LoS DTT Cae

a. Justn

TlVEy Publishing
W29936

BYTEIQ: SERVICE Vs. PRODUCT ORSURVIVAL Vs.

DEMOCRATIZATION?

Digvijay Singh Bizalwan, Sumedha Chauhan, and Gita Bajaj wrote this case solely to provide material for class discussion. The
authors do not intend to illustrate either effective or ineffective handling of a managerial situation. The authors may have disguised
certain names and other identifying information to protect confidentiality.
This publication may not be transmited, photocopied, digitized, or otherwise reproduced in any form or by any means without the
permission ofTotheorder
organ1zation. copyright
copiesholder. Reproduction
or request permission ofto this materialmaterials,
reproduce is not covered
contactunder authorization
lvey Publishing, lvey by any reproduction
Business rights
School, Westerm
University, London, Ontario,.Canada, N6G ON1: () 519.661.3208;(e) [email protected]; www.iveypublishing, ca. Our goal is to publish
materials of the highest quality: submit any erata to [email protected].
Version: 2022-12-05
Copyright ©2022, Ivey Business School Foundation

In Mav 2019, the entrepreneurial venture BytelQ Analytics (Bytel) was formed in Bhubaneshwar, a city
in the southeastern state of Odisha, India. BytelQ offered data analytics and data science capabilities to its
customers through services such as identifying risk and revenueopportunities, building data marts and
dashboards, offering loan management and origination, collaborating with health watch start-ups, and
training working professionals.' BytelQ's founders envisioned revolutionizing the world of finance and
banking by exploiting the potential of data. They also wanted to transform BytelQ into a product company.
Arya Ved¡brata, the chief executive officer of BytelQ. orchestrated a series of crucial product development
stages witha team that created the company's first product, BankStorii, in December 202 1. BankStorii was
designed to aid financial customers to select the best banking products.This new product could also help
banking institutions make strategic plans to stay ahead of their competitors. Vedabrata faced the dilemma
of which direction to choose.

new financial
As BytelQ was closing in on the last financial quarter of 2020-2 1l and moving toward the
model for BankStorii. He had two choices. The
year 2022-23, Vedabrata had to select a suitable business
first option was to make BankStorii available to India's mass users for free througha website platfora. By
pursuing this option, BytelQ could aspire to potentially capitalize on network effects, spillover effects, and
innovative business models. The company could also contribute to the increasing need to democratize (i.e.,
make available to the public) financial technologies,' which Vedabrata and BytelQ's co-founders
profoundly supported. The secondchoice was to make BankStorii a proprietary software. BankStorii could
be licensed to an exclusive group of big corporate banks that, in turn, could use the product to help craft
crucial business strategies for the future. This option could generate almost immediate financial revenue,
which the young company desperately needed.
Vedabrata was acutely aware of his company's pressing problems. He had to ensure that BytelQ employee
salaries were secure, turn BytelQ from aloss-making to aprofit-making company, create astable revenue
1 Tim Stobierski, "What's the Difference Between Data Analytics & Data Science?" Business Insights, January 5, 2021,
htps:llonline.hbs.edu/blog/postdata-analytics-vs-data-science.
2Barnali Patnaik, "Everything You Need to Know About Platform Product Management," Mindtree, February 9, 2018,
https://www.mindtree.com/insights/blogleverything-you-need-know-about-platfom-product-management.
3 "Using
Fintech to Dernocratize Financial Services," Mckinsey & Comnpany, March 7 2017,
https:/www.mckinsey.com/industries/financial-services/our-insights/using-fintech-to-democratize-financial-services.

Jun 2024 to Dec 2024.

This document is authorized for use only in Dr. Justin Joy's DTT 24 at Christ University from
 W29936
Page 2

compromising revenue.
model, grow BytelQ to the next level, and democratize BytelQ offerings without to Should he take
Was it possible to accomplish all ofthesegoals? Vedabrata had a difficult decision make. opportunity to
the safe option of giving out immediate and sure revenue? Would that choice overlook an
create a new unicorn start-up and reach a valuation of USSI billion?

BYTEIQ ANALYTIcs

Vedabrata, Sabitra Panda, Amrit Yadav, and Diyaa Ray the BytelQ founders--had each left their near
perfect jobs in the data science domain (see Exhibit 1for an overview of the data science field and related
domains). They allpreviously managed data science programs in big financial banks such as Goldman
Sachs, Standard Chartered Bank, JP Morgan Chase, and DBS Bank. Their passion was in driving innovation
through data. Over the years, they had realized and mastered the craft of unlocking the potential hidden
within data that was originating almost every second, in almost every organization, and fuelled by
digitization and digitalization.
The four entrepreneurs met up to explore the question "What's next?" Data had transformed into big data,
characterized by the 4Vsvolume, velocity, variety, and veracity. It was time to exploit data stored in
organizations' databases and to build innovative and revolutionary data science products. With this idea
and mission, Vedabrata, Panda, Yadav, and Ray founded BytelQ Analytics. Vedabrata became the chief
executive officer and chief innovation officer; Panda, the chief operating officer and chief data and analytics
officer; Yadav, the product manager; and Ray, the quality assurer.
As with most new companies, BytelQ had a flat organizational
reported to the chief executive officer and the chief innovation structure: All execution-level employees
officer. BytelQ was also a bootstrapped
company, as the start-up funding came from the founders, who contributed whatever
turned out, BytelQ started to generate revenue from day one by offering data science they could. But as it
working professionals. The company expanded its training programs to other Indian training programs for
countries (United States and Malaysia). Gradually, BytelQ transitioned its activities to the states and other
innovation. The founders leveraged their more than fifty years of combined experience in the heart of data
data science domain to create a proprietary data innovation framework (DIF). BytelQ analytics and
could use
aid other companies in generating intellectual property in the form of data science solutions to the DIF to
resolve their
present challenges and to make better decisions for the future.
By the end of 2021, the company offered four comnprehensive data analytics and data
The first offering was data innovation as a service, which allowed companies innovation services.5
from their collected and stored data assets. This functionality was made to drive unidentified insights
DIF. The second offering was data science as a service, which helped possible by leveraging the unique
companies
science and analytics programs through an end-to-end partnership (see Exhibit 2 for the create in-house data
for performing analytics and data science operatiens). Data ongineering BytelQ tools used
services*
which assisted corporations in setting up the proper data éngineering infrastructure was the third offering.
choose from either on-premises or cloud platforms for storing data effectively and and helped companies
making
for consumption when required. The fourth offering was the data catalyst, which data available
a data-first culture in their organizations. helped businesses to instill

Allcurrency amounts are in US$ unless otherwise specified: Mihir Shukla, "The Dermocratization of Technology."Forbes, November
7, 2019, https://www.forbes.com/sites/forbestechcouncil/2019/11/07Ithe-democratization-of-technology/?sh=7c2a0737967a.
5"Offerings," BytelQ Analytics, accessed November 14, 2022,
https:Iwww.byteiganalytics.com/services.

This document is authorized for use only in Dr. Justin Joy's DTT 24 at Christ
University from Jun 2024 to Dec 2024.
 Page 3
W29936
BytelQ's frameworks, models, and
training modules were all made available
believed that a perfect blend of its offerings to its customers. BytelQ
could foster cultural change in companies
from data in their value chains (see by including insights
BytelQ'sprojects undertaken throughExhibit 3for a high-level view of the above
these services). offerings and Exhibit 4 for

DEVELOPING BANKSTORII
Even though BytelQ was already offering four services,
to restrict the company to the services domain. Since theVedabrata and the other co-founders did not want
company's founding, Vedabrata knew the real
prize was building and catering to analytics-based data products.
the next big idea to create data products from the ground up. The Accordingly, BytelQ was on the hunt for
new opportunity and to assess its commercial feasibility. founders worked as a team to explore any

The first phase of the hunt involved sitting together and brainstorming. From those
recognize that India hadimmense financiat banking data. When that data was alignedsessions,
with the
the team came to
numerous Indian
national and international banks, as well as the status of financial literacy and inclusion in the country,
possibilities could arise. The team could create and commercialize an analytics product based on banking exciting
data.
After several discussions, the team agreed to develop a product based on bank data with the potential to generate
valuable insights. They did some initial ideation and decided to validate their idea with a small pilot test.
Fortunately, testing solidified their idea's viability, so they moved forward with product development.

The next phase was data collection. They acquired publicly available open-source banking sector financial
data. The team utilized MySQL from their data science toolkit to create an initial tabular data lake for this
step. They added financial data dumped by financial corporations operating in the country to make the data
more comprehensive (see Exhibit 5 for the data variables and their categories acquired during data
collection). With this data gold mine in hand, the team started to investigate. They performed data pre
processing steps on the data lake using the programming language Python, including filtering any erroneous
data, checking for missing data values, removing outliers, and converting the raw and unstructured data
into a structured data format. After this phase, the team had retrieved more than sixty data variables.

The third phase involved applying algorithms to the data and building suitable models. In this phase, BytelQ
finally exploited its core data analytics expertise. First, the team divided the data into training, validation,
and testing sets. They then used all appropriate machine learning and deep learning models on the training
data set (see Exhibit6 for an understanding of the relationship between the two terms). Using the validation
data set, the team selected the models with the slightest error after numerous iterations of training instances.
Next, the team evaluated the selected models for overfitting, a situation where a model gets extremely
conditioned to its training data set. Based on the testing data set, the team gauged each model's performance.
For training deep learning models, the team used the open-source software library TensorFlow. At this
point, their approach began to make sense. When they conducted data visualization through PowerBI they
unraveled the stories that the data were hiding.
Also in this third phase, a potential product user could interact with the system through a user interface.
The interface was similar to astoryteling dashboard. The user could visualize the comparative performance
across banking products on an interface analogous to the Gartner Magic Quadrant (see Exhibit 7). This
extensive exploration led to a highly comprehensive banking data tool, BytelQ's first product. The team
was thrilled with their achievement. Vedabrata asked the entire team to name the product. Co-founder Ray
took centre stage and named it BankStori, asubtle play on the words banking and storytelling.
 Page 4
W29936
COMMERCIALIZING BANKSTORII
BankStorii was allset to be deployed in the market and shown to the world, but rollout was not
easy. So, Vedabrata and the team took a step back to comprehend what going to be
Essentially, BankStoriiwould make possible near real-time banking product BankStorii could fully entail.
operating in India. Yadav explained BankStorii's potential to Vedabrata this comparison across all banks
way:
Say, you have a banking product in mind-housing loans.
visualize and compare all housing loans products of all Using BankStorii, any individual can
information will mean that a customer looking for housingbanking corporations in India. This
rationally. At the same time, if any bank wants to loans can make that decision more
BankStorii to see which key areas it needs to improve its housing loan product, it can utilize
BankStorii will aid them in choosing the right improve and where it stands out. Effectively,
leadership or product differentiation. business strategy; say, whether to go for
cost
And when Vedabrata further
possibilities. Vedabrata needed investigated BankStorii's potential, the team
to answer the central revealed the product's enormous
BankStorii?" He was faced with two question: "How
broad options: the platform pathway should BytelQ commercialize
pathway (the product). (the service) and the proprietary

The Platfo rm Pathway

The first option was to make
would reap the benefits of its BankStorii immediately accessible to the people of India and the
available at no cost in the form comparative analytical capabilities. BytelQ's BankStorii
of a web product to anyone
world, who
connection. On the BankStorii platform, users with a smart phone or a platform would be
savings accounts, time deposits, credit could analyze any banking product (e.g., computer and an Internet
the Indian population owned a cards) and then reach a financial decision. At the various types of loans,
BytelQ's radar for releasing smart phone device." Quite visibly,
its platform, which would
time,
the Indian market was the first 80 per cent of
percentage of people also had a bank account, cover banks throughout the target under
according to The World Bank's nation. The same
In 2019., the National Centre 2017 Global Findex Survey.?
for Financial Education, a
create a financially aware and not-for-profit organization whose mission was
Financial Literacy and Inclusion empowered India, surveyed India's literacy levels. The centre's to
cent. Given India's population, Survey found that the financial literacy level for National
which
people. BankStorii could become the right was estimated at 1.6 India stood at
billion in 2022, this was a very high 27 per
look for financial platform for them and for financially number of
information onthe platform to increasing their
much like Facebook, Indians, who could literate
Google, Amazon, Uber, or Airbnb in financial awareness. BankStorii''ss operated
company provided a dedicated virtual terms of a platform-based business.
complete a transaction. This model space for two parties-the vendor The
would also mean that BytelQ might and the customerto meet and
become a fast-growing start-up,
6Tish Sanghera, "80% Of Indians Now Have a
Bank Account. So Why Is Financial
17. 2018, httpsl/www.business-standard.comlarticle/finance/80-of-indians-now-have-a-bank-account-so-why-is-financial
Inclusion Low?" Business Standard, May
7inclusion-low-118051700150_1.htrml.
Asli Dernirguc-Kunt, Leora Klapper,
Measuring Financial Inclusion Dorothe Singer, Saniya Ansar, and Jake
and the Fintech Hess, The Global Findex Database
2017:
https://openknowledge.worldbank.org/handle/10986/29510
8 National Centre for
Revolution,
Financial Education, Financial Literacy
(Washington, DC: World Bank, 2018),

https://www.ncfe.org.in/reports/nflis, 130. and Incusion in India Final

9Platform Hunt, "The 9 Types of Report, 2019,
Software Platiorns," Medium, June 12, 2016,
of-software-platforms-473c74f4536a. https://mediurn,com/platform-hunt/the-8-types

This document is authorized for use only in Dr. Just yat 24 at Christ University from bn 2024 toDer
Page 5
W29936

rather than a vanilla business. Promising

accelerators, sustained a growth rate of atstart-ups at the Y Combinator, one of the most
least 5per cent per week.l0 The tern "vanilla successful start-up
to a type of business that followed a traditional business" referred
products and performed a linear set of activities pipeline business model by converting raw materials into
throughout the value chain.
However, this option meant that BankStorii would be available without a fee or subscription
therefore, BytelQ needed to devise other wavs to generate revenue streams. One source charge;
of revenue could
be selling advertising space on the platform to other businesses. BytelQ planned to use Google AdSense,
Since websites in the finance category that reached 50,000 visitors per month in Asia and Pacific countrigs
could make S15,000 per year with AdSense.!" Although reaching 50,000 visitors was not an easy task,
Vedabrata committed to bringing a minimunm of 10,000 visitors to BankStori within its first week after
launch. 1This project gave BytelQ some idea of weekly revenue generation. However, the projected revenues
were insufficient for BytelQto quickly get out of the so-called "death valley curve"-atypical initial phase
in start-up lifecycle wherein the company is not able to generate enough revenue to sustain expenses and
further substantiate its operations (see Exhibit 10).

The Proprietary Pathway

BankStorii a subscription-based software-as-a
The second option, the proprietary pathway, was to makecorporate banks which could integrate the product
service (SaaS)product. BytelQ would offer BankStorii to
(see Exhibit 8 for the types of banks that operated
as a highly sophisticated strategic toolinto their arsenals
customer base for BankStorii in the years to come.
in India). The banks would become the potential
However, BankStorii was a new technology and
still not market tested, so only a few banks could use
Technology Adoption Lifecycle Model emphasized
BankStorii in its initial years (see Exhibit 9 for the banks would be
could analyze the whole market as a series of different types of banks. The
here). BytelQ bank would have
acquired (one after the other), ranging from innovators to laggards, so each
sequentially
technological assimilation.
a different adoption attitude towards
innovators just after launch; the first half of
the early adopters in the
However, Vedabrata was sure to gain the to twelve months after launsh.
months after launch; and the second halfof the early adopters in the six would decide future customer
first six would also get market tested, which
Within this time frame, the product business, developed as a
become a traditional software product number of licences given
acquisitions. This way, Bank Storií would The
paid an annual subscription fee for use. Therefore, if a bank had five
propriety software, for which the banks using BankStorii.
number of employees least ten licences to
to the bank would depend upon the buy five licences. BytelQ estimated that it could sell at
BankStorii users. it would have to bank in the initial years of the
technology
each innovator bank and fifteen licences to each early adopter ongoing
priced at $1,350 annually, based on expenses and BytelQ's
adoption lifecycle. Each licence wasfees, BytelQ hoped to add a combined software integration,
maintenance,
situation, In addition to subscription
licensing fee and charged biannually per licence.
per cent of the annual software for the
and servicing fee mnarked at 2 estimate the revenue generated from the potential revenue
streams
Therefore, Vedabrata was able to growth every six months going
forward.
and to calculate revenue
first month post launch revenuewould make
BankStorii, sell to banks, get the
pathway-develop
The idea for the proprietary business. So, BankStorii would be a straightforward
and effortless business
BytelQ a traditional pipeline
model for BytelQ to implement.
2012.
Combinator," Paul Graham, September
Startup? Get Funded by Y
Graham, "Startup =Growth: Want to Start a
o Paul
http:l/paulgraham,com/growth.html#f1 On.
h t t p s : / / w w w . g o o g l e . c o m / i n t / e n _ i n / a d s e n s e / s t a r t .

accessed November 10, 2022,

Page," Google AdSense,
"Home

2024.
from Jun 2024 to Dec
Dr. Justin Joys DTT 24 at Christ University
authorized for use only in
This document is
Page 6 W29936

Rising Analytics Market

In 2019, the Indian analytics market reached a valuation of $45.4 billion and a year-on-year growth rate of
26.5 per cent. The market was poised to grow at 21.2 per cent compound annual growth rate by 2026.
Analysts estimated that analytics could reach a 41.5 per cent market share of the total information
technology (IT) enabled services (ITES) sector by 2026, up rom 23.4 per cent in 2019. BytelQ had
positioned BankStorii in the banking, financial services, and insurance (BFSI) sector, which was 13.9per
cent of the total analytics market. Excluding the IT/ITES and Consulting sectors' market share, BFSi
contributed 34 per cent.

BytelQ was not the only company to see opportunity in these numbers, however. Analytics-based
companies had been forming in the Indian market since 2000. Some analytics company veterans Mu
Sigma, Fractal Analytics, LatentView, Absolutdata, Smart Cubewere also among the most prominent
business intelligence and data analytics players."From 2010Oh ward,new-age boutique analytics start-ups
like Tiger Analytics,Brillio, and Tredence started to make their mark in the subcontinent. These highly
specialized and flexible companies offered various solutions for tools and technologies such as data
analytics, artificial intelligence, machine learning. data visualization, and data management. These
companies catered to other businesses operating in sectors such as fast-moving consumer goods, insurance,
health care, retailand e-commerce, manufacturing, finance, and BFsI.
Analytics-based companies, and others, were BytelQ's competitors. Yet, Vedabrata was aware that even
though these companies held a prominent place in the BFSI sector, none offered a product similar to
BankStorii's. BytelQ's product uniqueness raised the risk of product imitation post launch. Contemporaries
were unlikely to just sit tight while BytelQ swept the market with BankStorii. Panda and Yadav were
confident, however, that since BankStoriwas a data-based product, the competition would not be an issue.
Their data sources, they countered,would become the competition's most significant hurdle to overcome.
BytelQ knew it could have potentially,.capitalized on this unique edge in the BFSI sector and create aniche
of its own. Nonetheless, the start-up success rate in the Indian market stood at less than Iper cent."
BytelQ had made considerable investments in terms of time and people to develop BankStorii. No matter
which of the models the team chose, they expected almost 60 per cent of BytelQ's upcoming revenue would
started feeling
come from BankStorii. Vedabrata, known for his impatience and ready-to-act demeanour,
restrained by the puzzle of which path to pursue.

WAY AHEAD
with fifteen employees and two more data analytics
By January 2022, BytelQ was still a growing company
situation was not as optimistic as it might have seemed.
products in the development stage. But BytelQ's managing everyday expenses was difficult. Though
Specifically, BytelQ was a bootstrapped company, so managed to earn
company was lucky enough to generate revenue, it was never profitable. BytelQ only journey
the operations. This phase of Vedabrata's
entrepreneurial
S67,500 in 2021, which could not sustain years, Vedabrata and co-founders wanted
to make BytelQ
BytelQ's success. In the next five analytics and
was crucial for in the
generating intellectual property through data needed more skilled
innovative companies in
one of the most goal a reality, Vedabrata knew that the
company
data science space. But to make this
employees and a positive cash flow.
India, May 17. 2021,
"Analytics India Industry Study 2021," AlM Research, Analytics in
Kashyap Raibagi, Learning. March 2, 2020,
Companies in India," Great
https://analyticsindiamag.com/analytics-india-industry-study-2021.

Busines5 Analvtics
Great Learning Team, "Top 10 Joumal, May 31, 2019,
Even 1 Percent," Froe Press
https://www.mygreatlearning.com/blog/top-business-analyticS-companies-in-india.

India Success Rate Not

* FPJ Bureau, "Startup
https//www.freepressjournal.in/india/startup-india-success-rate-not-even-1-percernt.

2024 to Dec 2024.

Christ University from Jun
authorized for use only in
Dr. Justin lovs DIT 24 at
This document is
Page 7
W29936

If BytelQ chose the first optiontheplatform business model-it

would give the company more time to
grow. But the team would need alternative revenue streamns even though this option would generate enough
revenue and growth over time to potentially lift BytelQ into the unicorn club. Moreover, this option could
help to democratize financial know-how and technical capability within India and eventually around the
globe. Vedabrata was also hopeful that this step could prompt other individuals, businesses, or institutions
to help BytelQ in its long but transformational path.

If BytelQ chose the second option-proprietary licensing to corporate clients--it would generate almost
instantaneous revenue-BankStorii's customers were an obvious and easy target. This revenue would help
BytelQ hire the right talent and establish the infrastructure necessary to grow the conpany. The BytelQ
team that created BankStorii could monetize their hard work starting in 2021, but the model's initial scope
of growth would be limited to afew banks operating in India. The founding team"'s dream to democratize
financial technologies would also be limited. 15
Vedabrata was in adifficult position: the future was an uncharted path. Both options had promise but not
without risk. Should BytelQ offer a platform or a product? Vedabrata was deeply contemplating this
question and hoping that the team would find an answer at the forthcoming meeting.

Business Review, July 27, 2018,

Science," Harvard
Democratization of Data
Cornelissen, "The
B/07/the-democratization-of-data-science.
Jonathan
15