Top 10 Lessons: What Launching a Neo Bank Taught Me About the Real Fintech Game
Part 01
I spent past few years of my career as a product manager helping launch a Experian Smart Money checking and savings account. I learned more in that launch than in years of reading playbooks.
Here’s the uncomfortable truth I’d put on a whiteboard before writing a single line of code: consumer fintech is an engagement business wearing a banking costume. The product looks like a bank. The economics are something else entirely. The companies that survive understand the difference. The ones that die spent their runway optimizing the costume.
So here it is — the unfiltered version, for the founders, PMs, and operators building in fintech right now.
Below are the top 10 lessons that cost me the most — with the scars attached.
1. Your backend partner will make or break you. Cheaper is never cheaper.
Banking-as-a-service providers — your Marqetas, your Galileos — are not a line item you optimize down. They’re the foundation the entire product sits on.
When I was evaluating partners, the cheaper quote was tempting. But if a provider is undercutting everyone, there’s a reason, and that reason surfaces at the worst possible moment: a settlement delay, a compliance gap, a support ticket sitting untouched for three days while your users are locked out of their own money.
We lived it. We had to switch our banking provider right around the one-year mark, and by then the cost savings we’d penciled in at the start were already gone. We paid twice — once for the cheaper option that didn’t hold up, and again for a re-platform that ate the better part of two quarters of roadmap and pulled most of the engineering team off everything else. A re-platform is one of the most disruptive things you can put a young fintech through, and we chose it because the original “savings” were never real to begin with.
Pick the partner who’s proven — or at least well-documented — and reasonably priced, not the one with the lowest quote and the slickest deck. And here’s the very first thing your team should actually do: read the platform service provider’s API docs. Not the marketing slides, the docs. How clear they are, how complete, how recently they were updated — that tells you more about what working with that partner will actually feel like than any sales call ever will. Shiny slides are designed to win the deal; Tech docs are what you live in for the next few years.
2. Double Digit Paid Memberships and Fintech are two different games
This was the most counterintuitive lesson, and it cost us real time to unlearn.
We tied the Savings accounts to a premium Experian membership. On paper it was clean: bundle the financial product with the membership, drive both at once. In reality they’re playing different games with different scoreboards.
The friction in plain terms: a premium member already paying a monthly fee ($24.99 or more) gets credit insights, financial tooling, and a higher APY. That math works for them. But for someone who only wants free checking and savings, paying a high membership fee just to unlock a better rate makes no sense — the other benefits hold no value for that user.
The result: free users overwhelmingly didn’t upgrade. The membership became a gate for neo bank, not a magnet — and a gate is the last thing you want between a user and the moment they fund their account.
Attaching an account to a membership can absolutely grow your membership. Just don’t confuse that with growing your fintech. They aren’t the same business.
3. The North Star is interchange. Everything else is a vanity metric in disguise.
If you take one thing from this post, take this — and then let me argue against myself, because the smartest people reading this are already objecting.
The real fintech game is interchange. It’s the economic engine underneath nearly every consumer fintech that actually survives, and it only shows up when people use their card repeatedly, habitually, as their primary spending account. Which means your real North Star metrics are monthly active users and average monthly balance — not signups.
Now the objection I know you’re forming: “You’re telling me to build my entire economic engine on a revenue source regulators have been actively compressing for fifteen years?” Fair. Durbin capped regulated debit interchange. The $10B-in-assets threshold means the day you scale into ‘regulated’ status, your per-swipe economics get worse, not better. Networks fight over fees constantly. Interchange isn’t bedrock — it’s a fault line, and pretending otherwise is how you get blindsided.
I still believe it’s the right North Star, with eyes open. Here’s why: even compressed, interchange is the only major revenue stream that scales in lockstep with the behavior you actually want — engaged, primary-account usage. Every alternative I’ve watched founders reach for (membership fees, lending you’re not ready to underwrite, “premium” tiers nobody buys) either doesn’t scale with usage or arrives with its own regulatory baggage. Interchange’s ceiling is lower than the hype suggests. Its floor, tied to real engagement, is steadier than anything else on the menu. You plan for the compression. You don’t abandon the engine.
4. Speed of payments is a fraud weapon and a revenue stream at once.
ACH returns — the R01s (insufficient funds) and R10s (unauthorized) — are a quiet killer. Cross the wrong thresholds on your return rates and you draw regulatory scrutiny that alone can sink a young fintech.
We attacked it two ways. First, we integrated an AI model we called a Payment Success Indicator — a score, informed by a member’s actual usage patterns, estimating how likely a payment return. Below a confidence threshold, the payment didn’t go through; the user waited or adjusted transfer amount.
I want to be honest about the cost of that control, because the version of this story where it’s a clean win is a lie. Blocking a payment a model merely predicts will fail means you will sometimes block a legitimate one — and a false positive on someone’s rent is not a rounding error to that person. We frustrated real users. We had to tune the threshold constantly, watch for any pattern that looked like it disadvantaged a group of users, and accept that “reduced our exposure” came with “annoyed a meaningful slice of good customers.” It was still the right call. It was not a free one.
Second, we researched on same-day ACH, which clearly shwed how it can cut our return rates — faster settlement means less time for balances to vanish and disputes to form. And here’s what too many fintechs overlook: faster payments are also a revenue line. Most successful fintechs charge for card-based faster transfers. Skipping it leaves recurring revenue on the floor and, in my experience, lengthens your path to net positive.
Speed reduces fraud, increases engagement, and generates revenue. Few levers do all three.
To be continued…