Whatsy - WhatsApp Business Solutions

The hard part of lending isn't lending. It's knowing who to say no to.

Most digital lending startups skip that problem. They target customers who already have a credit score, already have a bank relationship, already have a formal borrowing history — and compete for them on speed and interest rate. It works until the incumbents notice, at which point you're a thinner-margin version of a bank with a nicer app.

Fibe — formerly EarlySalary, founded by Akshay Mehrotra and Ashish Goyal, with 46 million+ app downloads, 9.8 million+ loans disbursed, and over Rs. 48,000 crores distributed to 3.3 million+ customers as of March 2026 — is interesting because it started at the wrong end of that problem on purpose. Its founding customer was a 25-year-old with a salary, a smartphone, and no credit file at all. The question worth asking is what you have to build before you can lend to that person profitably.

The Bet That Started It: Underwriting the Thin File

The original EarlySalary product was a small, short-tenure advance against an upcoming salary. On paper it looks like a niche consumer convenience. Structurally, it was a data acquisition strategy.

A young salaried Indian entering the workforce is, to a traditional lender, close to invisible. No credit bureau history, no collateral, often no prior loan of any kind. The conventional answer is to decline — the file is too thin to price, and the ticket size is too small to justify manual assessment. That's not a moral failure of banks; it's arithmetic. Underwriting costs the same whether the loan is Rs. 20,000 or Rs. 20 lakhs.

Fibe's answer was to make underwriting cheap enough that small tickets clear the bar, using alternative signals — employment data, transaction behavior, device and app-level patterns — scored by models rather than by people. Each small loan then returned something a bureau score couldn't: observed repayment behavior on a customer nobody else had data on.

That accumulating behavioral dataset is the real asset. The salary advance was the product. The scoring system it trained was the business.

Why the Rebrand Was a Strategy, Not a Logo

"EarlySalary" is an unusually descriptive name, and that turned out to be the problem.

It told you exactly what the company did, which is useful when you're explaining a new product to a first-time customer. It also permanently filed the company under a single, small, and reputationally awkward category — the advance-against-salary product, which sits uncomfortably close in the public mind to payday lending. A customer who had grown past that need had no reason to think the company had anything else to offer them. Worse, the name capped the story: you can't credibly sell a Rs. 10 lakh personal loan, a credit card, or a loan against mutual funds under a brand that says "salary advance."

The move to "Fibe" is the sort of change that reads as cosmetic and isn't. It detached the company from its entry product so the customer relationship could outlive it. That matters more in lending than in most categories, because the entry product is almost always the smallest one. A lender that can't grow with its customer has to reacquire a new cohort of 25-year-olds forever, paying full customer acquisition cost each time. A lender that can grow with them earns the second, third, and fifth loan at nearly zero acquisition cost — and the fifth loan is where the money is.

The generic name was the price of that optionality.

Purpose-Driven Financing: Moving to Where the Decision Happens

The most revealing piece of Fibe's current strategy is the shift beyond general-purpose cash loans into financing attached to a specific need — education fees, healthcare, insurance premiums, travel, e-commerce, rooftop solar.

The instinct for a lender with a working app is to keep selling the same undifferentiated product harder: more marketing, more channels, cheaper rates. Purpose-driven financing rejects that, and the reason is worth spelling out. A cash loan requires the customer to first decide they want to borrow — an abstract, easily deferred decision, which is why it costs so much in advertising to trigger. A loan attached to a hospital bill or a semester's tuition doesn't require that decision at all. The customer has already decided to spend; the only open question is how to pay.

This changes the economics on both sides. Acquisition cost drops, because the lending offer meets a customer at the moment of a decision they're already making rather than trying to manufacture one. And underwriting improves, because a loan with a known end use and a known counterparty is a materially different risk from cash into a bank account with no stated purpose.

The catch is that each vertical is its own build. Education financing means integrating with institutions and their fee cycles; healthcare means hospital networks; solar means installers. It's slower and less glamorous than scaling one product nationally — which is precisely why it's harder for a competitor to copy in a quarter.

Borrowing Rails Instead of Building Them

Fibe's co-branded credit card with Axis Bank, alongside its insurance and deposit offerings, points at a discipline that not every fintech has shown: knowing which parts of the stack are worth owning.

A card program is not simply a product launch. It's network membership, issuing infrastructure, settlement, dispute handling, and a regulatory perimeter that belongs to a bank. A fintech that decides to build all of it itself spends years and enormous capital arriving at parity with rails that already exist and can be partnered into.

The alternative — bring the customer relationship, the distribution, and the risk models, and let a bank partner bring the licence and the rails — compresses that timeline to a fraction. It costs economics on every transaction, and that's the honest trade. What it buys is the ability to widen the product shelf at the speed the customer relationship matures, rather than at the speed a licence application moves.

The general principle underneath: own the part that compounds, which is the data and the customer relationship. Rent the part that's a commodity, however impressive it looks on a slide.

Where the Model Actually Strains

The things that make Fibe interesting also define its specific vulnerabilities.

The most significant is that unsecured small-ticket consumer credit is a cyclical business wearing a technology company's clothes. Growth in a benign credit environment tells you almost nothing about underwriting quality; models trained on years of expansion have never been asked the only question that matters, which is how they behave when employment softens in exactly the young, early-career segment the book is concentrated in. Every digital lender's real test is a downturn, and the test is pass/fail.

The second is regulatory. Digital lending in India has moved from a light-touch regime to an actively supervised one — rules on lending service providers, first loss default guarantees, data handling, and collections practice have all tightened, and continue to. An NBFC-fintech operating partly through partnerships is exposed to changes on both sides of that structure. Nothing about this is unmanageable, but it is a permanent institutional cost rather than a one-time compliance project.

The third is cost of funds. A digital lender competes with banks that fund themselves with deposits at a structurally lower cost. Speed, distribution, and better scoring of thin-file customers can outrun that gap while the segment is underserved — but as banks improve their own digital underwriting and move down-market, the gap stops being an advantage and starts being a headwind. The fixed deposit offering is a visible attempt to address exactly this, and it's the right instinct.

None of these are fatal. They're the ordinary costs of lending to people the incumbents skipped. But they deserve naming plainly.

The Actual Lesson

What Fibe's arc demonstrates isn't that digital lending is a good business. It's that in lending, the durable asset is almost never the product you launched with.

The salary advance was small, unglamorous, and eventually outgrown. What it produced — a scoring system trained on customers no bureau could describe, and a relationship with them that started years before any bank was interested — is what everything since has been built on. The rebrand, the move into purpose-driven financing, the partnered card: each is an expression of the same underlying logic, which is that the value sits in knowing the customer earlier and longer than anyone else, and the products are just the ways of monetizing that.

For any business studying how to enter a market the incumbents have written off: the temptation is to see the underserved segment as the opportunity. It isn't. The opportunity is the information you acquire by serving them when no one else will — and whether you build something that can still use that information once those customers have grown up. Fibe's most consequential decision may turn out to be the one that looked most cosmetic: changing its name so its customers wouldn't have to leave.