One year at a fintech startup in India: learnings from a credit risk perspective

With the Indian fintech space heating up, it’s important to keep these lessons learnt on the job in mind.

I complete 12 months at Creditexchange today. As I walked towards work, it was not unusual for me to start looking back at the last one year and ask myself some of the more obvious questions. With the Indian fintech space heating up more than ever, especially from early 2015 to now, it definitely was the right decision for me to move into this space, but I do ask myself everyday - am I on the path to disruption of the lending industry in India?

To answer that question, I would have to go back further to the end of 2012. That was the time when I started Bainary, a data analytics product for the financial services industry. This was the first time I started getting my hands dirty as an entrepreneur. Though I did not succeed in getting this off the ground and moved back to learning more from the industry in nine months, this is what I learned from those 9 months:

1. Product is important, so is timing. You can never compromise on a good product, but you must ensure that the timing is right. If you wait a bit too long to get the best product out, you may have missed the bus.
2. Get the 1st customer ASAP. No product is good unless you have 10 people validating your thought. Get those first 10 customers to use it and sit with them to get their feedback. That's most crucial for any startup, in any industry.
3. Know your strengths and work on your passion. I had always been a Credit Risk person, and deep down I would always be - that's my passion. Any deviation from that would kill my company. I must focus on something which is aligned to my passion.

With a bit more grey hair, I joined one of the early startups in fintech, LoanCircle, in middle of 2015. Backed by some of the best early investors, it seemed to me the most obvious choice for me to back myself in the fintech space. Heading Risk for LoanCircle was like a dream. But that journey was short lived - just eight months and then, fortunately, Creditexchange happened. As I look back, I believe those eight months taught me the most about the Indian market and the Indian fintech space and helped me avoid the same mistakes at Creditexchange. As the oldest and the first member of the team, it was not an easy decision, with the risks being higher than ever. But I believed in the business model and the founder (and I will tell you why some other day). For now, here's what I have learnt at Creditexchange that has helped us build something I have tremendous belief in:

1. Do your homework, understand unit economics. Risk and finance are two different things. That was the first thing I learnt as I stepped here. A fintech startup, especially one into lending, needs to understand its cash flows, its unit economics on Day 1. This is something I had never even discussed before. Today at Creditexchange, we have one of the most detailed, well-laid-out financial models I have come across, and with two months of launch, what gives me more confidence in the model is the fact that our costs, our revenues and our growth are well in line with what we had put into the model. Credit risk has an impact not just in terms of losses, but also on overall costs - Underwriting(how many customers I approve every 100 applications), Acquisitions (how many customers get funded every 100 applications) and Opex (how many customers we talk to who should not have).
2. Good teams do great work. Hiring at Creditexchange has always been driven by one thing - passion to build the business as your own, even if it requires one to take inordinate amount of risk on a personal and professional level. Today, we boast of some of the most experienced people working together to build something that they understand to the minutest level. Startups in India talk about an average age of 25 or 28 years in their company. Well, here at Creditexchange, I feel proud to say that we have a core team whose average experience is 12-plus years - across Product, Tech, Risk, Operations, Marketing and Sales.
3. Fintech is less tech and more fin. Compared to any other industry, fintech works very differently. This is not a space where growth, even if it comes at huge losses, would work. I cannot stress more on the fact that credit risk and cost efficiencies are crucial to build a Fintech business. If one looks at the most successful fintech companies in the US, none of them has compromised on the portfolio quality to grow.
4. Be flexible on partnerships. Today as most fintech companies approach banks for a partnership (as they all need capital to grow their business and ensure that their models are validated), none of them are ready to share their risk models. At Creditexchange, I took the lead to take a different approach. We went to a large Indian bank and opened up our entire underwriting in front of them. The result was phenomenal. We entered a large Rs 100-plus crore partnership with a relatively insignificant risk-sharing arrangement - unparalleled in the consumer lending space in India.

Here are some of my more pointed thoughts from a credit risk perspective on what we have done different:

1. There is no machine learning on Day 1. Many fintech start-ups talk about Big Data and Machine Learning algorithms right from the time of inception. Well, I don't believe in that. Your first cut underwriting needs to be very simplistic and intuitive. That's how large companies like American Express and Capital One started. Without enough data, the machine would not learn, rather throw some inconclusive and statistically incorrect output (I can give enough examples from an exposure, utilisation and balance perspective to prove this).
2. You don't need 1000s of data points to predict risk. As many fintech start-ups talk about 1000s of data points in their risk models, I feel proud to say that we have limited risk in our underwriting framework using way lesser number of data points and would always continue to do so. The key for us is to analyse and understand what helps predict risk and not just throw everything into the laundry bag.
3. Share your risk model with everyone: I believe it is very important to get validation of the first cut of risk model to learn faster. At Creditexchange, we have shared our risk and underwriting across industry experts to do exactly that. In 12 months from now, our Risk Models would look very different from what they are today. To ensure we learn the most, it is important that the starting point is correct. This can be best validated if it is shared with lenders, investors and partners.

There is a lot of learning still to come. But I do believe that we are on the right path and the Indian fintech space is witnessing some great idea. What needs to be seen is which of these teams can pull it off through the right execution.

(Disclaimer: The views and opinions expressed in this article are those of the author and do not necessarily reflect the views of YourStory.)