The 10-day fintech test
How to prove customers want it, in the time it takes to write the business case.
Most fintech products that fail do not fail because the technology did not work. They fail because nobody checked, cheaply and early, whether anyone wanted the thing in the first place. Of the venture-backed startups that shut down since 2023, 43% cited poor product-market fit as the primary reason, according to a CB Insights analysis of 385 categorised failures published in March 2026. That is not a technology problem. It is a sequencing problem: capital and months of build committed before the question of demand was properly asked.
The order most teams get wrong
The instinct inside a bank, a PE-backed platform or a well-funded startup is the same: get the business case approved, assign a team, start building. Testing feels like delay. Approving a budget feels like progress. That instinct is backwards. A structured test of demand, run properly, takes about ten days and costs closer to the price of a piece of market research than a technology build. Skipping it does not save time. It moves the cost of finding out to a later, far more expensive date.
What ten days actually needs to prove
The first five days establish whether there is a problem worth solving. That means confirming the customer problem is real and specific enough to act on, that the proposition built around it resonates with the people who would actually buy it, and that there is a genuine, testable signal they would pay for it, not just say they like it in a survey.
The second five days establish whether it can actually be delivered. That means testing whether the product can reach those customers through a channel that already exists, rather than one that has to be built from nothing; whether the intended pricing survives contact with what competitors already charge; whether the regulatory perimeter is understood well enough to know what licence, permission or partner is genuinely required; and whether the product is operationally feasible to deliver at the volume being planned.
The rule that makes this useful rather than theoretical: if a team cannot get a real signal, not an opinion, on all seven of these within ten days, it is not ready to commit capital to building.
Two markets, four decisions
The UK offers the clearest contrast inside a single market, a few years apart.
Monzo, launched as Mondo, did not start by building a bank. In October 2015, eight months after founding, it capped its first public test at 3,000 prepaid Mastercards, limited initially to 500 iPhone owners in London, deliberately kept inside prepaid-card regulation rather than the far heavier rules that come with deposit-taking. The company described the exercise plainly as a test of demand, not a launch. It worked. By the time Monzo received its restricted banking licence in August 2016, more than 200,000 people had signed up across its Alpha and Beta programmes and nearly 30,000 prepaid cards were already in circulation, according to the company's own account at the time. Only once that scale of proven demand existed did Monzo take on the cost and complexity of becoming a fully licensed bank, in April 2017.
NatWest took the opposite route with Bó. The digital-only current account spent roughly 18 months and an estimated £100 million in development before launching fully built, with its own brand and app, in November 2019. Six months later it was gone. It had attracted around 11,000 customers, reported at the time to be mostly friends and family of the team that built it, not the wider market NatWest was trying to win from Monzo and Starling. NatWest folded the underlying technology into Mettle, its SME banking platform, and moved on. The lesson is not really about execution, which is hard to judge from outside. It is that nobody had established, before the capital was committed, that a meaningful number of people wanted a NatWest-branded digital bank badly enough to switch to it.
France offers the same contrast. Nickel did not build a branch network to prove its model. It struck a deal to distribute current accounts through the existing bar-tabac network, tobacconists that already had a counter, a till and daily foot traffic, and launched nationally on 11 February 2014 already live in dozens of shops. Within eight months it had around 470 outlets carrying the product and 45,000 accounts opened, according to a French government report on financial inclusion published that October. By the end of its first year it had passed 70,000 clients. Meaningful outside capital, including Partech Ventures' €10 million round in 2015, arrived only once that low-cost distribution test had already produced a large base of real customers. When BNP Paribas acquired an 89.1% stake in July 2017, Nickel had more than 630,000 accounts open, according to BNP Paribas's own announcement of the deal. The test was not a lab pilot confined to a handful of branches. It was the real product, sold through infrastructure Nickel never had to build, proving demand before the large capital arrived rather than after.
Orange took the harder road. The telecoms group built Orange Bank from the ground up, launching in 2017 with an internal target of 5 million customers by 2023. By the time it agreed to sell the business to BNP Paribas in June 2023, it had reached only around 2 million customers and had lost more than €800 million since launch. Orange had advantages many fintech entrants do not have: an existing customer base, a trusted brand and deep pockets. What it had not established was that mobile customers actually wanted their bank to come from their phone operator. Convergence between telecoms and banking, it turned out, had limits that a six-year, nine-figure investment did not change.
The turn
The pattern across all four cases has nothing to do with sector, size or ambition. NatWest and Orange had more capital, more brand recognition and more existing customers than Monzo or Nickel ever had at launch. What they did not have was proof, gathered before the money was spent, that the specific product they were building solved a problem people would actually pay to have solved. A ten-day test cannot tell a board whether an idea will succeed. It can tell them, cheaply, whether the idea is worth finding out about properly. Applied early enough, that distinction was worth roughly £100 million to NatWest, in hindsight.
A pattern worth naming
In transformation and consulting work, the same habit shows up repeatedly. The business case gets rigorous scrutiny. The technology plan gets a full governance review. The actual customer problem gets a paragraph in the appendix, assumed rather than tested. Approving a budget looks like decisiveness. Running a ten-day test before approving it looks, to some steering committees, like hesitation. It is the opposite. The teams that insist on testing first are usually the ones that move fastest afterwards, because they are not spending the next eighteen months discovering, expensively, what ten days would have told them for nothing.
Aulay's point of view
The ten-day test belongs earlier in the process than most organisations currently place it: before the business case is written, not alongside it. For a PE-backed platform weighing a fintech bolt-on, that means treating demand validation as part of due diligence, not a workstream that starts after the deal closes. For a corporate innovation team, it means no capital release without a documented, evidenced answer across all seven dimensions, not a slide deck of assumptions presented as findings. For a founder, it means resisting the pressure, from investors or from conviction, to start building before the test is actually finished.
If your organisation is about to commit capital to a fintech idea, product or venture and has not yet run a structured test of demand, that is the conversation worth having before the business case reaches committee, not after.
References
1. CB Insights, "Why Startups Fail: Top 9 Reasons," research report, published 5 March 2026. Analysis of 431 VC-backed startups that shut down since 2023 (385 categorised); poor product-market fit cited in 43% of cases.
https://www.cbinsights.com/research/report/startup-failure-reasons-top/
2. Monzo, "We're Ready," company blog, 30 October 2015. Alpha prepaid card launch capped at 3,000 cards, initial 500 testers in London.
https://monzo.com/blog/2015/10/30/we-are-ready
3. Monzo, "We Are Now a Bank," company blog, 11 August 2016. Restricted UK banking licence granted; over 200,000 sign-ups across Alpha/Beta, c. 30,000 prepaid cards in circulation.
https://monzo.com/blog/2016/08/11/we-are-now-a-bank
4. Monzo, "Welcome to Monzo Bank!," company blog, 5 April 2017. Full banking licence restrictions lifted.
https://monzo.com/blog/2017/04/05/banking-licence
5. FinTech Futures, "RBS closes digital bank Bó six months after launch," trade press, 1 May 2020. Closure details, c. 11,000 customers, CEO Alison Rose statement.
https://www.fintechfutures.com/digital-banking/rbs-closes-digital-bank-b-six-months-after-launch
6. AccountingWEB, "Demise of Bó signals rough patch for digital banks," trade press, 2020. Estimated £100 million development cost over 18 months; launch November 2019.
https://www.accountingweb.co.uk/tech/tech-pulse/demise-of-bo-signals-rough-patch-for-digital-banks
7. Ministère des Solidarités, Conseil national des politiques de lutte contre la pauvreté et l'exclusion sociale (CNLE), "Compte Nickel — 8 mois après le lancement," government report, 16 October 2014. 470 tobacconist outlets, 45,000 accounts opened.
https://solidarites.gouv.fr/sites/solidarite/files/2023-07/COMPTE_NICKEL_CNLE_OCT-14_V2_1_.pdf
8. BNP Paribas, "BNP Paribas completes the acquisition of Compte-Nickel," press release, 12 July 2017. 89.1% stake acquired (rising to 95% by end-2017); over 630,000 accounts at completion.
https://group.bnpparibas/en/press-release/bnp-paribas-completes-acquisition-compte-nickel
9. Banking Dive, "BNP Paribas deal signals end to Orange's challenger bank status," trade press, reporting Orange's 28 June 2023 announcement. Losses over €800 million since 2017; c. 2 million customers against a 5-million target.
https://www.bankingdive.com/news/orange-bnp-paribas-exit-france-spain-retail-hello-bank/654261/
Frequently asked questions
- What counts as a real signal of demand, rather than an opinion?
- A real signal involves the customer doing something that costs them something: paying a deposit, joining a capped waiting list ahead of others, pre-ordering, or giving up time for a pilot. A survey answer with nothing at stake is not a signal. It is a guess with a percentage attached to it.
- Isn't this just an MVP?
- Not quite. A minimum viable product usually tests whether something works once it is built. A ten-day demand test is narrower and faster: it tests whether customers want it, whether the business can reach them, and whether the economics and regulation hold up, before any product exists at all.
- Does this apply to regulated products, where even testing needs permission?
- Yes, with adjustments. Testing interest in a proposition or willingness to pay does not usually require a licence; testing the regulated activity itself, such as taking deposits, generally does. Establishing exactly where that line sits for a specific product is itself one of the seven things worth knowing before capital is committed.