Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
Selling Shovels in the Fintech Goldrush | Nageen Kommu (Digitap)
387 Plays2 years ago
"In a gold rush, the best business you can do is selling shovels to the gold diggers".
In this episode, Nageen Kommu explains how his company, Digitap, provides the essential tools for India's fintech lenders, helping them analyze data to make smarter decisions in a booming market.
Nageen Kommu is the Founder and CEO of Digitap.ai, a bootstrapped and profitable B2B fintech company. In just five years since its founding in 2019, he has scaled the company to a revenue of
₹8.4 Crores (approximately $3.43M) and serves over
200 clients, including industry giants like KreditBee, Navi, and BharatPe. A graduate of NITK and IIM Ahmedabad, his journey is a masterclass in resilience and strategic pivoting.
In this conversation with host Akshay Datt, Nageen unpacks his journey and the secrets behind building a profitable fintech startup.
Key Insights from the Conversation:
- Solving for the Underserved: Digitap was built to solve a massive inefficiency in India's credit market, targeting the 400 million people who are "new-to-credit" or have thin files, making them invisible to traditional lenders.
- Alternate Data as the Engine: The company's core technology analyzes seven different types of alternate data—including bank statements, device data, and e-commerce activity—to build a comprehensive risk profile for borrowers.
- The Bootstrapped Anomaly: In a sector dominated by VC funding, Digitap made the deliberate choice to bootstrap. This instilled a culture of financial discipline and a customer-friendly,
pay-as-you-go business model with no heavy upfront fees. - Pivoting from Failure: Nageen's first venture, a sports-tech platform called Dribl, failed to gain traction and shut down after two and a half years. He leveraged this experience to identify a more lucrative market and launch Digitap with a clear, validated hypothesis.
- Human-Centric AI: Nageen advocates for a balanced approach where AI is a powerful tool to augment and amplify human capabilities, not replace them entirely, especially in handling nuanced financial decisions and building trust.
YouTube Chapters:
(00:00) Introduction: Selling Shovels in a Fintech Gold Rush
(02:40) The First Venture: The Ambitious Idea and Painful Failure of Dribl (Sports-Tech)
(11:49) The Pivot: How Failure Led to Finding a Real Problem in Fintech
(14:55) Building the First Product: Creating Simple Onboarding & eSign Solutions to Survive
(24:00) The Core Engine: Building the Alternate Data Suite to Assess Credit Risk
(25:50) The 7 Data Sources: How Digitap Analyzes SMS, E-commerce, and Location Data
(30:55) The Psychology of Data: Why a Customer Who Needs Credit is Willing to Share Information
(33:40) The Business Model: How Digitap's Bootstrapped, Pay-As-You-Go Model Works
(38:25) Growth & Traction: Scaling to 70+ Clients and Reaching Profitability
(41:00) The Future Vision: Expanding Beyond Underwriting into Advertising and Global Markets
Hashtags:
#FounderThesis #StartupIndia #Fintech #Bootstrapping #NageenKommu #Digitap #Entrepreneurship #VentureCapital #RiskManagement #AI #MachineLearning #CreditScore #FinancialInclusion #IndianStartups #AkshayDatt
Recommended
Transcript
Introduction to DigiTap and its Mission
00:00:00
Speaker
Hi, I'm Nagin. I'm the CEO of DigiTap.
00:00:15
Speaker
In a gold rush, the best business you can do is selling shovels to the gold diggers. In this episode of the Founder Thesis Podcast, your host Akshay Dutt is talking with Nagin Komu, the founder and CEO of DigiTap, which is tapping into the fintech gold rush we are currently witnessing
Nagin's Entrepreneurial Journey and Early Challenges
00:00:33
Speaker
in India. DigiTap is in the business of selling shovels to fintechs, helping them ingest and analyze data for better decision making.
00:00:41
Speaker
By using DigiTap, a fintech startup can get data about consumers on seven different parameters, which allows them to take smarter risk management and customer engagement decisions. In this conversation, Nagin talks about the journey of his first startup that did not find success and then finally discovering product market fit in his second venture. He talks about how they built the product to solve an industry-wide problem of making sense of customer data and how the moat has allowed them to scale up profitably.
00:01:11
Speaker
Listen on, and if you like such insightful conversations with disruptive startup founders, then do subscribe to the Founder Thesis Podcast and any audio streaming app.
Consulting Experience and Transition to Entrepreneurship
00:01:27
Speaker
I joined the UI as a consultant from the campus, again, because of my telecom expertise in Alcatel-Lucent. So they decided to put me into the telecom vertical itself. And then I started providing consulting or getting on to consulting projects in the telecom space.
00:01:43
Speaker
So we had like projects, various projects varying from sales and distribution in the telecom sector to revenue enhancements, to cost reductions, to customer experience, digital transformations. So work on all these aspects with various telecom operators, advising CXOs in these telecom operators in terms of providing them with certain advice and as well as implementing the advisors that we have given them.
00:02:09
Speaker
So I had the opportunity to work, not just in India. In India, GEO and Airtel were our major clients that we worked with. But apart from that, I also had students in Asia-Pacific, Southeast Asia, and as well as Middle East. To be frank, during my MBA days and even during UI, I never thought that I had the entrepreneurial spirit in me, that I really wanted to.
Vision for a Sports-Tech Platform: Dream11 Meets ESPN
00:02:29
Speaker
I always felt comfortable working for someone and always felt comfortable that there is a certain take-home that I'm getting, and I'm not risking too much.
00:02:38
Speaker
However, me and my friend with whom I did my MBA, so we were having a lot of discussions around the sort of experience that one was having as a sports fan, just to tell you I'm a huge sports fan myself. And so I follow Manchester United in football, a huge fan of them. Similarly, Roger Federer in tennis. And then of course, being an Indian, I'm a big Indian cricket fan, fan of the Indian cricket
00:03:02
Speaker
and I also follow a lot of Formula One as well. Now, this love for sports and my love for technology were the two things that actually made me take the plunge into the entrepreneurship. And this happened when I was having some discussions with my friend and we both then figured out that there was a huge gap in terms of how the experience of a sports fan in the online space was having and what sort of an experience a sports fan was having in the online space.
00:03:27
Speaker
So we decided to do something about it. What we actually then thought was that we will build a platform which combines sports and technology. And these are the two biggest passions of my life, sports and technology. And that's the main reason why I actually ventured into competition.
00:03:44
Speaker
What did you envisage? Was it like a social media platform? Was it a content business? What was it? So it was a combination. What we were trying to build was not just one platform. So if I have to just tell you in a nutshell, it was a combination of a Dream11, Facebook, an ESPN, YouTube, and a Sportskeeda in one particular platform.
00:04:04
Speaker
So what we envisaged was that not all sports fans are unique. So every sports fan has a certain emotional attachment to the team or to the person that he follows. Now some sports fans are very vocal and outspoken in nature where they try to express their opinions quite a bit on the social media platforms.
00:04:24
Speaker
Some sports fans are not that vocal. They take more pride in watching videos of their teams doing good, doing well. Some sports fans are more into analytics. They like more in terms of stats and analytics. Some sports fans also want to get involved in terms of trying to get the hang of things and they try to see how best we can. They also try to get involved in the sense that they feel that I also need to get that
00:04:49
Speaker
of winning and all that creating fantasies and all that now if you see individually there are different platforms to cater to these needs of the users creating fantasy teams and stuff like what we wanted to do was really have a amalgamation of all these elements in the platform but personalize it according to the need of the users or the fan if i come in then as a fan as my emotional attachment to my team i might have a different experience to when you come in on to that particular
00:05:19
Speaker
And that differential experience is what we wanted to achieve through our tech. So that was what is in a nutshell it was. So if I come in, if I'm a Manchester United fan, if Manchester United won that particular game, if I come in onto the platform, I will be shown a lot more videos of Manchester United because I'm kind of a person where I want to see again and again my team winning.
00:05:39
Speaker
If you are also a Manchester United fan, if your team won, but you are more outspoken in nature, so you will be shown more content around social media, what's happening, what your friends are talking about the game and things like that. So that is the main purpose of the differentiation or the personalization that we wanted to bring on the experience of the fan.
00:05:58
Speaker
And in terms of the revenue aspects of things and all, we were actually a pre-revenue. I had that particular point in time, but we also had a path laid out in terms of how we could achieve revenue once the users came out of the platform. Advertising lead revenue. Advertising lead revenue, broadcast, license-based revenue, subscription models, ticketing solutions.
00:06:18
Speaker
See, basically the thought process was that if the fan is there on the platform, then what are the fan needs, merchandising, subscriptions and things like that. So there were multiple revenue models that could have been possible. Why didn't it work out? So one, he estimated the market wrongly. So we started off with the football in India and there was a reason for that as well. So it's just.
00:06:38
Speaker
It's not as if we blindly started to do that. See, for that particular aspect to work, there had to be an inherent rivalry between the fans. And that's when your emotional attachment to your team comes out. If you choose cricket, and we did quite a bit of research going into the emotional aspects of the fans and all, right? In India, if you see, we are not fans of cricket, and there is a huge difference here. We are not fans of cricket, we are fans of the Indian cricket team. That's the difference, right? If you are fans of cricket,
00:07:07
Speaker
then you would have been even watching Saurashtra versus Goa in the Ranji Trophy. But you are not, you are just fans of the Indian cricket team. Right now, yes, there is a huge buzz being created by IPL, so you have fans of RCB or an MI that came in. But even if 10 years back, if you consider, we were just fans of one particular team and fans of those 11 players.
00:07:28
Speaker
Football inherently is not like that. You would find fans of Manchester United, fans of Barcelona, Real Madrid and there is a strong rivalry especially when we were starting off dribble, Ronaldo was in Madrid and Messi was in Barcelona so there is a huge rivalry between Real Madrid and Barcelona fans itself. So that piece of the underlying structure
00:07:50
Speaker
of a particular game was extremely important for our idea to kick in. Which is why we strongly believe that we chose the right sport but not the right geography. We actually overestimated the market in India. We felt that there are enough football fans in India for this idea to take shape. That was a mistake that we didn't. We tried to reach out to the various football fans through fan groups and all that. But we frankly did not get the level of traction that we wanted to have.
00:08:18
Speaker
Isemo also felt that you were trying to do too much. The way you describe your idea, it is Facebook, come Dream11, come Twitter, come Instagram. It just sounds so ambitious that there is no way a startup would be able to do all of this without raising a lot of money, hiring a lot of engineers and product managers.
00:08:38
Speaker
If you actually see the product itself, we were able to manage to do this. The personalization piece is something that we didn't have at that particular point in time. But that particular platform had everything that Facebook had. That particular platform had everything that Dream11 had. It had everything that a sports presenter like a sports kid, a news presenter like a sports kid had. So we were publishing news. We were publishing articles. We were publishing everything. And like you had a content team to publish news. Yes.
00:09:07
Speaker
And the advantage of working in sports is that there will be a lot of people who will be willing to work for free, who are actually the fans of the sports. So we had a lot of contents, we had articles. So from a product point of view, it wasn't challenging. What we, as I told you, if we had taken the right geography and put it in the right geography, we would have gotten a lot more traction on the platform than India.
00:09:29
Speaker
But there's nothing restricting you to India. It was, say, you would have two European football clubs, fans also on the platform that could also happen, who are in Europe. It could. So this is a digital platform. So yes, there is nothing restricting. There is no physical touch points here.
00:09:46
Speaker
But the platform initially needs a little bit of a marketing push because people need to know that there is a platform of this nature. The marketing push that is needed say in Europe versus what is needed in India was very different and that's where the funding that startup needed comes into play as you rightly said. So the money we needed was more on the marketing side rather than on the tech and product side because we did manage to do justice on the product
Pivot to Fintech and Leveraging AI-ML
00:10:11
Speaker
side.
00:10:11
Speaker
But if we had to go on to say Europe or South America, for example, or Latin America, we needed a lot of push in terms of the marketing. Now the second problem that happened with this particular idea was that this happened back in 2017-2018. We are actually looking out for some funding and at that particular point, none of the VCs, we would have spoken to about close to 120-150 VCs.
00:10:33
Speaker
firms, angel investors, ranging from angel investors to large VC firms. The common feedback that we faced was that they did not understand this industry. Had it been an ad tech or a fintech or a health tech, they said, yes, we understand that industry. Yes, we know what works in that industry, what doesn't work in that industry.
00:10:54
Speaker
to an extent that even in sports they only understand gaming which is your mobile gaming industry because that is the only thing that we have seen flourishing outside of India is the gaming industry or a fantasy industry. We did manage to raise funds, we managed about 100k USD, raise 100k USD which is an initial fund to just get us going but that was again from people who were passionate for sports followers not friends and family, it's actually from professionals
00:11:22
Speaker
So we had the CEO of shadi.com investing in our firm. We had partners from EY who knew me the kind of work that I did for them. So it was predominantly they invested on the idea and the team that we pay. So it was and it was from professionals. It wasn't from any friends or family sort of around. So yeah but that was a small amount of 100k where we tried to do a few things but it didn't work out and then we had to then close off on the area.
00:11:48
Speaker
Okay, I think 2019 is when you decided to shut it down. Absolutely, yes. So we've given about close to two and a half years to the idea. We just did not see the traction that we wanted to have. And that's where we've decided that the best course on action for us to move along. But what we did was that we figured out that there is a good core being that we have developed. It's really our goals, concerns.
00:12:12
Speaker
governance approach in AI ML space. So at that particular point in time, I was actually looking out for ideas to pursue in the AI ML space. Even during triple in the two and a half years that we ran it, I was taking care predominantly of the product and the technology side of things. So I got close to what technology can do, what was AI ML can do. So we eventually decided that after having multiple calls and multiple
00:12:38
Speaker
chats with my friends in the fintech domain, we decided that there are significant areas where IML can be successfully to solve some of the problems that the fintech players are facing. And that's when we started developing certain solutions for them. And then since then, we have been working predominantly with a lot of these NVSEs and fintech players.
00:12:59
Speaker
So did you have some early backers in the sense that somebody who said, yes, if you build a set for me like this, I will buy it. What gave you the confidence to invest in building solutions for fintech? That's what I want to understand.
00:13:12
Speaker
Yes. So as I was having those discussions with some of my friends in the Windex page, one of the early backers that we had where, so the CEO of a company called Creepy, so he's a friend of mine. And so we had lengthy discussions with him. And then he said that if you select these sort of solutions, we would be able to then get those solutions. And then we have a lot of need for such solutions.
00:13:34
Speaker
So we were actually discussing a lot around how alternate data, like when we say alternate data, alternate to a bureau score would help in terms of underwriting individuals and what sort of a role can alternate data play in the underwriting of individuals. So that conversation started off there. So he was actually looking out for any sort of solutions that will provide him additional data on the individuals that he wanted to underwrite as on his platform.
00:14:02
Speaker
So we were thinking of setting data sources, how best we can mine those data sources, what sort of information can we get from those data sources. So the conversations were happening around in those lines. And to start with, we figured that those are really more complex solutions to build. And just to sustain us until those solutions become mature, we started in certain onboarding solutions as well, which would have given us some traction with that particular FinTech.
00:14:28
Speaker
will be able to sustain us for a certain period of time for us to be able to build more complex solutions. So that's how we started. It was always the intention of building certain more complex alternate data solutions. But we started building certain onboarding solutions as well to provide us with certain initial traction and money so that we could push the power zones and then provide or at least pay for the techies to be able to develop these more complex solutions.
00:14:55
Speaker
So what did you build? What was the first product you built? So as I told you, it was predominantly on the onboarding solutions. So the first product was on building an e-sign solution. Like how does an individual sign a document without having to print it, download it, print it and then upload it. There are other based technologies that you could use for him to sign a document online.
00:15:16
Speaker
and then onboarding solutions or other validation using paperless XML methodology and then solutions around developing Nash-based solutions. So these are some of the initial solutions that we developed because these were quite easy to build. They didn't have a lot of these complex AI, ML or computer vision sort of technologies to be needed. These were largely like workflow automation issues. Workflow automation, correct. These were the initial three solutions that we started off with.
00:15:45
Speaker
I want to ask a few questions about these solutions before we go to what you built next. Is there a law which tells what kind of an e-sign is legally valid in a court of law? Is there some regulation around it? There isn't. It's a grey area and I wouldn't call it a grey area also because a lot of NBFCs believe the ID Act 2000.
00:16:06
Speaker
A lot of NBSEs that we work with, they quote the IT Act 2000. What does it say about e-signs? It says that as long as the person has been authenticated through an OTP, he can sign the document. And whatever document has been signed by an authenticated OTP mechanism is a valid document.
00:16:23
Speaker
But that's only from an information technology point of view. So even in E-Sign, there are these two methodologies of doing E-Sign. One is getting an authenticated DSC certificate using your other credentials, which is a little more authentic and legally more valid. And DSC is what you need if you're a director of a company, right?
00:16:41
Speaker
if you are in direct traffic, correct. So any document that is signed with the DSC is a valid document. So even a director of a company can sign those documents on behalf of the company. Now the same sort of a DSC certificate you would get as an individual as well. You don't need to be a director using other credentials of yours. It's a more valid legally binding sort of a signing methodology.
00:17:04
Speaker
But a lot of BMCs and FinTechs that work in the market, they also quote that just by doing an OTP validation of the phone number, you are still valid in doing the signing of the document. And the major difference between these two is in the cost. An ABAR-based signing can cost you anywhere between 10 to 12 rupees. As an OTP-based one, you can do it anywhere with 1 to 2 rupees.
Legal Nuances in E-sign Solutions
00:17:27
Speaker
Why does an ABAR-based validation cost more, e-sign cost more?
00:17:31
Speaker
Because there are entities like UDI and certain e-sign service providers which connect to you, they have their cost of it. So it is a multi-lay solution wherein you'll have to first hit the e-sign service providers like NSTL and all. Then they in turn have to get the details and authentication done from UDI, which then is the cost to say tenfold cost. And what about these global solutions like DocuSign and all? Are they valid in Indian court of law? Questionable.
00:18:01
Speaker
Because DocuSign just makes you log in with your email ID. I don't think they do anything more than that. That's the exact point that I'm coming to. A lot of these contacts, they actually mention that any sign that is being done through an OTP, either on your email or on your mobile number, right, is a valid signature.
00:18:21
Speaker
Now post that what sort of a stamp you put on the document is up to you, whether you will just mention that this document is signed in a printed form by Nagin or you would want to take my signature on the screen and then you just have to put that as a stamp on the document is up to you. While see the main purpose of signature is basically saying that it is this person who is actually authenticating this document saying yes I am willing to do whatever is mentioned in this particular document.
00:18:48
Speaker
So those methodologies are plenty. One is either using other or the other is basically using an OTT based authentication mechanism. So all these documents and all these are also per se valid from that point of view. But we haven't seen any code providing a judgment on the contrary. So if there is a judgment that happened to the contrary, then there is a precedence that is set. And then people might then say all the other ways design is a valid form of reason. But we haven't had any judgments to the contrary.
00:19:16
Speaker
OK, so this kind of a simple sign has not been challenged yet, basically. So therefore, it's not been challenged. Got it. So I understand the service that you provided you integrated with an NSTL kind of a body, which is a e-sign service provider who further get from you the details and then that one time password is triggered to the user. And then that's how it's verified. And other verification is the same workflow. Somebody enters their Aadhaar number and then you would send an OTP.
00:19:46
Speaker
Correct. It is the same workflow. You get a lot more details about the person in terms of his address and all those details verified because you'll have to do the KYC for the platforms. The KYC just can be done through Adha because it includes address, date or word which are the main requirements. It is.
00:20:02
Speaker
So, you need a POI and a POA for any KYC. So, ARBAR forms your POA, which is your proof of interest. And a proof of interest is always a PAND card that they usually take. Okay, okay. And what happens in the NACH, that Nash Band-Aid? Nash Band-Aid is nothing but, so once you have decided that I'm going to take, say, 10,000 rupees from this particular platform as a loan, a platform might decide, depending on the riskiness of that particular individual, that I want to have a Nash setup done for this customer.
00:20:31
Speaker
Now what NAS is basically is nothing but the platform will ask the customer to enter into his bank account either by providing a debit card details or by providing net banking credentials and then give them the ability to deduct the monthly EMI amounts from their bank account without the intervention of the user. Say your monthly EMI is 1000 rupees for the next 10 months because you've gotten a loan.
00:20:57
Speaker
10,000. By doing the NASH, you are actually telling the FinTech firms that you are authorized to take 1000 rupees every month on this particular date from this particular packet. NASH also has two things. One is basically the mandate registration and the auto debits.
00:21:13
Speaker
So the mandate registration, what happens is you are actually authorizing during the mandate registration that you can go and take the amount from my bank and post that every month they then send something called an auto debit where they say, okay, on the 9th of September, please from this bank.
00:21:32
Speaker
So that also is from the customer's bank account and then gets transferred to the fintech spec. How do you earn? You again have a transaction fees, like every successful transaction you charge.
00:21:43
Speaker
Yes, so on each of these transactions, we charge a successful transaction fee. To some of the clients, we also charge a one-time sort of an integration fee that we also had to bring in the integration fee. Initially, we never had that integration fee. It was always a pay-per-use subscription model. But a lot of what we have seen is that a lot of companies integrate just for the sake of integrating it and then they don't use ourselves.
00:22:06
Speaker
So just to bring in that seriousness in terms of if you're integrating when you have to use our service. So if you tell them that the integration itself will cost you a lakh, for example. And if they really do not have the intention of using it, then they would not even spend time in the integration phase. Because not just for them, we also spend some time in the integration phase. We also spend our resources and all that.
00:22:28
Speaker
Yes, so those are predominantly the revenue sources for us. And for a company, is there an alternative to these products or is it build versus buy
Advantages of Third-Party AI-ML Solutions
00:22:38
Speaker
from you? What are the options for it? When you talk about most of these workflow related products, it's always a decision of build versus outsource.
00:22:46
Speaker
they can always build. So there's nothing stopping them from building them. In case of these alternate data sources and some of the data sources that we mine and then generate these scoring models to help the companies to identify the riskiness of a customer. So the major advantage or the major characteristic of those solutions is that the more the amount of data you have, the better the models will be. So if you try to build in-house,
00:23:14
Speaker
you will never have enough data whereas when a third party builds them they will have data from all the organizations that they work with and it will be a diverse data and then they will be able to address a lot of these edge cases and the models themselves will become that much richer when they work with that diverse set of data
00:23:32
Speaker
So my suggestion always is that when you look at complex AI-ML model-based solutions, the critical raw material for those solutions to work is the amount of data. It's always better to rely on a third-party vendor to get those solutions in place than these workflow solutions. Flow-based solutions are all more like you can build on. There isn't anything that is super critical. So tell me your journey after these three, like Isai and Adhaal verification and Nash after these three.
00:24:00
Speaker
Sure. So once we have built these, we then started looking at a lot of these alternate database solutions. These three started giving you revenue, like besides credit fee, you got on more clients also. Yes, we got on more clients. We started onboarding clients like Navi, Techologies, and Viteis, and clients like that. Then, of course, there was a major hurdle of COVID lockdown, one that happened in 2020 March.
00:24:23
Speaker
During that time also, we were able to focus a lot more in terms of maturing a lot of our solutions. So that is when actually we had a good amount of time because frankly, yes, we got hit because we didn't have any other clients, we couldn't onboard any other clients.
00:24:40
Speaker
But that also gives us an opportunity to focus on what next. Because what happens is if you have too many clients that you are servicing, which is one of the problems we are having right now, is that you always tend to address the problems that these guys are facing. And you do not have time to work on the next big product that you want.
00:24:58
Speaker
The COVID actually helped us in that because it gave us that opportunity to work on the next big product. So we've then started working on solutions like device analytics wherein we in the solution we actually have an SDK that goes into our client's app which reads that only the financial SMS's of the customer on the device and then we are able to generate a complete financial profile of the customer.
00:25:22
Speaker
Okay. You would be able to generate monthly expenditure because every time he's swiping his card, he's getting an SMS saying you spent. Not just on the card. So we'll be able to give you information on credit cards, CASA statements, CASA accounts he's maintaining, utility bills he's having, any investments or insurance he has bought, loans he has taken, wallets, wallet spends he's doing, what wallets is he using. So plenty of information of that.
00:25:48
Speaker
And we also started working on a few more of these solutions. So as a strategy, we started putting down what are the various data sources that you'd want to have. So we've identified seven data sources that you'd want to look at for underwriting a potential customer. So the number one is the bank statement. So if a customer provides us with a bank statement, then we pass through the bank statement and then are able to generate a lot of information for our clients.
00:26:16
Speaker
So this is like, it could be an image or a PDF file, you will parse that. It could be an image or a PDF for, yeah, we will parse it. Then device analytics that we just spoke about. So the second source of information was SMS and soft device data. The third source of information is employment data. So we are now, we also have solutions where we will be able to tell if the customer is employed or not.
00:26:37
Speaker
just without any customer intervention just by taking his phone number or his name and company name. So those are the solutions that we worked on. Then the fourth one is basically look at his e-commerce data. Look at his behavior on Amazon, Paytm or Flipkart platform and then analyze how he's spending on those platforms, what sort of delivery addresses he's using on those platforms.
00:27:00
Speaker
just so that the clients can get an understanding of predominantly the current address of that customer based on the delivery addresses that do that. Then the fifth one is basically telecom data, which is basically validate that particular phone number, what is the name and address of that particular phone number as per the telecom information and things like that.
00:27:20
Speaker
So the sixth information is on location. So if we are able to get the location of the customer, then what does the location tell us in terms of whether he's a risk customer or a rich customer or a non-risking customer? Yeah, so these are the six sources of information. I'm actually missing one, but let me get back to that later. So we've added seven sources of data, and then we started getting what sort of information can we build from these. So that's down with the journey that we had.
00:27:48
Speaker
So this is like a comprehensive solution you sell to a client or a client can pick and choose any one of these. Client can pick and choose any of these solutions. Okay. The more data they want, the more solutions they will pick up, basically. Correct.
00:28:03
Speaker
nature it's not like a platform although we also have a platform that we are now building in wherein we can offer all these solutions in one go and then say as per the solution you are distinct your score is X as per the solution your score is Y so your overall score is say X and overall score we either recommend to that you provide him with credit or we don't
00:28:27
Speaker
Okay, so this is like a credit engine that you're building then? Correct. Okay, let's go a little deeper into some of these data sources that you spoke about. I understand bank statement, the customer will upload it. I understand device analytics also that you will be able to read SMS's on the device. How do you do employment verification?
00:28:48
Speaker
See again, employment verification, we do it through two sources. One is a UN as a source. So every employee who is a centroid employee, if he's getting spied, so he would have something called universal account number. So UN is one source where we identify that if he is having a UN account and what is a UN account and based on that we identify whether he is an employed person or not.
00:29:10
Speaker
So the UN database is searchable, like you can search on the UN database for phone number. There are certain trade secrets that we follow here, but that's predominantly the source of information. And the other source also that we rely on is the PF source, like whether this company has credited any PF to this employee in the last three months. Again, that information is something that we predominantly look at to identify the employment of the customer.
00:29:36
Speaker
Yeah, so the body running PF, they give you access to data or how does that happen? The data is available. So you can actually go and check the data and it is done in the public domain.
00:29:49
Speaker
Yeah. Okay. This was the fourth thing, employment verification. Then you said e-commerce spends. Okay. So that is predominantly what we do there is we ask the customer through SMS, not just through SMS. We explicitly ask the customer to provide his credentials on these platforms.
00:30:06
Speaker
He is willing to provide us his username, password or an OTP if he has logged in through his mobile on these platforms. So then we do this platform data and then we generate a lot of details in terms of what all has he been shopping? What audience has he generated? How frequently is he shopping? How much amount is he shopping for? And then we also get his addresses. What addresses is he using to get his goats delivered?
00:30:31
Speaker
From that point of view, we are able to then identify his current address as well, which is a major pain point for a lot of our customers, a lot of our clients that get the current address, especially when they are doing a digital business. They don't understand, they don't know what is his current address. So from that view, we are able to identify what is current addresses for those students.
00:30:53
Speaker
I'm just not able to feel that I, as a customer, would ever agree to sharing my password like this. This is, yes, you are right. Absolutely. You will be surprised with the numbers though. But see, we are not talking about customers like you and me. We are talking about customers that most of these NBSs and fintechs, which are customers that are not serviced by the backs.
00:31:14
Speaker
These are customers who are hungry for credit. They are good customers, mind you. So they pay well, but it's just that they don't have the opportunity to somebody coming in and providing credit. They don't have that access to that credit. These customers are willing to give this information in return for, say, a decent credit. And location data is basically the GPS-based.
00:31:37
Speaker
It is predominantly just by the GPS. So getting the location data is not at all a challenge, but once we get the data, then again through publicly available sources. So we identify, say, just to give you an example of Mumbai, so whether he is living or whether he is living in Colapa. There is a huge difference between a guy who is living in Dharavi versus a guy who is living in Colapa.
00:32:00
Speaker
So we have publicly available sources where we get certain details about the location he is working in or living, what are the rental rates there, what are the property rates. It is an commercial area or residential area. So certain information of that nature will definitely help in terms of underwriting customization.
00:32:19
Speaker
through all of this, you are able to give one single score to a customer or do you have like multiple scores that you give or what is the dashboard? So the dashboard looks like this. So we give them in terms of various variables, even in case of device analytics or a bank statement, we give them where it's like, what is his monthly, what is his, what is monthly credit? He's getting into his bank account. What is a monthly debt?
00:32:41
Speaker
Does he have any risk parameters like any credit card transaction declines? Any loan payments missed? Any credit payments missed? Any utility payments missed? So all this information is being provided as well. So not just one particular overall score and on top of it also provided the score. Is this product live? You have customers who use this?
00:33:02
Speaker
All the products that we have just discussed are all live. We have multiple customers using these products and some of the customers that discussed initially like Credit V, Navi, said they do use some of these products and we also have several other customers like Motilav Aswal, Mass Financial, Bharat Pay, some of these clients, Northern Arc, who are also using some of these solutions.
00:33:25
Speaker
Currently, we are, as I mentioned, building a platform which provides all the solutions in one book. We currently have a player who is using that to an extent, but we definitely are looking at building that as well. So essentially, for a lending company, you become the back end for them. Correct. Almost everything needed to do underwriting to onboard the customer, you can provide like a one-stop shop for a lending company.
00:33:53
Speaker
Absolutely. Amazing. Okay. And you spoke about how when you're looking at EIML-based applications, then data is important.
Ensuring Data Privacy in Model Improvement
00:34:02
Speaker
And when you work with the third party, then you get richer data. So does the data coming from one lender help another lender? Say, Credit B is sending you like, say, thousands of profiles on a daily basis. Does the data of those thousand profiles help, say, a Navi? Because the same person might apply for a loan to Navi also.
00:34:21
Speaker
No, so we don't do it at an individual level, saying that person has applied for a loan in Credit B and he has not repaid in Credit B, so we tell Navi that don't give a loan to this individual. So it doesn't work that way because the individual data is sacrosanct to that client.
00:34:38
Speaker
We don't deal in terms of cross pollination or cross reading of the data. That is, in fact, most of the data that is being taken from in these cases is being purged immediately for us. What will help us is only to make our models better. So rather than dealing with, say, two different models, one for a credit, one for a different company, we deal with the same model, which will enable richer and richer while dealing with the data from credit, and also while dealing with the data from some other credit.
00:35:07
Speaker
Moreover, it's not just the data that we are dealing with, it is also some sort of customization that these clients ask us to provide. So there are certain variables that some client might come and say, can you provide this variable? Now, when I make that variable available to other clients also, they also then see the usefulness of that. So that's some sort of industry best practices that then come in and then reside with digital. It's much more easier for them to get the rich data.
00:35:34
Speaker
Essentially, by virtue of having so much data, your algorithms become better, but the data itself is not shared. No, so the data itself is sacrosanct to that particular client. We do not even mix the data. The data is residing separately for each individual client. The data is being purged immediately in the cases where the clients are requested for it. Some clients ask the data to be residing there for a few days so that they will be able to do a dead reconciliation, but otherwise,
00:35:59
Speaker
Data is all perched. But there could be a plane building a shared database, like what's simple. Civil is essentially just a shared database, right? See, civil and all also became rampant in India only after legislation that came in saying that all the fintech companies or BFSI companies need this information with civil unless and until there is an external push of that nature or there is an internal understanding between all these organizations.
00:36:26
Speaker
So that yes, certain sharing of this information of that nature will help one another. This is not going to happen. Because every company feels that their own data is important for them and they are not willing to share that data with anybody else. Are there alternatives to your stack? One is of course build-in-house, but are there other software companies providing this kind of stack?
00:36:50
Speaker
to fintechs and lenders. There are companies which we encounter as our competitors in each one of these solutions that we have mentioned and each one of the sources of these solutions that we have mentioned. However, I have not seen a company that is doing everything in a wholesome manner. We are providing a platform of this nature, the credit engine platform of this nature, where we can plug in all the seven different data sources at once and then provide a comprehensive score on all the set of areas.
00:37:17
Speaker
Like companies typically build like a credit engine in house. Most fintechs, that's the USB of a fintech that they build technology in house. So you would also be building a credit engine for them or that credit engine is different from what you are doing. Which is the reason why.
00:37:33
Speaker
When you ask me the question of whether you provide a scoring model or you just provide the underlying data as well. We do provide a scoring model, but the reason we provide the underlying data is simply because most of these fintechs want to have their own create engine. So what they do is then take the underlying data like what is the salary amount coming from banks, what is the entry credits coming in from device analytics.
00:37:56
Speaker
What is the employment data coming in from employment verification? What is the plan? What sort of a plan is it easy on a postpad or a prepared plan that is coming from the mobile data, from the telecom data? All that information they can individually use and then build a real engineer of their own. That is the case with most of the big NBSes.
00:38:15
Speaker
Tell me what kind of traction have you seen? What is the kind of top line that you did last financial year? Or what is your target for this year or next year? Just to understand how you've grown. See, we started off with just one client in 2020. Now we work with 70 plus clients that are active on our platform.
00:38:35
Speaker
and some of the clients that I've told you are Paratpay, Navi, Cradby, Modula, Swal, MassFinance, some of these clients. Then, basically, in terms of the traction, so we are just to provide you a quick highlight or a glimpse of it, we are a profitable company on an evita basis. And we would be doing somewhere about close to 2 million ARR by the end of this financial year. So, do you need to raise funds or there is no need?
00:39:02
Speaker
Currently looking at, so some of our top clients provide most of the revenue for us. So at this particular point in time, we are just trying to balance that equation out in terms of having a better healthy mix of venue in terms of our customers. And then we look for further funding. And that funding would be for what? For going beyond India or for building more products?
00:39:23
Speaker
It will be understanding what's outside of India for the set of things that we have already built, and also understanding what can be built outside of India. It will predominantly be for building solutions outside of India. And how do you do your sales? Like how do you acquire these 70 customers?
00:39:39
Speaker
Predominantly, so we've initially started off with just two people in the sales team. So we are our 10 plus people in the sales team. We started off with our initial connects that we had, the two people that started off. And right now we have a big sales team. We rely on rules like LinkedIn and all to connect with our potential clients.
00:39:59
Speaker
And we also send out certain emails through Mailchimp and all those, all the regular B2B sort of marketing tools are being used. And in this particular financial year, we have focused a lot on branding as well. We've been attending a lot of seminars. We've been part of webinars and events that have happened, physical events, sponsor podcasts.
00:40:19
Speaker
Yeah. So this has definitely helped us in the sense that people started recognizing digital. So even before go and tell what digital is, they say, we already know we've been receiving new dealers connected in these events. We've seen webinars. We've seen your article that are coming in the media.
00:40:36
Speaker
So that helps in just getting the name out into the market. But then, obviously, the real piece of converting a particular client is understanding where the gap for that particular client is, and then providing him with what he needs. We differentiate ourselves among our companies, definitely in the support that we provide. So that is where the individual sales team, the one-to-one approach
DigiTap's Market Expansion and Future Strategies
00:40:57
Speaker
works. How big could GTAB become? What is the kind of addressable market here, if you can help me understand? What is the potential of the business?
00:41:05
Speaker
If you look at the sort of solutions that we are developing, currently we are looking at a very pigeonhole market where we are saying these solutions have a use case that we understand is there in the underwriting or BFSI market. But what essentially we are doing is we are gathering data.
00:41:22
Speaker
We are analyzing data and we are also planning and building products in that phase, in that particular direction as well as we are looking at understanding data to address advertising and personal opportunities to firms. So how do we build a platform that can help firms in terms of advertising to their customers? This is really a market history to be able to do customized targeted promotions and offers and notifications and so on. Yeah. Okay. Amazing.
00:41:51
Speaker
And that brings us to the end of this conversation. I want to ask you for a favor now. Did you like listening to this show? I'd love to hear your feedback about it. Do you have your own startup ideas? I'd love to hear them. Do you have questions for any of the guests that you heard about in this show? I'd love to get your questions and pass them on to the guests. Write to me at adatthepodium.in. That's adatthepodium.in.