy traffic during
the night rush hour. At the same time, it was seen that the lunch and snacks session orders from the same area weren’t delayed. The concentration of
orders were higher during lunch as compared to dinner. Therefore, we created 2 different service areas for the same outlet. But the area from where orders
were delayed was deactivated at around 6:30 pm keeping in mind the traffic issues which were there only during the dinner session. The outlet now
operates with a controlled late order percentage.
 OPTIMIZATION OF SERVICE AREAS

The points that can be seen in the picture

above are in the form of a polygon.
These basically are the coordinates
(latitudes and longitudes). The brighter
white dots are dragged rather than the
darker ones to change the service area
of an outlet because on dragging the
darker dots further points are generated.

This same area is being fed on all the 3

active brands and Zomato as well.
 MANPOWER BALANCING

Balancing of manpower is very essential for ensuring

smooth functioning of the outlets. Orders are delayed
at some outlets specifically due to shortage of
manpower or presence of not so experienced makers.
This is when balancing comes into play. Delivery boys,
experienced makers or managers are transferred from
outlets having excess manpower. Manpower balancing
is done so that we are able to utilise our overall
manpower effectively.

Each maker and DM has his or her own rating

based on which it is decided how experienced and
efficient the maker is.
 SCANNING ANALYSIS
❖ Scanning process ensures that the packets that are being used follows certain rule, to ensure that the customers doesn’t get stale food,
by making the outlets adhere to FIFO rule. For streamlining the entire process, we assign a unique code to every packet which we call
UID. Now with this unique ID and the entire scanning process which outlets have to follow while receiving and opening that packet, we

can keep track of every packet.

❖ Whenever the outlet will try to open a packet of new manufacturing date while he has another packet of the same inventory with an
older manufacturing date, he will get a pop up, saying please open the packet with the older manufacturing date, if the outlet decides to
open it anyway, it will be counted as a FIFO Mistake and the outlet will be penalized for the same.
❖ Let’s say if a packet is expired, whenever the outlet will scan that packet to open it, the system will prompt him that the packet is
expired and it will be wasted automatically.
❖ Now, for all this to happen outlets have to strictly adhere to this scanning process. So, we started off by, making a framework which
shows different data points of this scanning process. The follow up was done on a daily level, by our data POCs.
Scanning Data:
 SCANNING FRAMEWORK
The framework is live, and month till date data, can be seen on the dashboard. This helps the outlets manager to keep
track on their respective outlets.
 PRE-RUSH PLANNING (PRP)
❖ PRP stands for Pre-Rush Planning. Rush hour here means, the peak time of business, we have two rush hour slots 11pm-1pm and
7pm-10pm. It is done for certain items to save time in Rush-hours during live order preparation. It might also be done in some cases to
enhance the quality of certain items to provide better service to the customer.
❖ So, we basically have to do prediction based on past data of sales. We look at 9 weeks historical consumption data for every inventory
individually for each outlet, each brand and their level of sale. Using this, we calculate per rupees sales (PRS) for every inventory
outlet, inventory, brand, seller wise for the next 7 weekdays.
❖ PRS = (overall consumption quantity outlet, inventory, brand, seller wise)/(total sales outlet, brand, seller wise)
❖ From all the PRS inventory, outlet, brand, seller wise we select the median, median+1, median+2 PRS depending on recent sales trend
of that inventory. Then the respective PRS gets multiplied with the sales prediction of that weekday outlet, brand, seller wise to get the
predicted quantity for that inventory on that weekday in unit kg/pc. This predicted quantity is for the whole day
 PRP (Continued…)
❖ In this system, the query checks every 10 mins if the stock is sufficient or not. So here we have a Check till Time and create PO till
Time. Check till time refers up to the time the query will check the current stock is sufficient or not at that instant.
❖ PO amount check = Estimated Consumption Left till Check till Time +

Total Pre order Quantity + Threshold Quantity of that Inventory

(Below which it will be OOS) - Current Usable Stock

(total stock amount within its shelf life) - Total Existing PO Quantity

- min(Current Day Current Hour Consumption till now, Hourly consumption

from past data)

 PRP (Continued…)

● If the PO amount check is negative then PO will not be created. But if it is positive the PO will be created using
the formula: -
● Min. PO amount reqd. = Estimated Consumption Left till PO till Time + Total Pre order Quantity + Threshold
Quantity of that Inventory (Below which it will be OOS) - Current Usable Stock (total stock amount within its shelf
life) - Total Existing PO Quantity - min(Current Day Current Hour Consumption till now, Hourly consumption from
past data)
● Every inventory has a specific batch size mentioned according to the recipe. The PO amount that will be created,
will be the next nearest batch size possible.
● We also have primary slots for each inventory, it’s that time at which we check the consumption for the
maximum time, we do this to avoid the POs to be created during rush hour.
● An initial study shows that our inventory management project had tremendous impact in the first week itself.
Overall numbers of Product OOS has improved and sales loss has been minimized.
 RESULTS AND DISCUSSIONS
● When we started doing late order analysis for PAN India outlets, around 30-40 outlets had late order percentage more than 20% daily.

Now it has come down to 3-4 outlets with a percentage of more than 15%.

● An initial study shows that our inventory management project had tremendous impact in the first week itself. Overall numbers of Product

OOS has improved and sales loss has been minimized. (as compared to the earlier weeks when the project was not in place)

● FIFO was violated at the outlets earlier when scanning analysis wasn’t started. After having kept a track on the figures for 6 months now,

we have noticed significant improvement in numbers (as compared to 6 months back) related to manual opening of packets (25% vs
6%), packets opened after 1am (12% vs 1%) and packets left unopened at closing (4% vs 1%). We have also started a pilot project on
automated closing at some of our outlets in Delhi.

● Effective roster management has helped outlets perform smoothly. Introduction of break shift has made maximum manpower available

during both the rush hours. Besides, from a company point of view, we were able to bring down our paid off (extra amount paid to
manpower for working more than 26 days a month) numbers from Rs. 15,00,000 to Rs. 2,50,000 in a span of one month.
CONCLUSION:
● Our end goal has always been delivering good
service to our customers, while minimising the
costs, thereby increasing the overall efficiency
of the system.
● Manpower cost and Inventory cost are two
biggest cost function of our company, so the
manpower cost was optimized by man power
balancing and roster planning. In order to
optimize the inventory cost, an automated
mail is configured for the outlets in order to
tell them how much quantity is to be wasted
for each inventory, depending on the time at
which it was prepared and shelf life of the
items. Track of every wastage entry is kept,
and outlet is penalized for any extra wastage
entry done by them. Any variance in closing
and livestock update is also monitored.
CONCLUSION:
● The overall manpower cost and inventory cost
has been optimized over time, and the overall
efficiency of the system has increased
eventually
● To improve the service and experience at
customer end, we minimized the late orders.
By making the outlets adhere to PRP and
scanning process, we ensured that customers
doesn’t get any kind of stale food.
● Manpower balancing was done to cut down on
manpower cost.
THANK YOU!