Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
A Masterclass on Digital Lending for "Middle India" | Madhusudan Ekambaram (KreditBee)
363 Plays3 years ago
"I told students, 'Please break our system. If you break it, the product is free for you.'"
This counterintuitive approach reveals the genius behind Madhusudan Ekambaram's early strategy. Instead of avoiding failure, he gamified it, turning his first users—students—into a dedicated team for stress-testing and fraud detection, building an incredibly robust platform from day one.
Madhusudan Ekambaram is the Co-founder and CEO of KreditBee, one of India's leading digital lending platforms. He has scaled the company to serve over 70 million users , disbursing ₹21,000 crores in loans in FY24 alone. Under his leadership, KreditBee has underwritten over 40 million customers and achieved significant profitability, with profit after tax growing nearly 200% in the last fiscal year.
Key Insights from the Conversation:
- The Co-lending Model: KreditBee scales capital efficiently through a "skin-in-the-game" co-lending model, where its own NBFC takes a portion of the loan risk alongside partners, building immense trust in the debt ecosystem.
- Winning "Middle India": The company's core strategy focuses on the massive, underserved market in non-metro cities, with over 77% of its customers coming from this demographic.
- The Data Quantum Moat: The true competitive advantage is not just the algorithm but the sheer volume of data from underwriting over 40 million customers, which makes their risk models uniquely powerful.
- Regulation as a Strength: By operating with its own NBFC license from the start, KreditBee built a resilient business that was well-prepared for India's evolving regulatory landscape for digital lending.
Chapters:
- (01:34) - From Corporate Life to a Fintech Insight
- (06:28) - The "Zero-Cost" Go-To-Market with Students
- (11:26) - How We "Hacked" a BNPL Product
- (17:52) - The "Skin-in-the-Game" Co-lending Model
- (28:27) - The Fundraising Journey: From Xiaomi to Premji Invest
- (35:54) - Navigating the COVID-19 Crisis
- (45:45) - The Underwriting Moat: Why Data is King
- (54:59) - Customer Acquisition Beyond Google & Facebook
- (1:05:03) - Why Regulated Fintechs Still Need Equity
#KreditBee #MadhusudanEkambaram #FounderThesis #FintechIndia #DigitalLending #StartupPodcast #NBFC #Lending #RiskManagement #Underwriting #StartupIndia #Entrepreneurship #VentureCapital #Fundraising #Scaling #Profitability #BNPL #IndianStartups #BusinessStrategy #Leadership
Recommended
Transcript
Introduction and Inception of Credit B
00:00:00
Speaker
Hello everyone. I am Madhusudan. I am the co-founder, CEO of Kraid B.
00:00:33
Speaker
Bajaj Finance, the largest listed lending NFC in India, has a market capitalization of $55 billion, even though it owns just a small portion of the total market. This would have given you an idea of how massive the lending business opportunity is in India. Madhusudan E was working as an e-commerce manager for a global brand, helping them grow their online sales when he came to the realization that if consumers were provided easy access to loans, then it would help increase the sales of the store.
00:01:02
Speaker
This insight led him to quit his job and start a lending fintech company, Credit B. Credit B is one of the very few lending fintechs that also has its own NBFC license. This episode of the Founder Thesis Podcast is a masterclass in understanding unsecured lending to consumers. Madhusudan talks to Akshay Dutt about starting up, building a strong product, building both demand and supply, and about the importance of risk management in a lending business. Listen on to get educated about all things lending.
Career Shift and Global Insights
00:01:34
Speaker
Oh yeah, after my engineering, I think I was totally core into software development, R&D, telecom. That was my world. But let's say five, six years into that, I was more into product and business. So that was a switch. I think largely between 2013 till 16. That's when I came out to 2016 when I came out to start the startup.
00:01:52
Speaker
Prior to that, I was managing or heading this e-commerce vertical for a company called Huawei, which was like the brand e-commerce store. Let's say Samsung would have its own Samsung.com where they would be selling the products or Xiaomi would have MI.com and stuff like that. It was a similar brand e-commerce store, which I was leading for about 9 to 10 countries, largely in Southeast Asia, Middle East,
00:02:16
Speaker
and Latin America. So I was responsible for the business, P&L of the entire business and whatnot. So what was more interesting was the overall traveling what I was doing. I was literally living on my suitcase. I had to turn around to all these countries. Introduction to that different clients and different people, there were certain common
00:02:34
Speaker
factors what the customers are looking for and of course there was very reasonable specific things as well. So that kind of led me to understand the people requirements and how there are certain things that the customer experience they're looking at the quality of experience which remains the same. You grow any part of the world I think that's what is more appreciated by the customers. While there are region specific stuffs what one has to look into but there were very common stuffs which one has to do in their own businesses. So
00:03:00
Speaker
That was learning. It's easier said than you experience it, but those experiences can really strengthen your fundamentals. And that's where I took
Fintech vs D2C: Market Challenges
00:03:08
Speaker
this call off. It's now or never because I was almost like 30 years old by the time. I would imagine that this experience is essentially like running a D2C brand where you're looking at everything from generating demand to serving the demand through shipping and logistics and handling returns and all of that. I'm wondering why you chose FinTech after this? Why not look at a D2C player?
00:03:28
Speaker
How did the credit B idea originate? As I told, sir, this was 2013, 14, 15 and some part of 2016. During that time, not just in India, most of this Southeast Asian market or even in the Latin American market, the only way somebody could get credit was through the credit card. Okay, so the credit card was the only way what you can key in and then buy the products or shop online and you didn't have any other things.
00:03:53
Speaker
Sir, I started talking to many of the lenders with the overall view of trying to get them onboarded as an lending partner. So I could increase sales and whatnot. I was convinced with the idea because that will easily give you an uptick of 10-20% of your sales the moment the credit is available on the store.
00:04:09
Speaker
And now when I'm talking to all these guys who are bajaj finance of the world, who are there on the ground pretty strong and the real apprehension from all these guys is how do I underwrite people? How do I lend to a guy whom I have not even seen once? How entire this digitization of the process works and whatnot.
00:04:24
Speaker
The more and more I got into that, and I looked at the markets in US and in China. So this entire fintech was growing at a very astronomical space. That triggered me to studying a lot about the Indian market, how we could do this entire lending digitally. So that's where I started spending more time. We came up with a clear idea of solving two things
Innovative Lending Practices
00:04:43
Speaker
at that time. One is digitizing the entire loan process. Now, digital loans are the norm. I'm talking in 2016, where the process was still
00:04:51
Speaker
with a human intervention, there was a fixed model. Digitizing the entire process was one of the key goals. The second goal was interestingly using a lot of alternate data models to underwrite the customers. So what we saw then was the overall bureau penetration, like your great bureau penetration was roughly 20-25% towards the market. Rest, 75% of the Indian market are not covered by these bureaus. Therefore, you need to really work upon a lot of alternate models to solve for it.
00:05:20
Speaker
I think these are the two problems what we wanted to solve. Started building on this product and the product was an overnight success for us. Was it a B2B? Did you want to do it B2B? Like going to companies like Huawei and offering BNPL on their point of sale at the checkout? Or did you want to do direct-to-call?
00:05:37
Speaker
So even idea was always D2C. So it should be direct to customer. So we wanted to always empower the customer to either take a loan or shop online, or utilize that particular amount for his education. So it was always the idea was D2C. So we were never on the B2B model or B2B to see what one would say as a checkout financing model. I don't think that was the thought.
00:06:00
Speaker
See, end of the day, you own the customer, then you are the king. So that's how this particular market works. The more you are creating layers before you interact and claim that this customer is my brand customer, the interfaces are not really helping this particular thesis. Though that would be a cash burn heavy approach, right? If you want to like directly acquire customers, like the customer acquisition would need a lot of cash burn. Did you anticipate it would need a lot of cash burn? Were you prepared for that?
00:06:28
Speaker
you discovered. Let me give you some thoughts. 2018 is, although we started working mid of 2016, we conceptualized the product, we started running it on a very small scale. One clarification, this is like unsecured personal loan. That was the core product.
00:06:45
Speaker
BNPL is what we wanted to start with. BNPL would mean like at checkout, right? So were you doing that? Not necessarily. So now there are various versions of BNPL. For example, today, BNPL has multiple things. One is of course, what you talked about the checkout finance.
00:06:59
Speaker
The second is we have launched our card product. All right. So you provide a line onto the card and you put offers on top of the card and the card can be utilized. It becomes like an universal product for your entire BNPL problem. And you can go on to swipe the card offline, online, across any merchants. So that's a different way approach of the BNPL. So the card we launched is not a credit card, but very much acts as a credit card because there's a line that is there backing that particular card.
00:07:28
Speaker
And therefore, so one could use it across the merchants and whatnot. So various ways of approaching that. There's a third model as well, we'll discuss later in our discussion. But there are many models of approaching this entire BNPL piece. So you launched as
Targeting and Market Strategies
00:07:42
Speaker
get alone in five minutes. Something like that would have been the business. Since we were digitizing the entire loan process, that time, so there were many things that was not in place yet. For example, the KYCs, the overall or the digital KYC mechanisms were not in place.
00:07:57
Speaker
or if you want to set up the e-nash for auto debit and everything. So they were getting almost developed. Indestack itself was developing during that time. We, our intention was to ensure that ultimately we come out with a product which has a complete digital, but we started with some kind of a physical angle. So whatever the key way is. One day, something like that. Yeah, something of that sort. So this is how we started. And during that time, this was underwritten by whom?
00:08:24
Speaker
Were you underwriting it? So that was the entire idea. So we never acted as a marketplace or a lead gen model from the beginning. The idea was to underwrite because that was the, that's what we thought as a core essence of this business in terms of developing all those alternate models to underwrite because that was the USB of the business. And we started kind of, then you would have needed an NBFC license to launch this. Yeah.
00:08:48
Speaker
So initially we partnered with an NBFC and we got our NBFC and by the onset of the business we applied for our license and we got it in 2017 early. Early 2017 we had our NBFC. By that time the business was also just taking baby steps and within that time we had our own NBFC in hand. How did you acquire customers? Was it like digital? Tennis?
00:09:09
Speaker
Yeah, that was an interesting part. So what we decided was, so since we are doing it digital and we wanted to go this road, went to many bankers to discuss about the idea and whatnot. Everyone's thought was India is a company, India is a country, nobody is going to pay you back. You guys are going to bust all your saving, the money or bootstrap funding, what you guys are doing to your business and whatnot. So what we were trying to do was that we wanted to start with a very controlled cohorts.
00:09:35
Speaker
What I mean by control cohort was we did not launch in 16 and 17. We were not like pan India, pan customer, anybody can come and apply. That was not the approach, but we started with the student community. So we started with some of the employee community.
00:09:50
Speaker
Yeah, education and consumption loans. The students were actively buying laptops, they were buying, they wanted financing for their projects, then their semester loans is what we had done it. Along with the parent, we were like writing for a bike loan and whatnot. So what happens here is that when you go after that particular student community, so the idea was that the marketing was almost zero. I'll tell you the idea here. So all the students are present in the same campus.
00:10:15
Speaker
Okay, so go to a campus, then we used to have a campus ambassador or one of the student who come takes up the part time job with us. And he would go to, he would not to each of the student hostel rooms and then educate about what the product is and how they should go for and whatnot.
00:10:31
Speaker
That's how the thing started and your customers are always present in that particular campus and then you could always go and meet them, understand with them. And the best part of that particular community was we used to incentivize them to provide us with the issues on the app. Or if this app could be the data could act or any other issues that they can report, we were incentivizing them.
00:10:52
Speaker
And students spent a huge amount of time in testing out that particular product. We'll come out with various issues and say that, yeah, Abkhaz, it's a brick or can be broken. It hangs here. So that the student involved method really helped us in making our app much more robust before we took it to the outside market. All right.
00:11:10
Speaker
We were largely financing their semester loans, bike loans along with the parent and some of the consumption like what the laptops. So where was that need? The moment they come into the colleges they had this particular need and parents loved it because it was always offered as an EMI and that was pretty successful. Did you ask them to specify that I need a loan to buy a laptop or did you just tell them how much loan do you need and that money would get transferred to their account?
00:11:33
Speaker
But since it was student, so there was at the point of transferring cash was limited. So we never wanted to give out cash in their hands. So the idea was that we are tied up with Flipkart, Amazons of the world at the time, and student would come and then basically provide us with the details of this is a specific product, what he looks at. And that product is what we used to kind of procure from Flipkart and request Flipkart to even do the delivery. So we were financing Flipkart rather than financing the student directly.
00:12:00
Speaker
It was always towards that purpose. So I think that was pretty controlled way how we started this year. And because it's going to the campus, I think some of it is not too challenging. Going to campus, the pin code is very well known. I mean, when we started with all these, the metro cities, the elite colleges. So already there was a good amount of logistics to do the deliveries and whatnot. So that's all the product itself. There are so many ifs and buts. What if they receive a defective product, they want to return and all of that. So then they would come to you and then you would facilitate or how would all of that?
00:12:30
Speaker
So it was always educated to them saying that so they would not pick the product they wouldn't come to us and say that I need a let's say a little laptop so they wouldn't come and say that they would say that flip they would go to Flipkart and they would basically go to that particular product and they used to copy that link and then paste it on our app. It was a simple copy paste or share mechanism they used to share that particular link to our app.
00:12:54
Speaker
So, we would only go and order that specific item on Flipkart and therefore, it was always all the returns, refunds and anything that they were looking for was through the Flipkart and not with us. So, that education is pretty important when you deal with this kind of a B and P scenario because there's always an element of confusing between who is the lender, who is the e-commerce marketplace, where the model is.
00:13:17
Speaker
Yeah. Yeah. So I think that's what we did because by the step, they were actually like a hacked together BNPL approach. Yeah. So that's how we started. And this gave us enough time to experiment on our alternate data models because of all I'm trying to underwrite student was more like without any income. So there is no income and therefore now you are looking at the parent and then you are trying to underwrite them indirectly.
00:13:43
Speaker
pan of the parent and like income proof from the parents and stuff like that or like how would you what kind of what was your approval mechanism? Yeah, so it was always dependent on the family income. But what we were trying to achieve when we were doing the student was to basically tackle the fraud.
00:14:00
Speaker
So the fraud that could happen on the app. See, end of the day, it's an app interface where somebody comes and interacts. Your engines decide that, OK, you want to give a loan or not. Nobody is sitting and evaluating that. So now frauds can happen in terms of impersonation. I use somebody else's mobile. I use somebody else's data. And then I do it. So then how do you prevent that
Fraud Prevention and Risk Management
00:14:17
Speaker
kind of fraud? So it could be a simple impersonation. Second thing is a group fraud. So I collected pan across some people. And now I try to basically get into the system or hack into the system and then get a loan.
00:14:28
Speaker
So this is what we incentivize students heavily. You please break our system. You break the system and then the product will be free for you. So whatever you have got it, that will be free. So therefore, there was an expensive testing that happened on this one to mitigate the fraud. So that's what we were largely achieving. Again, let me tell you this. If you could mitigate fraud, digital fraud that could happen on your lending app, more or less you are underwriting 50% you are done with that.
00:14:52
Speaker
The rest of the handwriting is, once I have understood that, let's say, Akshay has come onto my app, the moment I just ascertain that it's just Akshay what is claiming to be, and the documents what is given does belong to Akshay, and let me now start the evaluation. That's 50% of the work. Any fraud happens before that. So that's something what we are trying to mitigate by this particular pilot, what we did with the students. Yeah. Amazing. OK. And how much did you, how many products did you lose this way?
00:15:20
Speaker
Not much. The student market was pretty... I wouldn't have a count on top of mine, but it was 0.5% was the entire loss rate, including the credit loss and whatnot, which was pretty beautiful. Yeah, really healthy in terms of what had happened. But it is a pilot, right? So there's only so many campuses, there's only so many good colleges. We are not like going randomly to any college and everything. So we selected a good amount of campuses where you have these kids with some intelligence and then they also get enough feedback. What was your connection strategy?
00:15:50
Speaker
Collections pay. See, it was largely that students paid back and we used to, anyways, always there was a parent who was a guarantor to the loans of all the semester loans or bike loans and everything. Although student kind of front faced as a user of that particular credit for his education purpose or for his vehicle purpose or anything. But ultimately the one who was responsible was always the parent. So therefore any collections was with the parent. Otherwise it really look wise to go to a student where he himself is not earning. Yeah.
00:16:19
Speaker
And they would do like a bank transfer. The parents would do that. Always. There was zero cash collection process. We never collected cash and they used to pay on. The idea was that largely what happened is it was directly paid by the students. They used their parents' debit cards and then they paid it and whatnot. The parent coming and paying by themselves was very limited.
00:16:40
Speaker
Although they used all the resources from the parents, but parents themselves were paying less because the student would go and introduce the app to the parent. That I want this. Yeah. Online payment and all of that. Correct. Correct. So in that case, it was always front-ended by the student, but backed by the parent. And you have to bear that 2% intercharge rate. If someone is paying back the loan through debit card, then the payment gets recharged of 2%.
00:17:04
Speaker
No, so what happened is the debit card prices were slashed heavily during that time as well. Yeah. And one more thing was since these were like the ticket size was hardly about like 10,000, 15,000 rupees, the EMI was less than 2,000 rupees.
00:17:18
Speaker
So the charges, whatever was applicable for is for the EMIs which were more than 2000 rupees. So those charges were pretty minimal at the time, so to be really worried about. And then even the bank transfer was one of the mechanism what they were doing it there also it was on the average EMIs were 1500, 2000 bucks, under 2000. So that was working pretty much without much of the extra charges which has to be bought.
00:17:41
Speaker
So now let's move forward to 2018. So by 2017, you have got an NBFC license, but how are you funding it? How are you giving out loans? Are you taking debt to give out loans? Yeah, I think the models have evolved. So it was a hybrid approach what we followed. There is a platform which integrates with multiple lending partners or NBFCs, let's say. So one of the participating NBFC is our own in-house NBFC.
00:18:06
Speaker
So this is a set of bunch of NBFCs, which are lending on that particular marketplace. And the marketplace is the one which owns the app. And that's how the larger idea was. So at that time we had one or two partners. We, the scale was pretty small. We were doing about 20, 25 crores or so on a monthly basis, which was largely funded by our NBFC. Then there were two more partners and we were doing the supply side of the stuff. And that's how the thing started.
00:18:32
Speaker
Who does underwriting then? If the NBFC is funding it, they rely on your underwriting.
00:18:38
Speaker
No, not exactly. So basically, the way it has to happen is it was a joint underwriting. So there used to be multiple contours and multiple, the underwriting guidelines, what NBFC would basically come out and then put it in place saying before we even start the engagement or saying that this is the set of customers, what we are looking at. This is what should be the hygiene factors. And the platform would basically deal largely with all the fraud
00:19:02
Speaker
issues. So the fraud, digital fraud things has to be solved by the platform. So that's what the platform was looking at. And then since it was a digital integration, so they had their own risk engine at the NBFCN and our risk engine basically, which is detecting the fraud and what, which would do the initial filtering and then push the cases to this particular risk engine of the NBFC they would give an approval. And then the loan is free. Although it happens, it happens totally real time so that the customer feels can also probably everything is seamless.
00:19:30
Speaker
But it's always that there is a specific digital parts of the issues are solved, mitigated by the platform. And the unwriting on the credit itself happens with the, of course, so the platform will pull the bureau reports and whatnot, and then pass it on to the NBFC. So they would look at the contours, what they are looking at it, and then they would say, respond with whether they want to go ahead or not.
00:19:50
Speaker
The Bureau report was also being handled by you and recorded. And therefore the interest rate would not be shown to the customer before he finishes submitting his KYC. After KYC only the risk engine will give an interest rate. No, we had kept it already pre-decided. So there were buckets of the customers who are created depending on the risk profile. For example, somebody has a Bureau score of 750 and above. Somebody has between 700 and 750. We don't deal with the customers who are having a Bureau score of 700 and below.
00:20:19
Speaker
We only deal with people who are 700 and above, so there are certain buckets. And then there are new to credit guys who don't have any kind of a bureau footprint. Then there is a region specific, there are metro cities, there are non-metro, then within non-metro, there is tier one, tier two. Then now it comes to the profession saying that salaried, self-employed, within the salaried, what kind of profession, then within the self-employed, what nature of self-employment. So based on all these things, there is a metrics based pricing, what comes out, basically what you call it as a risk-based pricing, right?
00:20:49
Speaker
This risk-based pricing is primarily agreed between the NBFC and the platform, what is involved. And this, the only thing is now what happens is to bucket that particular customer to that specific metrics where he falls into in terms of risk and then the similar pricing will go. So there is no price negotiation or price discovery that would happen on the spot there because this is something like it's a layout.
00:21:19
Speaker
Just let me add a little bit about it. So what happens today for us is our own NBFC does cool ending with the partner NBFC.
00:21:28
Speaker
So there is a platform which does all the fraud detection and whatnot. And post that, once the cases are passed, it is actually passed to two NBFCs. There is my own in-house NBFC, which does a co-lending along with the partner NBFC. In such ways, we have about 12 to 13 co-lending partners. It is not a platform participation, but they're co-lending with my own NBFC. So once the records files are passed on to, it is underwritten by two NBFCs.
00:21:56
Speaker
One is it is underwritten by my own NBFC and then in the partner NBFC, we jointly underwrite and we provide the loan. It can be even like a 10,000 rupee loan or 20,000 rupee loan, but about 80% of it will be funded by the partner NBFC, about 20% is funded from my in-house NBFC, but it's a co-ending arrangement. And again, there is a beautiful master guideline about co-ending practices by the regulator and that's what we follow there. So this gives NBFC faith because you have skin in the game.
00:22:25
Speaker
very much because it's jointly and written by two NBFCs. Therefore, we are taking the in-house NBFC takes the equal interest of whatever the other is taking. So there's a game at which the NBFCs are participating. But a connection is handled by the platform. The NBFC, the colander doesn't need to bother about that.
00:22:44
Speaker
it's very specific to each of the colliding partners. So for example, so there are NBFCs which are strong at a specific region. And so maybe the initial collections would be handled by my own NBFC. So largely what happens is the collections is also done by my own NBFC. So the platform is more of for us, let's say it's a lead generator for most of the activities of the leading KYC
00:23:06
Speaker
or even the collections happens from my NBFC because we are a colliding partner with other NBFC. And the partner NBFC, see for example, they're very region specific, regionally established NBFC. So in such regions, we leverage on their collection as well. So it's a, when you come as a colliding partner, it's not just on the credit front, but also on the collections and the operations front as well. And that's what, where we leverage depending on each of our strengths.
00:23:31
Speaker
Probably when it's in the 0 to 30 bucket, then you would do it. The moment it crosses 30, then you would pass it on to the partner who is strong. So pretty dynamic even in those aspects on how we deal with dynamic in terms of the products, what has been given. Every lender, every NVFC has their own preferences in terms of how they would try to get this engagement. We try to derive a certain standard sort, but yeah, it's specific because they also look at their own comforts there.
00:24:01
Speaker
And you could have also gone the other way where you just build your own book, take debt. You could, for example, pick up debt at 10% and lend it at 18% and that 8% spread is yours. So why not go that way? Why go the coal ending way?
00:24:19
Speaker
Yeah, see the idea here is that the way the debt ecosystem that works in India is collective confidence one has to get into. For example, if you are just, I'll give you an example, right? Let's say you have a single NBFC and then you're just raising debt. Any lender who comes onto your balance sheet lender. So he wants to look at how the company is performing, what's the process are running. So how diligent are they in terms of complying with the processes and whatnot. Yeah, they will do their own underwriting before giving you loans.
00:24:47
Speaker
Yeah, so now the idea here is that once what the leverage that the platform provides to us is that since we have a great name, so the names like Yara Capital, which is coming live on this one. So which recent announcement was Punawala Fincorp, right? So it's from the Punawala group, which went live on the platform. There is Chola Mandalam group.
00:25:06
Speaker
So when those guys are operating because on the platform, so they are party to all the processes, what we are doing it on a regular basis, they are party to the fund management, they are party to the credit cost, what is happening. So there's a good amount of authentication of not just the credit cost or one could state, but also to the processes, what is followed, how much diligently the company is working and whatnot. So that gives a pretty holistic confidence to any new incoming debt partner.
00:25:35
Speaker
End of the day, this business, one has to solve for supply and you have to solve with multiple NDFCs, banks, and the collective confidence is much, much higher than having a single book. We are not SBI, right? So we are not a brand like SBI. We are in the process of building a brand.
00:25:51
Speaker
And in that process, the confidence from the overall ecosystem should be pretty strong. And that only comes in once there is a good transparency in terms of the processes, what you are running, the way you're dealing with the business, which comes pretty strong once you have all these third-party people understand and then participate in your process.
00:26:12
Speaker
If you like to hear stories of founders, then we have tons of great stories from entrepreneurs who have built billion-dollar businesses. Just search for the Founder Thesis Podcast on any audio streaming app like Spotify, Ghana, Apple Podcasts and subscribe to the show.
00:26:33
Speaker
But how did you fund it? Even to run your own NBFC, you still need funds, right? So you're only doing 20% of the loan, but that 20% is also a big amount. So how do you answer that? We have 56 lead partners on my NBFC. We have banks which has lender, Bank of Paroda, S-Bank, HSBC, BIA Bank.
00:26:52
Speaker
any of the banks, and then there are corporate NBFCs, likes of Vivritti Capital, Northern, any of this corporate end, this one, AU Small Finance Bank, FinCare Small Finance Bank, right? So there's a bunch of those people, then there are corporate NBFCs. We issue bonds and debentures, what we call it as non-convertible debentures. We raise capital from that. We have ECB, external commercial borrowing, what we have raised. So this is a
00:27:15
Speaker
pretty hybrid approach what we follow in capitalizing our own nbfc and as well as multiple partners on the platform so as i told it's not that we participate alone it's that entire ecosystem that comes into play 56 net partners 12 to 13 of co-leading partners you would roughly say about like 70 partners or who are working on the supply side of trying to cater to the demand which is going on monthly basis so that ecosystem is the basis to do this business here
00:27:45
Speaker
You have not diluted anything so far. No, we did. Last round was our series B round. Last round was almost a 145 mil race by the likes of Premji Invest who came in, one of our investors. Then there is TPG backed. NewQuest was the investor, but it has been acquired now by TPG. We have banks. ICICI Bank is one of our investors. They're the equity investors.
00:28:08
Speaker
Largely, if you look at it, it's the PE funds and the large financial houses or even the asset management companies like Misubi are the investors. So, it's PE and mutual funds, banks, financial houses. So, these are the guys who are our equity investors and that's how we capitalize through the equity. Yeah. Okay. Okay. Tell me the whole fundraise journey so far. What was your seed round and whom did you raise it from all the way till the season has been?
00:28:30
Speaker
Seed round was pretty interesting. So there was this particular company which was doing lending. So they are US listed, NASDAQ listed company. And they came in and they wanted to look at a strategic partnership, which we said no. And they were looking like funding us to make us as a branch of their company and whatnot.
00:28:49
Speaker
But anyways, so we were like, we are not into that game. So we are largely looking to run with our own journey. So the seed funding started. We did raise about one and a half million dollars as a seed money. So somewhere. So we were too lucky with that.
00:29:04
Speaker
So these guys are Southeast Asian market guys who are listed in NASDAQ. During my business, I had contacted them to come and be party to that particular e-commerce, what I talked about. So then I was talking very heavily about how big is Indian market and whatnot. So they had an understanding that, yeah, this guy is pretty serious about it. And then he has done enough of work. So that came out about one and a half million dollars.
00:29:26
Speaker
that was our seed money, then it was followed by this MI Xiaomi. So MI guys, so what happened is at one point of time, we were one of the largest sellers of their phones on a platform. So we wanted to understand every other day it used to go to their
00:29:41
Speaker
warehouse and then say that we need more phones. And a little bit intrigued came, why do you need more phones after there's a business? Then they answered our models and they thought it's a very interesting model. So it was, they invested in us. So that was our series A kind of round. And that was about, I think we raised close to about six, seven million at the time.
00:30:06
Speaker
And last round was with PI, Premji, Invest, Motilal. And you also have to give exit to MI, Xiaomi people. So we gave them a total exit. Now it's a different cap table what we have. So that was the series B, what we did. And then we should be out in the market for our series C Fundrise. But anyways, this is how the journey has been. So your series C would probably be a unicorn round, right? Like you would already be like considering the last round was 145 mil. So the next round would obviously be like a unicorn.
00:30:36
Speaker
Yeah, so hope so. So I'm keeping my fingers crossed. Definitely. So we are looking for that. Yeah.
00:30:42
Speaker
Okay. Amazing. Okay. Yeah. Let's go back to the evolution of the business. Like 2018, you started doing deadening on the platform. By the time the product itself become totally digital, when it became, when we achieved the totally digital by piloting with multiple control cohorts, we launched it in 2018. It was an overnight success. So Kratby was launched and then totally overnight success that month on month, the growth was over a hundred percent. We were running out of money to lend. That was how the growth was. There wasn't enough money to lend out.
00:31:12
Speaker
So we were the early movers and then we gave up this product and then customers were the word of mouth was pretty high. So the growth was tremendous. So there was a very good early movers advantage for us at the time. So how did you fix that? So you had that kind of hacked together solution of where you placed the order. How did you fix that? Oh yeah. The way we did that was there was another model what we started at the time. We were giving either we were loading the money to the Amazon wallet of the customer or flip the wallet of the customer.
00:31:41
Speaker
or there was a gift voucher which they could, which was attached to their specific Flipkart account. So once they go to checkout by themselves, that gift would be that either the valid amount would come in or the gift voucher would come in. So they would never be able to end cash it. So this is how we had fixed the particular problem at that time. And also we started with the personal loan. All right. So these were the two products what we launched the trade be with. As I told you, they would be for salaried individuals.
00:32:06
Speaker
Salaried and self-employed. So working up to a line and salaried. So this is where we started it. So now I used to just keep, one would be like the full amount is transferred to the bank account. Like they apply for one lakh loan, they get one. And that was not for students. So by that time we were done with the student while we did the pilots.
00:32:24
Speaker
and where the product was there. So this was to the open market now, open market, anyone, everyone could come into the app and whatnot. That is how the app was prepared. All right. So otherwise I would say that if you had not gone through that journey of trying to pilot it with the control cohorts, do enough testing, it would have been just a cash would have been out of the banks from us.
00:32:44
Speaker
Mistakes would have been very costly at that stage. Quite a cost. Now, I'll give you an idea, right? The way we were lending at that time was that I used to only keep the next month salary of my employees as the savings of the company. Rest all money is out in the market. So whatever you collect, you just again lend out. In one specific month, there was this... Before the MI investment? Before the Xiaomi investment? Yeah, before the MI investment.
00:33:10
Speaker
So that's how we are doing it. At one specific month, one of the lenders had promised saying that he would give out some loan to our NBFC based on that. I had even landed out the salary what I had kept. That did not come in. The date it was missed by the lender. I was totally running out of money. Then the next month I had to kind of give a big lecture to my employees saying that, yeah, it was like fifth of the month.
00:33:31
Speaker
Generally, first is when the credit rate is over, everyone is now talking about Kia ora, and now I was in a situation where I have to say that business is doing great, we are doing great, but I've just landed on all the money. And then the employees and it was a small
00:33:47
Speaker
team at the time, we were just about 40-50 folks at the time. Everyone is talking within themselves. So now I had to do that and of course by 10th I was able to arrange the money again and then start the salary. That happened for one month and never did that mistake ever. So always I showed that you buy employees the same salary to savings miracle.
00:34:07
Speaker
like this kind of an approach is not able to justify or nobody will appreciate what you're doing. So anyways then the so once the equity rounds came in and then there was also the debt community started trusting us so there were multiple more partners who added and today I think we are on the reverse side of the maybe the supply is so sufficient for us that now we are still in the demand.
00:34:29
Speaker
We invested, what was your monthly dispersion? I think we had almost hit, I don't know if we had closer to 70-80 crores monthly or maybe 100, somewhere that. 100 would be the idea.
00:34:44
Speaker
capital efficiency perspective is phenomenal. Like with just one and a half million fundraise, you scaled it up to 100 crore monthly dispersal is amazing. It was at a very good scale. So there are multiple partners where we had raised debt, there was good amount of debt that was there. And then there are partners. I think somewhere that I wouldn't really, this happened in what, 2018. So when they came in, so that was the series A around what we concluded.
00:35:08
Speaker
Okay, okay. So how did the trajectory change after Xiaomi? What would the evolution post the Xiaomi investment? Oh, as I told, so post that I think they did both Xiaomi and Shunwei, they're investing now, right? I think both of them put together had infused close to about like 50 to 20 mil. So both MI and this one. So it was well capitalized. Now we had an equity of more than 100 crores. And we could leverage that further. So
00:35:34
Speaker
We were doing almost, so just before the pandemic, March 2020, we were doing about 800 crores of lending on a monthly basis. That's the scale what we were reaching. And then that's where we started with our series. So wherein all these people came in and then we went for a 150 million fund raise again that further capitalized the company and there was sufficient funds. Yeah.
00:35:54
Speaker
Okay. Okay. Like when pandemic hit, then how did that impact you? What was the pandemic? And by 2020, the product lines were the same. Oh yeah. Yeah. We were doing BNPL and the PS. So those are the primary products, what we were doing it. And the business was all good. We were profitable by 2019.
00:36:11
Speaker
Yeah, FY 19 we were already profitable, FY 20 we were profitable, but March, life was looking like a fairy tale story. Things are all in place, our credit costs are at control and whatnot. Then happens the COVID. So the March 23rd, 2020, when there was a total lockdown was announced.
00:36:27
Speaker
the exposure what got impacted for us at the time was the disbursements what we had done from September 19 till March 20 roughly about 4600 crores. So I come from middle-class family these numbers were still very pretty large right now that's the exposure what we have done in this last six months which was exposed to the covid
00:36:48
Speaker
and it was a nightmare. It was a state of confusion. One doesn't know what would be the state of this. Everybody, everyone else is confused and everyone else is just giving different theories. I was doing this kind of a, not exactly a podcast, but it was more of a panel discussion. So we all went into homes and then the only thing what we were doing was coming on to zoom and then starting talking to each other. I think the only discussion was, right? So everyone is given their own theories, their own strategies and then whatnot.
00:37:13
Speaker
I think it was the next repayments that started happening in the month of April which gave us the confidence we are saved because we didn't know how many people would turn out and pay back and whatnot. There were more we had to offer to our customers which is roughly about six months of holiday period in terms of the EMI payment which has to be offered.
00:37:32
Speaker
Now, despite we offering all that and everything, so there was a good amount of customers who came and started paying back. And that was a relief for us saying that we have done a good job in identifying the customer slightly. And that's how the things were. All right. So we ended up, generally we have a loss rates of three and a half percent. So on any steady state month, the COVID wave one, we had to take a hit of close to six and a half percent. So.
00:37:56
Speaker
roughly it doubled. It took some time to come to that 6.5% but so the impact as of today when I see it, it's literally a little bit doubled, close to double our usual delinquency. That wasn't too bad for us because there was enough profits what we had generated. We had a good capitalization of both equity and debt and the profits what we had created prior to that particular COVID was able to take care of all the losses what we'd incur because of this COVID impact.
00:38:22
Speaker
And in that way, once we started getting the sense of it, I think things were back into normalcy. Again, everyone started ramping it up. Then COVID-19 too happened. It wasn't as bad as wave one. So typical 3.5% went till about five, 6%, five, five and a half percent. That wasn't much. We knew the drills. We knew what happens. The customers will get their confidence back. They would come back after the lockdowns back to normalcy and that's what happened. And then they start paying back.
00:38:50
Speaker
That particular COVID wave 2 was almost negligible to know what we had. Wave 3 was zero impact, either from the credit cost perspective, business perspective. I think people have basically understood that these waves might happen, but business as usual or life as usual has to continue and not to react over dramatic over those issues. So I think by wave 3, there was a very mature industry, mature customers who are already there in the market. Okay. And have you launched a credit card product also?
00:39:17
Speaker
Yeah, so we launched last August and in this close to about 7-8 months of operations, we have already issued close to about 1.4 million cards.
Credit B's Product Success
00:39:27
Speaker
For the customers, monthly we are issuing almost like 250k cards, 250k new cards is what we are issuing on a monthly basis. So that product has really picked up pretty well since we launched.
00:39:37
Speaker
Tell me about how that works. How does the product work? This is like a say like any other of the new HVintech cards It's the same as that like a slice or all of these. Yeah, see the segments are different So the idea here is that see if you only look at the product we'd be a card product or the PL product I think these products are there for hundreds of years So at least the overall PL itself is there probably thousands of years the card maybe about like 50 60 years phenomenon so I don't the
00:40:04
Speaker
The real innovation is not on the product itself, what we have launched, but the innovation comes into how much of what kind of customers have you launched? So for me, my customer base is mid India. When I say mid India, more than 75% of my customers are from non-metro cities. When I say middle class, the people who have annual salary of about five, six lakh rupees. So that's the middle class, mid income, mid India, middle class. So this is a segment what we get it.
00:40:31
Speaker
Compared to many other competition, probably they are after the metro, salaried segments. So everyone has their own segments with which they are doing it. So the ones who are doing it for the salaried, metro, customer base, the focus would be very much on the convenience because they already have three, four cards from the banks. Now somebody is adding at the fourth card. So now you need to really put in a lot of convenience factor, differentiating factor and bottom.
00:40:57
Speaker
Now, when we come to my segment, more than 85-86% of the crowd or my customers never had a credit card in life. So I'm the first time issuer of a credit card. So that's a different ballgame because overall credit card penetration itself in India is about 3-3.5% maybe. And what I believe is that there is a good 12-15% of the population what can be covered with a credit card. And that's the opportunity probably all the card guys are after.
00:41:22
Speaker
I adjust that we have split our customers very differently that my course essence is solving for this particular set. So I'm after them. So somebody's off with the salary. It's that. So that's how the game is played out. Yeah. And for a customer, what is the experience? Like they do the K Y C, then you decide the limit. Like you're getting a card with 30,000 limit or 60,000 limit.
00:41:40
Speaker
Yeah. So again, the, we provide a credit line so largely and card is just a, it's a prepaid card. So it's not exactly a credit card, but in the hands of the customer, it is more like experience is like a credit card, but for sale, it's a prepaid card. All right. So with the creative backing, the customers.
00:41:58
Speaker
Why not? Because that means your money is blocked then. Even if the customer hasn't spent it yet for a prepaid instrument, your money is blocked with that customer. So the idea here is that the line gets that the card has to be managed entirely through your app. So therefore you basically load money through your app so that it's more so we on the fly or on demand is when the lines are created rather than you load it and always keep it alive.
00:42:24
Speaker
interaction between the card and the app itself. So that's as good as you're requesting for a personal loan or a kind of, and then that is available on the card. All right. So that's how that particular functions. And the interest becomes payable only once they transfer it to the card or once they swipe the card, like when do they have to start?
00:42:42
Speaker
There is an interest period for about 45-40 days. So if they are using that particular product, so our incomes are largely from the MDR incomes. We treat it as a personal loan only when they withdraw the cash. Otherwise, as long as they're utilizing it, swiping it or purchase, it typically provides them an interest period for about 40 days.
00:43:03
Speaker
30 plus 10 days is where the interest rate period actually lies. Yeah. Okay. Okay. But why not just issue a credit card only? Why is there a regulatory reason for that? Or that would be frictionless, right? For a consumer. So the idea here is that credit card itself is product what the banks offer it. And probably then in that case, the underwriting is from the bank in terms of providing that one.
00:43:27
Speaker
What we provide is a credit, which is underwritten by the NBFC. So that's basically why we do a prepaid card with the credit line offer and not a credit card problem. Okay. Got it. So there's no bank partner for this instrument.
00:43:43
Speaker
There is a bank partner. So this is a card which is backed by BI wallet and therefore RBL bank is our partner there. What is BI wallet? So there is a, the cards are basically attached to the BI wallet and it's a prepaid instrument. Okay.
00:43:58
Speaker
Yeah, pre-paid instruments. So that is offered by the bank and it's a co-branded card with RBL bank and the lines are already to that particular PPA validator. So this is the back end of how we deal with this. Okay. And it would be like a RUPE or a MasterCard or a Visa. Visa card. We are today on Visa network. Because you are underwriting. So therefore this is the way to offer this to customers. Whereas that traditional credit card needs to be underwritten by a bank. Correct. Correct. Yes.
00:44:26
Speaker
And also somewhere we want to have the control via the app and not just an independent product as a card. So therefore they both interact together well for our purposes. Okay. What is your dispersal split? So you have three products, as I understand you have the PPI wallet through a card is one product. Then there is a traditional personal loan product where money is transferred. And the third is the voucher, like you finance the voucher. So what is the split between
Underwriting and Data Utilization
00:44:51
Speaker
these three?
00:44:51
Speaker
75% is PL product about 20% 20-22% would be on the card product roughly about 2-3% would be on the BNPL product. Considering card is not BNPL but card also is serving for the BNPL but non-card BNPL would be about like another couple of percentage.
00:45:07
Speaker
And I saw that this is zero cost CMI for this e-voucher. So is that because the partners, say Amazon or Flipkart, they are... There are certain incomes where we make a par revenue. And this would be like three installments. We give you three installments depending on what's the ticket size of the voucher. There are people who even buy iPhones, which is a LAC plus product.
00:45:27
Speaker
And therefore, that has to be one split close to about nine months, 12 months. Otherwise, if the product ticket size is at about like 15,000, 20,000, it goes anywhere between three and six months. For us, the average is about six months. The overall average for our PL product or anything is about six months.
00:45:44
Speaker
Okay. So would you say that data is a vote for you? The underwriting. So the overall segment to which we underwrite. So that's the mode because we don't find a lot of competition with the segment what we deal with. Although there are, look at this proposition, right? So the moment you go to the salary, super prime, Metro customer, your CACs are blowing up. Your ability to price them is pretty low because there's too many competition and even the banks are offering multiple products.
00:46:11
Speaker
So what is easier there is risk management is pretty easy to manage in that segment. But the path to revenue, path to profitability is always a question. On the other hand, what we solve for the segment is quite contrary. So non-metro, self-employed, middle class. So that segment is almost quite a wide segment, almost 400 million people, what we estimate who are great worthy, whom can be served in that particular segment.
00:46:38
Speaker
But what that brings in the complexity, right? So the complexity of underwriting, the complexity of managing your credit cost and everything. So since we solve for the complexity, and then we have built these engines to underwrite them, almost we have underwritten over 40 million customers through these engines. And therefore the engines become more efficient, the more data that they are crunching. And that's the kind of mode what we have created in solving for that particular segment.
00:47:03
Speaker
What I would say is that lending as such is not a single app business or it's not a winner take all kind of a market because across the globe you would not find only one bank in one country. So there are always any country you take it, there are multiple banks. The reason for that is there are different needs what each of them serve. There is a different niche what each of the banks and that's a similar case in India as well. Although from a very broader perspective it might look like
00:47:26
Speaker
There's so many apps and then they're just trying to doing the same PL product. There's the same card product, but what everyone is trying to find is their own niche where they are solving. And that's the differentiator. Yeah. How does your risk underwriting algorithm work? What are some of those counterintuitive learnings you might have had about risk underwriting? Obviously I know a lot of it would be proprietary, but if you can share some stuff around that.
00:47:49
Speaker
No, I think so. Let me put it this way. Let's take that the more easy way of trying to underwrite was why not if there is a bureau score, which is one of the very key determinant of stuff. In your case, I'm guessing most of your audience would not have a bureau score, right? More than 90% of my customers have a bureau score of 700 and above. That's how it is. Due to credit, audience is just 10%.
00:48:12
Speaker
Yeah. 10% on my book. Yes. On my evening. So now what happens here is that even with the bureau score, so let's say I would give you one example, right? So the guys who are between, let's say some range, anything that we can pick it up something like 750 to 770 is supposed to pay back and behave better than let's say 700 to 730.
00:48:33
Speaker
So just by going by the logic of this course, what if it has reversed? So what if that is reversed for a case, what we find when we have dealt with this large community of people, about 40 billion guys, what we have underwritten, about 6 million of the guys have taken the products from us and those ranges have differed. So now the idea is that even though we take that, then there is a need for us to develop our own scorecard because even I don't want to blame the bureau saying that they are not doing a good job, but the idea here is that
00:48:58
Speaker
then nobody can give one single score for entire India. I can give a score for this particular people, but not the India is a very diversified dynamic ecosystem. You cannot put one score and say that these are all good. So for my customer base, the bureau score mapping actually is developed for I need to have a different scorecard, which kind of
00:49:18
Speaker
let's say I even have about 20 scorecards which makes the decision, I'm just talking about a bureau based scorecard and the need for it. Because these differences are there, I have my own scorecard which kind of further splits and then gives a proper range of proper behavioral based scoring to that particular person. So now once such things happens, these comes as a mode. So otherwise, anybody who can come in and then take 750 and then say that, yeah, 750, it doesn't work that way. So if it goes for that, probably the delinquency should be double for us.
00:49:47
Speaker
So you have this code from 1 to 100 or 1 to 800 or something like that. How does it work?
00:49:52
Speaker
See, I think every credit underwriting models have several scorecards. Now it all comes down to whether you want to just represent a single one score and say that this is the overall ultimate score. And it may not be necessary for a lender who is trying to give one score because we are not communicating that score as a result outcome to either the customer or to any external agencies. For example, like all the credit bureaus have to give one score, although they will have multiple scorecards.
00:50:20
Speaker
at their own end but ultimately they all put an effort to come with that one score. So what happens or what goes into the different scorecards are so there are something like the fraud basis scorecard, what we call it as an anti-fraud scorecard. So this has nothing much to do with whether a person is basically has eligible or not eligible it is to mitigate
00:50:40
Speaker
whether in digital learning what happens is you don't see the customer like how we are chatting together today we are seeing face to face I know that it's Sakshai he's responding to me so now it's a device and the customer who are interacting with each other now the customer may be
00:50:55
Speaker
totally different from what is he or she claiming to be or they may be the same person but the data may not be matching to what they claim to be for example I let's say there is no not Akshay but somebody else is claiming to be Akshay so that becomes one fraud there is impersonations that happen wherein somebody is trying to have access to for whatsoever reason they might have access to KYC's of some other people and they're trying to do that and generally the looking at the households sometimes a man of the family has access to his mother father
00:51:24
Speaker
who are a little bit aged, their KYC documents, or he might have access to his spouse, he or she might access to have the spouse data. They may not be totally fraud in nature, but the idea also comes into line that when I'm talking about my spouse, it's as good as I'm talking about myself. Overall understanding also will be there. So those are impersonations.
00:51:45
Speaker
When it comes to a group frauds where some of the guys get together and try to do the frauds, then there are software frauds or a bots which are used to do the fraud. So that basically has to be recognized early and then they have to be rejected or put into way. So that comes as an anti fraud scorecard. So that has nothing to do with your bureau based scorecard.
00:52:03
Speaker
Alright, so now the next thing comes is an income estimator, right? So there is, I'll just give you an example. Now for a salaried person, it becomes all the more easier to understand what is his income because there is a bank statement available and then you would have an idea about what's the salary that they're drawing because there is an input credit that happens to his accounts on all the timing and whatnot. But now you take a nature of a self-employed person. With the self-employed nature, what happens is that
00:52:30
Speaker
So with the self-employed nature, imagine this, there is a Kirana Dukhan. So the Kirana Dukhan has a lot of transactions happening via the cash. Cash is still a major economy in terms of transactions that happens in the media level. And now you have a lot of surrogate methods that one needs to develop to understand what could be the overall income and what not.
00:52:50
Speaker
So, there is no accurate way of trying to say probably sometimes a lot of this Kirana Dukan owner itself may not know how much was the input and output. He has a broader idea of how much the sales happened and what not, but they might not have had a ledger which is accurate and everything. So, everything comes as an estimation.
00:53:07
Speaker
Then income estimator is a model that comes out and probably gives a scorecard on what is claimed by a customer saying that let's say a customer comes in and says that he earns about 40,000 rupees on a monthly basis. Now the income estimator model has to be done to understand whether that how much are we off from our estimations from the surrogate methods and provide a scorecard saying that is my estimation closer to that the self-stated income of a person or not. So that comes as in that particular scorecard over there.
00:53:36
Speaker
Then, so these are various different things where the scorecards itself evolve. And then there are multiple things comes as a strategy scorecards, which takes two, three scorecards as an input and add few more rules or few more pointers to that. And then that becomes one of the strategy scorecard generally involve two, three more scorecards and then put its own logic or thing and provides a scorecard. So.
00:53:59
Speaker
Each of the aspects has to be defined. So first is on the fraud angle. The second is on the estimation of the cash flows. The third is on the ability to pay, how much is he kind of intended or whether he has enough of cash flows that he can pay back or not. So that comes as a ability to pay estimations. Then comes your, as I last time, probably we talked about a bureau based scorecard because
00:54:20
Speaker
the exact bureau may not be working for every specific set of the customers. Then comes the willingness to pay back, looking at his own repayments, both on us and offers. So how is he behaving on the, on our own platform versus how one is behaving on the outside platform. So that becomes as a willingness to pick a scorecard. So these are different scorecards that, that we, that always come into play. And those scores are largely used by these, what we call it as risk policies, which use that particular score number, and then
00:54:49
Speaker
feeds into a decision tree to either accept, reject or maintain the same credit lines what has been offered. So that's how our typical risk modeling or our risk agent works. And most of the fintechs would give you a simple and an idea about how this works as I stated.
00:55:05
Speaker
But what really makes a difference for any specific company would be while this is all theory. So these models scorecards only behaviorally, once you're fed with the quantum of data, right? So the quantum of data really matters. So you could develop a scorecard and probably are just tested out with 5,000, 10,000 customers. That may be a very limited data for these scorecards to start predicting versus somebody was tested with a million customers data versus with tens of millions, 30, 40 million customers data.
00:55:34
Speaker
So it varies totally. So that's the mode. The theory of developing the scorecards or the risk model can be given to you by any chief risk officers of any company or it can be taught in any of your data science classes. Somebody who has done his master's in data science would give you the similar outlay of how it works. But what really matters is how much of data have you fed into these models so that they are behaving and then predicting accurately.
00:55:57
Speaker
And that information is when it starts making sense. And that becomes a mode for the companies versus which is a grown up company versus a mature company versus a bootstrapped or a very nascent level of growing companies. So what I would basically leave you with is that any structured format of data is always important for underwriting models. And whether it comes as your main business line, main business functions, or you depend on somebody else to get that particular data, for example.
00:56:25
Speaker
Let's say there are even the pay later products. For example, there are Flipkart's pay later product or Amazon's pay later products. All right. So that totally rely on the customer experience on their platforms, how much they have bought, what are the returns ratios and what is the value of average value of their orders, frequency of orders and whatnot. So that data is all the way critical again to underwrite and they use that particular data to underwrite. Data is the king. So now it all depends on how you collect the data, but you cannot discount the kind of data what is available with each of these business models.
00:56:55
Speaker
you.
00:57:24
Speaker
What is the average ticket
00:57:26
Speaker
Okay, got it. So you told me last time that now supply is no longer a constraint and demand is a constraint. Wouldn't this mean that you should go below 700 scores? Like you currently told me you have a cutoff of 700 plus bureau score and only 10% of your borrowers are new to credit. So considering that you need to ramp up demand, wouldn't these
00:57:50
Speaker
metrics into change. I'll give you one idea. I don't know if I in what context probably I gave that idea of demand as a constraint. What probably I would have given a context is that we have solved for supply side. All right. So because of the overall platform side, what we are there. So any month we are ready to even serve the demand, what we are anticipating for let's say six months from now on. And that is the shape every lender has to be. I am not basically saying about the constraint on demand because India is a very growing economy. There's a
00:58:16
Speaker
huge pool of customers who are coming into the, it's largely an end population. So every time you have people coming into the economy who are financially can be given a credit and whatnot. And that particular thing, I don't see there is any demand constraint as such. But while the demand is there, unless you have a supply side solved, you'll never be able to meet such demands. So probably I would have given a context about how well or how prepared are we in terms of solving that supply side stuff.
00:58:42
Speaker
Now coming to your next question, why not below 700 kind of a folks and then so why only go for that? Let's say everyone has their own model of whom they are serving and whatnot. So we are largely fine to sell the new to credit market rather than to say the somebody who has already proven subprime.
00:58:58
Speaker
A new-to-credit is probably the guy who has even graduated from, let's say, IIT, IAM, and then today he has joined his job, maybe with the top four consulting firm or in an investment banking job, but he is also a new-to-credit. So versus somebody probably who has, let's say, because of his totally no income, not things, not so economically favoring things, he might not have ever borrowed itself. So all these guys come into
00:59:23
Speaker
new to credit. So there is a good possibility of identifying based on their future and everything to become the credit versus trying to solve for a subprime segment, which is where probably these are the folks who have defaulted largely and they don't have a repayment disciplines and whatnot.
00:59:39
Speaker
So that is what we are trying to avoid. Now in a country like India, probably there are like the bureau penetration itself is roughly about 30%. For the entire India population, I would say that there's at least 400 million plus folks who are eligible to get a credit and they have not got that or not established their bureau records because of various reasons. Either they are just joined as a first job today versus somebody who never had a worthiness to borrow from a formal trade sector itself. So that population is pretty large.
01:00:05
Speaker
So going after that particular population is probably what's the focus, what we have today versus going and solve for a subprime category where somebody has already been defaulting and then now you're trying to figure out the reason for is defaulting and then trying to provide a credit. I don't think that is what we are trying to specialize here. Yeah. Okay. And how is the way in which you're going after new to credit? What strategy are you using to tap that segment?
01:00:30
Speaker
No, strategy is simple, right? So they do come out of the platform. We don't, when we are advertising about our products or when these customers come in, we don't necessarily say that target saying that we need a specific set of customers or even for the most of these ad agencies or the ad networks. I don't think they have a way to differentiate whether somebody is a new to create or he's not a new to create it.
01:00:52
Speaker
to the lender to decide once these people do come on to the platform but having said that this population is pretty large and increasingly there's a good amount of customers who are coming onto the platform and we are able to assist again the assessment remains very fundamentally same so looking at what is his current job and everything need need not be if he's a 24 year old who has just graduated joined his job probably for last six months he is working and he shows a pretty much credibility then why not give
01:01:19
Speaker
There I don't think is a specific strategy because I don't think the strategy would work at the marketing or an advertising level. Let's say we are sponsors of RCB today. So digital sponsor for all for RCB, official digital partner for RCB.
01:01:33
Speaker
Now, how do they segregate saying that whether there's an NTC who is watching the cricket or it is a 700 plus, I don't think that particular thing works. When people come onto the platform, I think that's where your ability to solve for it, what actually matters. What is your customer acquisition roadmap? You are spending on both performance and brand, like the Royal challenger IPL team you're sponsoring is obviously for building brand value top of the mind recall.
Growth and Future Plans
01:01:58
Speaker
Do you also spend on performance marketing?
01:02:00
Speaker
Yeah. So we do spend on performance marketing. So only good thing, what is happening with us is close to about 40, 45% of my customers are organic because since we are a early mover advantage, what we had. So already there are close to about 40 million plus customers data or what we have underwritten in this last four and a half years of our journey. And there's a huge world of more that happens of these people. Second is the overall SEO that you have to work upon.
01:02:26
Speaker
These are the norms, right? But it's not that we have cracked some rocket science in terms of anything on the performance marketing. For everyone, I would say that performance marketing is what you can optimize is anywhere between 10 to 15%. One company which does it maybe 15% better than the other company. It's a good number and one should not forget about it. But the biggest advantage of getting the organic traffic also depends on the number of customers what you have served.
01:02:49
Speaker
It's as simple, right? So today I don't think probably Ola or Uber are doing any ads. So I don't think they're doing any ads because they have such a large user customer base and they have created certain brand name. It's a word of mouth. They know how about my friend is using it. My dad is using it. So why don't I use it? So.
01:03:05
Speaker
That particular thing comes only based on the base customer base, what you're carrying it. But of course we are not at reach that stage of Ola or Uber, but we do also need to spend on the performance marketing and whatnot. So about 55% around or 60% in some months is the paid traffic, what we get it, but 45% for us is organic. And that keeps our costs pretty low.
01:03:25
Speaker
And not to entirely rely on digital channels like Facebook, Google. If somebody is having that strategy, I think it's going to hit hard a lot because there you don't have a control on the CAG. Imagine the number of unicorns that has been created in this year. So in the last one year, and there will be similar amount or double the amount what should be created in next one year, one or two years.
01:03:43
Speaker
And when all this kind of money is pumped in, everyone is going for the customers, and they are going to bid on Facebook, Google of the world, which is a very natural choice to stop the customer acquisition and whatnot. But in that way, only these two companies get a lot and the entire bidding process, your tax are always going to go high. So one needs to have a very diversified strategy, not just rely on these channels. These channels are very sexy. You just put your credit card. You can get any number of customers in front of you.
01:04:12
Speaker
have control on the cost what it is spending. So, one needs to go for multiple strategic integrations, strategic partnerships in utilizing. I give you one example, right? So, we have let's say we have close to about 25-30 integrations across with many of the channels. Interesting channels are payments banks. So, the payment banks by nature of the license, they don't lend. We have the lending partner with Fino Payments Bank, which does its particular role, but
01:04:37
Speaker
There are customers who come and then look for a loan and we serve the loans for that particular customers who are looking for the loan. So now that becomes a very strategic partnerships. We are not intending to do any payments business and payments bank. Of course, they don't intend to lend and also their licenses don't permit them to the lending and we become the lending partner over there. So now those strategic partnerships are the ones which comes and saves on your CAC.
01:05:00
Speaker
Rather than, as I told you, just plugging your credit card here and there. Got it. Okay. What is the need for you to raise more equity? Like you are planning to raise more equity, right? What is the need for that money? No, the way it works is there is certain leverage at which you can lend. Regulator allows you to, for an NBSC, Regulator allows to leverage one is to 7X, as in if your equity is a hundred dollars, you can borrow over $700 and then you can lend. Although that's the limit what Regulator defines it, it's not the limit what market
01:05:29
Speaker
is comfortable with. For some companies that particular leverage can be just about 1x to 2x. Some companies it can go up to 3x and let's say a very mature company like Bajaj Finance may be operating at about 5x. So nobody has hit that 7x. It all depends on and this is on the unsecured lending part what I'm talking about.
01:05:45
Speaker
Very secure like housing loan and all, you could easily see about 45x of a leverage, even for us very small size companies. For an unsecured lending, it's at about 3x, 3.5x per number. Now, once you hit such limits, then you have to infuse equity so that you are able to serve and build larger loan books onto your platforms. All right. So that is one thing. Second thing is, does that mean that one has to keep raising funds every two years, every three years?
01:06:11
Speaker
it need not be because as the company's brand value and as the company's existence, their ratings, everything improve, your leverage also keeps increasing. So the company which is today at about 3x kind of a leverage with next two or three years of consistent performance, they can easily kind of be operating at about 4x or 5x.
01:06:30
Speaker
So that's the reason why the NBFCs or FinTechs, or even to the extent of banks, why they keep raising it. But banks have a different nature of managing this because they also have a deposit taking abilities. So they take the customer deposits and they basically use that deposit to further onward lend it. So they did not actually go for this equity raise always. This particular deposit taking is also dominated by only a few banks. There are many other banks which are there to create that kind of a brand value so that people's goods
01:06:58
Speaker
start saving over that. So they also have to go hit the market to raise equities. Yeah. What is your number multiple of equity that.
01:07:06
Speaker
So we are at about 3x plus. So that is where the comfort is there in the industry for us. But yeah, so if I have to go for 4x, then I'm more leveraging myself and market also starts restricting there at that level. But there are, we are at about 3x plus. But you don't need money for customer acquisition. Like each customer acquired is profitable for you. Like the cost of acquisition is less than the amount you earn from that customer.
01:07:29
Speaker
We are generally profitable in some months we are profitable at 1.2x of the loan as in like generally on the second loan we are definitely profitable but what portion of the second loan from the customer? So even if let's say I've acquired about 100 people, at certain months if 20 people take the second loan out of that hundred, I break even on that hundred people. So in certain other months it could be 30 right? So it ranges between for us is like 1.2 to 1.3
01:07:57
Speaker
3x of the number of loans what they take from here. At exactly just on the one loan cohort we don't make money but at 1.2x of the loan even if the 20% repeats or 30% repeats we are definitely breaking even from that cohort. But having said that we have about 80% repeat ratio so that's beyond that particular number. Amazing, amazing.
01:08:16
Speaker
If you like the Found A Thesis podcast, then do check out our other shows on subjects like marketing, technology, career advice, books and drama. Visit the podium.in for a complete list of all our shows.