Welcome to Unstarted. What is unique about this show is that it is about you. It is for founders, by founders. I am very excited for a very special episode of Unstarted today. Unstarted is about all the minus one to zero. But what do you unstart for? You unstart for a destination, which this gentleman who is my guest here today, Srikanth, who started Fractal Analytics. Hopefully, you all have heard of it as a most recent IPO two months ago. Srikanth, it's great to have you here. Congratulations on the IPO. People like you are a big inspiration for founders who start, because ultimately, it is about some form of a next milestone, and nothing more aspirational, frankly, than going public. I sold my company. Sometimes I wonder if I sold too early. You have been through the same vintage. So it's very, very inspiring to see that. And kudos and welcome.
You sold a company that inspired millions of others to start companies. So that was quite game-changing back then. Thank you.
Thank you. So Srikanth, let's go back. You know, one of the objectives of this show is that when people look at backgrounds, I keep telling them, when you look at somebody's background, it may be intimidating, but actually the story behind the story may not be. But actually, your background does look very intimidating. IIT Bombay? Delhi. And then IIM Ahmedabad. Then you started this company quite early. But I know that the 26-year journey has had lots of ups and downs, many of which will have a lot of lessons for the people here. So it would be great to just go all the way back. Where are you born and raised? What's the family background? And then we'll come to how you ended up starting a company which was, has been, I would argue, an AI company before people knew the A of AI.
Absolutely. And it was analytics at that time. But we should talk about that. I grew up in a small town in Assam, in a place called Duliajan, where there are oil wells. And growing up in a middle-class home, dad used to work for an oil company, lived in multiple parts of India, made friends. Every three years, we would move somewhere and have new friends to make. So I grew up as a guy who was very interested in knowing all of India, learning, and really not having clarity as to what I will do. Engineering, etc., because I felt I was good at math, I wanted to do engineering. Did not know that engineering is not as much math as I thought it would be. But yeah, it was like that. And during my engineering, I realized that I'm actually more interested in psychology, human psychology, understanding how human behavior works. And now when I look back 26 years, all the work I've done is about connecting these two dots, math to understand and predict human behavior. And that's really what AI has become. It's taken us a while to figure out everything. But when you look back, it is about math and human behavior.
How much of this is with hindsight? So we'll come back to that. So let's keep going through the journey. What then took you to, so you got an interest in psychology. Then you went to IIM. And then you did a couple of jobs. You come from a middle-class background, which many of us come from. It's very relatable. And a small town. Where did the entrepreneurial risk-taking, in Hindi we say kida, but also risk-taking come from? And especially given that background.
I grew up thinking I will never be an entrepreneur. I didn't have any entrepreneurial genes. My parents, my many, many generations, there'd be no entrepreneurs. And my idea was I'd work for a big MNC. And the simple reason behind that is I didn't think I had the capital, nor the ability to hustle and maybe even work the system like it used to be back then.
Generate the resources, basically.
Exactly. That didn't feel like that was me. I thought I'd work for big companies. And so then I met a few people that changed my worldview. I realized that maybe India is different. Now you can be an honest businessman in India. That was something that came through. And then during the dot-com, people like you, raising a bunch of capital, also told me that you can raise capital if you have great ideas. So I wasn't the only one falling for the dot-com bubble.
I did, yes. Tell me about these people who influenced you. What kind of, because one of the themes that we hear, and which I was not expecting when I started these discussions, was that a few entrepreneurs have told me that sometimes people around you know you better than you know yourself. And these are all people who never saw themselves as entrepreneurs. Who said, but people around us told us to do this. Say a little bit more about that.
So one assumption was that there is no such thing as an honest businessman. Correct. And I met Narayana Murthy at that time. And he came to our campus for a business ethics lecture, as a guest lecturer. And I saw the way he was building Infosys back then, it was a very tiny company. But something told me that India of the late 1990s is different from the India of the 70s and 80s. Which made me drop that assumption that you need either a lot of money or you need to be unethical in order to build a business. So that was one big thing that happened. And secondly, all my friends around me were very clear that I was very passionate. I was very clear that the internet is going to change the world in a very significant way. I knew that some business had to be built around that. I was talking about those ideas, but then when I saw all this around me, I felt like maybe I could be an entrepreneur myself.
And what was the core motivation, if I were to ask you? For me, by the way, because I was telling somebody earlier, I was a Goldman Sachs banker carrying the bag for these CEOs who are going public. And they would be in private jets. I would be in a private jet, but to carry their bag and make the presentation. And I said, I want to be that. So money and that lifestyle was clearly a motivator. That went away very quickly. So for you, was money a motivator at all? What else other than, yes, I can do it, but why do it?
It was about success. It was about creating something. It was never measured in money. I was never driven by money that much. But soon it became about building something that will be around 100 years. It's somewhat egoistic, I would say, but it is a quest for immortality, if I would call it that. Very good. Can you build something that will be there once you're gone? That was the thing. Well, it's not personal immortality, because you would be gone, but you want to impact the amount that you want to leave behind.
That's very good. Excellent. Then, actually, let's pick up one question, and then I'll come to how did you think of starting something like this at that time. So the question says, founder, this is from Kush Sembal. The question is founder conviction versus market readiness. It's a long question, but I will rephrase it to just say, how do you founders balance strong personal conviction with real market signals? And how do you know when the traction is real? So now walk us through how did you think about what you would start at that time? You said you were clear that the internet would be very big. So within the internet, what you would start? How did that? One of the things we talk about is why are you uniquely positioned to start it?
Exactly. So absolutely right. So we became entrepreneurs without having a very clear idea of what probably is.
There were six of us. At one point, it was almost going to be seven. Then it became six. Then it became five. Then it became three. Then it became two.
So there's a whole story there. But we were clear that we want to be entrepreneurs. We did not even have clarity on what problem to solve. We did some soul searching. And once that whole internet thing had gone away, and we realized that there is no money, we can't raise capital, we can't survive, we can't make revenue, that's when we said, what do we do? And I said, we know math. We understand human behavior. Can we use math to help companies make better decisions? Now, strong conviction, because I had seen usage of math in doing some structured debt work in ICICI before that. So we launched India's first CDO, collateralized debt obligations, back in 1999. It was the first.
Which took the world down in 2008.
Exactly. We created that in India for ICICI. And there was an amazing amount of math simulations, risk modeling, parsing out all the cash flows. And I thought, ICICI needs so much of math for it to do something like this. What if you can bring math to other decisions that companies make every day? What price to charge a customer, what promotions to run, how to understand your customers, and so on. And the key idea was that we knew that math would be helpful, but we had no idea if the world was ready. So to your point about market readiness, there was a lot of personal conviction because I was interested in math. And remember, whenever we are sort of deeply immersed in any problem, it feels like that is the most important problem in the world. It's never that important, but I was in that mode. And the market realities were very different. We went across the country in India, and we said, you know what? You can use math to make these decisions better. And there was no response. We went to ICICI, and we said, you are launching a 30-minute loan product. You need to assess creditworthiness in those 30 seconds. We'll build that risk model for you. And that's when we got the first break. We wrote a research paper. We built the first model. It was very successful. Then we went to HDFC Bank. It took us 18 months to convince HDFC Bank that this is a good way of doing it. So yes, we had the conviction, but the market was not at all ready. And we barely survived during that time. It was good that we had very low expense base, and we were not expanding talent, etc. So it took us about three, four years to, I mean, we were profitable in year three. But it took us three, four years to know that this business is going to take off like a rocket. But there were results. There was feedback from the market. So we went to Citibank, and we said, Citi, we can build a model that can help you cross-sell personal loans to your credit card customers. They said, we have 100 PhDs sitting in New York. They're very, very good. You guys have not even 10 people, and none of you have a PhD. How do you think you'll beat them?
We said, look, we are very good. Just give us a chance. We'll do it as a bake-off. And if you're better than your PhDs, then take us. Otherwise, don't pay. And when we built this model, they were phenomenal. And we had incorporated price elasticity and things like that into the model. It worked like a charm. And they then said, okay.
How did you guys know that you were better than them?
So they actually did a campaign. Yeah. And they saw our accuracy was-
Understood. So that's great. That's, sorry to interrupt, the results. How did you know that your capabilities were also better? You were just at the confidence- That was just raw- Founding- Yeah, yeah. No, was it raw confidence? Or was it backed by actually the quality of the people was very good?
Not at all. Actually not. It was just pure passion and some self-confidence, let's put it this way, some bravado. And what is there to lose, right? We're going to give it a full shot. We'll try very hard. And this was a nascent space. And we had done some work which felt that, okay, there was some proof that we could do better. And we had all these models, etc. But a lot of it was about-
Betting on this and proving yourself.
There are many things I want to abstract away from this. You mentioned that you went to college because you liked math, then realized after the fact that that wasn't about math, engineering wasn't about math. But you never gave up your passion for math. And then you got interested in psychology and went back. Even in your first job, you tried to implement the math with ICICI, I guess, second job. And then you started a company for math. I think that's a great, great lesson for people. So, and this has also come up in other situations where people are constantly compounding on their core passion. Now, sometimes the core changes, but in your case, you just kept building on it, building on it, which is fantastic to see. Second, business didn't go anywhere for some time. Same situation you were in, probably in the same years. By the way, one of our first clients was also ICICI. Because our consumer marketplace platform was not working, there weren't many internet users. We went and did reverse auctions for procurement. We went and did bank-to-dealer auctions where ICICI was a client, just to get revenue. And then to your point, that even if the scale wasn't apparent, the some feedback that you're getting makes it clearer. So, which goes back to this question that, that is not delusion, that you're seeing actual traction and customer feedback.
There are some results. You know that what you're doing is making a difference to them. You know that you're able to naturally do a better job than other people out there. When you're, it takes education, you're educating the market, but people are listening, they're learning. And somebody like HDFC Bank, once they figured out how to do this, they went for the next six years, they built a fully AI or analytics-first bank back then. It worked out really well for them.
Yeah, yeah, they did, probably them, and maybe Bajaj Finance at some point, the best cross-selling across products and stuff like that. There are two other things you said. There was some hustle, some bravado. It wasn't, you guys were convinced you could do it, but there was no evidence that you were actually better. But you had the confidence.
Because we had no idea what the 150 PhDs are capable of. And we knew that we're going to work very hard on this. But beyond that, it was just a guess.
Yeah, yeah. And finally, you said profitable by year three. And I keep telling people, revenue is your best VC. And profits is of course your best IPO in some way. So, great to hear this. It's been 26 years since you started. I want to hear how from those initial days, how things have evolved, how you have brought it to the stage of IPO. But I also know that there's been a lot of evolution, some painful, some joyous along the way, in even how the co-founder journey played out. So before we go there, let's first take this question on co-founders. My question to you is that how do I know if I should be a solopreneur or enlist a co-founder? So, you enlisted six or five, a total team of six. Talk about that. Let's talk about not just answering this, but overall the journey, how it has evolved to this stage.
So we started as six co-founders. And honestly, it was a super bad idea. When you start, you feel like more means more social proof, like more people are quitting with me. So I feel a little bit more secure that I'm not failing alone. But six people means decision-making is slowed down, lots of negotiation. It was like a Pope's meeting for a decision on the church or something like that. It was just crazy. So that kind of glacial pace at which you move. Now, the question is zero, one co-founder, no co-founder, or many co-founders? My answer after all the work I've done so far and all the companies I've seen is just two people. You and a co-founder, and complementary skills. If you are the CEO type, a CEO type means you can sell, you can hire people, you can inspire people. That's the CEO type. And CTO type is someone who can get things done, build the best technology, build the product, be relentless in terms of operational excellence. I'm creating two skill sets here. There could be more. But if you have these two, if you are one of them, find the other one. There's no one who's all of the above. And my understanding is that if you are a CEO type and CEO, CTO type, you could be everything. Even then, there are so many things that come during the way, during the entrepreneurial journey. You need someone where you can sort of brainstorm with, someone you can problem solve with because the job of a founder is very lonely. So if you have another person who you can really trust, it makes the, it de-risks your overall company, etc. And here, you're not about making, it's not about making the maximum money, right? It's never about making more money. It's about increasing the probability of success. So if you want to increase the probability of success, have a co-founder. That would be my.
So you're saying six is too many, but one is too few. And I was going to ask you that question, but it sounds like forget skill sets. Even if you feel you have all the skill sets, it's lonely at the top. And I completely agree with that because you are always in a, whether you are with your investors, whether you are with your employees, whether you are with your ecosystem externally, whatever, you're always selling. You're never real.
Yeah, and one thing I would say is that if two of you are both almost identical in skill sets.
Then it's a problem. Then you may run a, slight differences might become bigger differences.
Some complementarity is great.
Yeah, yeah, yeah. Or find it like me and my co-founder were identical. We were great friends. And then we didn't know what the complementarity was. Maybe a little bit that I would do more finance and product because of my background. But then we found our complementarities and we actually sat down and we said, okay, instead of, and it wasn't as much friction between us. It was who you said, decision by committee. One person should be able to take a call on one room.
Yeah. Right. And so that's how we said that was okay. And by the way, some of that we operate even now like that. Since one person take a call, otherwise it all becomes.
It's kind of building specialization over time, right? I become focused on one side and become really good at that. The other person focus on something else and becomes really good at that. We may come from similar backgrounds and therefore similar skill sets, but then we specialize differently. Yeah, yeah. I'll pull in one more question and go back to the story because I think this will come into your answer. So question. Hi, I'm Munthimu and I'm a product manager right now and I'm an aspiring founder. So my question is to kind of understand the rationale behind how do you kind of crystallize your ICP, right? Like how do you ruthlessly prioritize your product pricing and features for your ICP and not kind of accommodate the noise from the rest of the crowd? So the question is about finding your ICP and how do you not get distracted by noise, but let's pull back. You have been doing this for 26 years. You started with ICICI Bank and you started with HDFC and analytics. Everything in the world has changed in the last 26 years. So walk us through that story. Take your time. What has not changed? What has changed? How did the customer segments change? How did technology change? And then let's talk about the current stage of AI.
Yeah, initially it was about using math to make better decisions and we picked banking as the vertical because that's the vertical that we understood best. We built a pretty good business in India serving banks. We served almost all the big private sector banks back then and helping them build risk models, marketing models, and so on. And at that time, AI was in the stage of nascency where the best models were about building on structured data, which is things that are in rows and columns, and building simple statistical models, right? Like logistic regression. It's simple, but it's extremely powerful. Logistic regression was a very good way of building models and sort of helping companies make decisions. Then we moved to more basic machine learning techniques like random forest, decision trees, XGBoost, etc., which were better. And at that time came in the wave of deep learning related progress.
Just keep us updated on the timelines also. So 2000s was your first phase?
Yes. So early 2000s, 90s was all about expert models. 2000s, we built statistical models. By the way, I built an expert system in IIT. Can you imagine? 35 years ago and now all of this is your thing.
