All told, Aye has lent about Rs 1,000 crore ($137.1 million) since its inception four years ago, serving roughly 80,000 customers. All small-ticket loans, sure, but Sharma claims that 70% of Aye’s 103 branches have broken even and, on an overall basis, Aye has been profitable since the start of this year. Aye also plans to lend another Rs 500 crore ($68.5 million) this year, adds Sharma.
In comparison, Lendingkart has given out Rs 2,000 crore ($274.3 million) since 2014 and said it will achieve profitability before the year-end. Lendingkart says it will disburse Rs 2,000 crore this year alone.
“Today, the difference between the online and offline models is blurring. An only online acquisition can be termed as a purely digital activity unique to online models.
The rest of the processes related to Customer Relationship Management, operations, analytics, credit, and collections are rapidly converging between online, offline and omnichannel models driven by technology,” says Norwest’s Shah. As such, there’s an increasing consensus that offline or omnichannel NBFCs are still viable investments.
This is what investors like CapitalG are counting on. As a VC firm steeped in tech, it understands that digital does not automatically mean better for business.
“We looked at both digital and offline business since we believe that technology would play a critical role in the credit market and there will be convergence across fintech and offline businesses irrespective of where they start,” says Anand.
While digital helped fuel the pace of growth in consumer internet businesses, it isn’t the be-all and end-all in a business where the pace of growth is a grossly overrated metric. A metric that is potentially even detrimental.
Whether offline or online, there’s only one metric that truly counts in lending: non-performing assets. In its own research, CapitalG found that in the microenterprise segment, lenders which had more touchpoints during the loan process had better asset quality and lower NPAs, according to Anand.
Growth in the online business
Shah of Norwest, who is also scouting the market for fintech bets, found that the GNPAs in online businesses was higher—about 5-10% on a cohort basis. Traditionally, Gross Non-Performing Assets (GNPAs) are measured for the entire portfolio. But investors prefer looking at GNPAs on a cohort basis since GNPAs emerge with a lag. A cohort, in this case, is a set of loans that were taken at around the same time. This sort of comparison ensures that GNPAs of a certain cohort are tracked against the same set of total loans.
According to Shah, offline businesses fared markedly better than their online rivals, with GNPAs of 1-4% on a cohort basis, going up to 4-7% for unsecured loans. He added that this did not account for outliers in both business models, of which there are several.
It is also a myth that purely online models are cheaper than offline models. For instance, as a digital business, Lendingkart’s expenses—which include salaries, marketing, and tech costs—should ideally be lower than that of Aye’s. However, for every Rs 100 ($1.37) earned, Aye spent Rs 136 ($1.73), while Lendingkart spent Rs 150 ($2.06), according to their respective FY 2017 financials.
Role of the Third-Party Sources
This is why many online lenders that started off as purely online businesses—companies like Lendingkart and Capital Float—now employ field staff and contract third-party sources to acquire customers. For instance, 40% of all Lendingkart’s loans are sourced through third-party agents, according to a credit rating report by rating agency India Ratings and Research. But customers sourced through these agents don’t always end up being sticky.
“Having your own channels to source loans is important as sourcing through intermediaries typically results in lower repeats and poorer asset quality,” explains CapitalG’s Anand.
All this has led to a wider belief that slapping tech onto a finance first business is easier than the other way around. CapitalG is working with Aye to help them build their analytics capability. “We are leveraging Google advisors with expertise in analytics and machine learning to help build out their real-time dashboards using their underwriting and collection models,” explains Anand.
Sharma added that the use of artificial intelligence and machine learning would help reduce delinquencies and also help them identify the quantum of NPAs well in advance. In addition, machine learning can also help them simplify their onboarding processes.
In all of this, the metric of scale is falling by the wayside. But can it really be ignored?