A lending company raised $15.4M and spent it like a software startup.
War Games: I take a company whose story has already played out, run it through a checklist I use to judge African deals, using only what was public at the time, and see whether the checklist would have caught the problem before the money went in.
Lipa Later was an African fintech that lent money to shoppers. It raised $15.4M, expanded across four countries, and collapsed in 2025. A software company and a lender need money for opposite reasons, and Lipa Later raised the kind a software company needs. The warning was in a press release three years before the end.
Lipa Later made its money lending to shoppers, so it needed a steady supply of cash to lend and time to earn it back. It raised the wrong kind of money for that, and every new country it entered made the gap wider. Verdict frozen on January 2022 inputs
What the pitch showed, and what the business was
The same company, described two ways.
What the pitch said
- One-of-a-kind "buy now, pay later" across four countries
- Its own credit scoring, instant approvals
- Exclusive deals with Carrefour, Apple, Samsung
- Three to five new countries within a year
- Named by the Financial Times among Africa's fastest-growing firms
What the business was
- A lender that has to keep funding every loan it makes
- Money raised as a mix of equity and debt
- None of the stable, low-cost funding lenders normally rely on
- Store "exclusives" a competitor could out-bid
- Competing with M-PESA's instant credit, which Kenyans already use
The difference between those two columns comes down to one line in the funding announcement: the round was a mix of equity and debt. A software company builds its product once and sells it many times, so the money it raises pays for growth. A lender hands out cash on every sale and waits months to be repaid, so the money it raises is the thing being lent, not fuel for growth. Lipa Later raised money like a software company and spent it like a lender. That mismatch is what broke it, and it is the same problem I looked at in the Copia retrospective.
Where the money went
The mismatch, drawn.
That is the mismatch as a picture. The checklist is how the same problem shows up as questions during due diligence, while there is still time to ask them.
Running the checklist on 2022 information
Open each dimension for what was known then, and the question it raised.
Known in 2022: the round was announced as a mix of equity and debt, with no split given. Established "buy now, pay later" firms like Klarna, Affirm and Afterpay had already shown the model needs cheap, reliable funding to keep lending: a banking licence, a renewable credit line, or the ability to package and sell its loans. Lipa Later had none of these.
The question it raisedHow much was it lending out compared with the real cushion it held, and what happens when the next round of funding is not available on the same terms?
Known in 2022: the idea was borrowed from Western markets without checking whether it fit Kenya. Only about 14% of adults had a credit record, which makes lending decisions hard. And there were no credit cards to replace: people already used Fuliza, the instant overdraft built into M-PESA that many Kenyans reach for by default.
The question it raisedHow many customers came back for a second purchase, market by market? If someone can get instant credit through M-PESA, why choose this instead?
Known in 2022: the company's edge was meant to be exclusive deals with big retailers. In African retail those deals are commercial, not binding. A store switches to whoever offers it a better cut.
The question it raisedIs the Carrefour deal a real contract or a handshake, and what stops a better-funded rival from offering the same stores more?
Known in 2022: Kenya was already writing new licensing rules for digital lenders (the DCP framework), and this product fell squarely under them. The cost and delay of getting licensed was foreseeable.
The question it raisedWhat does compliance cost, and what is the plan if operating without a licence becomes impossible?
Known in 2022: the founders knew the market. But the ownership split and the terms of the debt were never disclosed. That missing information is the concern, not the people.
The question it raisedWhat conditions were attached to the debt, and who decides between raising more of it and protecting the company's cash when things get tight?
Known in 2022: the demand was real. Plenty of shoppers wanted to buy now and pay later. That part was never in doubt.
The point it raisedA real market does not fix a broken funding model. People wanting the product did not stop the company from running out of money.
Two reds, and the verdict is do not invest on these terms. With that locked in, the outcome is the test.
What happened next
The same problem, on a calendar.
Would the checklist have added value?
I had no stake in this company and no inside information. The only question that matters: if an investor or lender had run this checklist before putting money in, would it have shown them the risk in time?
Would have surfacedThe company would run short of cash as it grew, because it was spending its funding on setup while the loans themselves had none behind them.
Expansion and the $1.9M acquisition used up its cash, and it could not renew its funding in 2024.
Would have surfacedWhether customers came back for a second loan mattered more than headline sales numbers.
The company kept quoting sales volume. It never showed that customers came back.
Outside the toolThe checklist works from public information. It cannot predict the exact timing of a collapse or see private decisions like the acquisition.
The public record shows what happened and when. The reasons inside the company are not visible from outside.
What the checklist could not see
The exact split between equity and debt was never made public. Neither were the company's finances at the time of the acquisition, the conditions on its debt, or what was decided in the boardroom in late 2024.
None of that was needed to reach the red verdict. The two main warnings came from the funding announcement and public market data alone.
What to take from this
How a business is funded matters more than how fast it is growing. A lender funded like a software company is a mismatch no pitch deck will point out for you. This is the pattern to watch for in any market.
A real customer need does not save a broken business model. People wanted this product. It failed anyway, because wanting a product is not the same as a business that can pay for itself.
None of this needed inside information. The warning was public in 2022. A checklist earns its value by forcing these questions before the money goes in, not after the company is gone.