Brands
Discover
Events
Newsletter
More

Follow Us

twitterfacebookinstagramyoutube
Youtstory

Brands

Resources

Stories

General

In-Depth

Announcement

Reports

News

Funding

Startup Sectors

Women in tech

Sportstech

Agritech

E-Commerce

Education

Lifestyle

Entertainment

Art & Culture

Travel & Leisure

Curtain Raiser

Wine and Food

YSTV

ADVERTISEMENT
Advertise with us

Machine learning algorithms will guide lending in future, says Clix Capital CEO

Machine learning algorithms will guide lending in future, says Clix Capital CEO

Wednesday June 13, 2018 , 9 min Read

Bhavesh Gupta, CEO of Clix Capital, says lending institutions are sitting on massive amounts of data, and can sanction loans on phones by using algorithms to check customer credentials.

There is always time to innovate like a startup. Clix Capital was formed in 2016 after Pramod Bhasin and Anil Chawla, both former leaders at GE, along with AION Capital, bought the commercial lending and leasing business of GE Capital. For the last two years, the company has been working like a new age fintech company using algorithms to study customer behaviour, and predict payments. The company is working on products across their enterprise and consumer verticals.

This is where Bhavesh Gupta, CEO of Clix Capital, thinks technology can play a major role in winning new customers. He draws a distinction between digitisation and automation, telling YourStory that modern-day lending is all about using data to predict outcomes. The TC report considered people in the age group of 20-69 years and with an income of over Rs 2.5 lakh a year to come to the estimate of 220 million credit-eligible consumers. The report also forecasts that the addressable market will increase by 14-16 million consumers per year, reaching an estimated 295 million by 2022.

No wonder that Clix Capital feels the India technology story in lending will change their fortunes, and that of many Indians.

Edited excerpts of the interview:

CEO of Clix Capital, Bhavesh Gupta

YourStory: How has digital evolved in lending institutions in India?

Bhavesh Gupta: I have been in the lending industry for years and technology usage in India only began in the last decade. Today, technology has democratised the entire consumer side of banking and financial services, which is mostly customer experience. We, as an industry, have moved from branch-based transactions to branchless transactions; mobile banking is taking off in a huge manner. It took India less than a decade for technology such as ATMs, point-of-sale systems, and mobile banking to increase in penetration. I believe the financial industry will be mostly branchless in the future, and every institution is preparing for a future that is technology led.

But while the consumer side has seen the adoption of technology, which is basically the liabilities side of finance; things are slower on the lending side. Processes were completely manual globally. Loan origination and mapping was not done through digital processes; it was all on an excel sheet. In the last 10 years, companies have pulled out their loan origination data and are sorting it out with data collection tools. However, for me, this was not technology-led lending; it was still part of automating workflows. This data organisation - digitising data - happened till 2012. The knowledge-oriented approach began only after 2012.

YS: What happened after 2012?

BG: All the data was in loan origination libraries and stacks. That’s when lenders began to realise that something could be done with the data to engage customers better and create profitable products. Most lending institutions still don’t track consumers. But today most accounts are linked to Aadhaar and this allows businesses to understand customer borrowing and interest payment habits. Lending institutions are sitting on massive amounts and they can sanction loans on phones with a pre-qualified offer by using algorithms to check customer credentials and give you a loan tenure based on your past interactions with the institution.

This is happening big time; we are also applying a lot of technology on the same. The whole lending process is predictive today; it’s harnessing the power of analytics, thanks to smart algorithms. Lending institutions are becoming smarter in the process.

YS: Talk about how data is being used and how data is being collected for analysis?

BG: Customers have moved from branch-based banking to using phones, net banking, or SMS. All this is data and for lending institutions; this is a significant migration that has happened. In the next five years, I don’t see branches existing. There will be more customer experience centres and not branches.

How are we using data? We use several alternative sources of data and supplement banking transaction data to help our clients monetise customer behaviour. We track – as an aggregate - how one pays bills, we study social profiles, their credit card payments. One of our products has 10,000 contracts made based on data crunched with machine learning algorithms. The problem our client had was that they had enough data to understand why customers buy mobile phones, but did not know how to monetise that data. So, we came in to study credit worthiness and looked beyond standard CIBIL scores that determined the credit-worthiness of the customer. We said we will create “lookalike” data, which means that we can create a pattern of the individual’s data beyond traditional means.

We decided to look at, for example, mobile bill data that throws in data points about how many calls and purchases were made. This is very rich data where we can give customers a better experience when applying for a loan. What is that experience? The whole point was to make the customers come on to our website or our partners’ websites and get a loan sanctioned in minutes, thanks to our algorithms that can look at data from the customer’s uploads and analyse them for credit-worthiness.

Our algorithms compare the “lookalike” data with that of the external data, CIBIL data, that we bring in and we match the two to approve loans to our customers. Our delinquency rates are very low, thanks to the technology, and we are going to extend this product further to all our clients.

YS: Is modern lending all about analytics?

BG: We take a lot of pride in building analytics. We have people working for us globally. It took us six months to crack a model, test it, and implement it. Our data can tell us who we are we declining a loan to and who are we giving a loan. There is some correlation to delinquency based on location and their purchase patterns. Today ML models are still in an early stage and cannot be used to say that delinquency rates are low. This will happen with homogeneity and scale of data. Once there is a massive data set, ML models can figure out delinquency rates accurately.

YS: Are you applying analytics for enterprises?

BG: For large customers, we apply ML at broader levels. We are looking at who is likely to borrow and we collect a lot of data that shows whose profitability is going to grow. We also track which company is hiring senior teams, expanding operations, and adding more branches. We also compare this to borrowing in years and repayment patterns.

With data, our managers are able to win clients faster. So to your question on analytics, we are using forecasting models and bringing in ML to understand our lending better. We are not a company that looks at just P&L data of enterprises. We are looking at insights far deeper than just the financials of a company. I am doing all this with just 15 people. Once you use an analytical tool, you don’t need lots of people. The future is about algorithms that matter for decision-making.

Analytics is what cracks customers. In the future, technology can reduce opex costs; an agent need not go home to verify a customer. Since processes are digital we can use triangulation logic – by asking customers questions - to arrive at a thesis that such a person exists and s/he resides in a particular geography. In the future, a lending business no longer needs proof of income as technology can predict the person based on intelligent questions asked. How can this happen? This can happen with our proprietary data, CIBIL data, and open API data from banking institutions.

YS: What are your macro predictions in 12 months?

BG: There is a lot of data that says macroeconomics is bad and microeconomics is good. The headwinds of dollar and oil prices going up show that interest rates can go up and industry can struggle because of this. On the micro level, the pangs of GST have gone away, rainfall is predicted to be good, and the capex for large corporates is going to be good. But I am neutral in my view about the economy. I believe consumption is growing, a lot of money will be pumped in because of the runup to the general elections, and the government will ensure liquidity in the market.

Bhavesh says he is clear on how technology is impacting and will continue to affect lending. There is more data to believe why it will help in growth. Today, the total household debt in the country can nearly double to Rs 78-94 lakh crore by 2022, from Rs 37 lakh crore last year if the country follows the growth trajectories of other countries.

Clix Capital, one of the fastest growing NBFCs in India, uses technology to help clients monetise data. According to CARE Ratings, the company had assets under management (AUM) of Rs 802 crore as of December 2017. The capitalisation – stock and retained earnings - remains comfortable with CAR of 80.16 percent for Clix Capital and 60.39 percent for Clix Finance as on March 31, 2017. The company, has operations in 10-15 cities, serves customers in close to 100 towns, and has partnerships with fintech players, telcos, ecommerce firms, and startups.

According to a report by credit information company Transunion Cibil, banks and other lending institutions have tapped only into a third of the estimated 220 million credit-eligible consumers till now. The rest of the 150 million consumers, who are eligible for credit, are still untapped, providing an opportunity for sustained and prudent growth for lenders over the next five years and beyond.

According to PwC, a majority of consumers across all age groups prefer to apply for credit online rather than in person or on the phone. At the same time, digital lenders benefit from the cost and scalability advantage over traditional lenders through digitalisation, and can access new customer segments in areas that were not before reachable by banks and their branches. Clearly, Bhavesh – and Clix Capital – have made the right bets.