Every fintech PM has had this week.
You walk into the Monday roadmap review with 22 features. Engineering needs 15. Compliance wants 8 of them delayed. Sales wants 6 of them yesterday. The CEO has just read a VC deck and wants to build an agent. By Friday the roadmap has 27 features and nothing has been cut.
I’ve been that PM. I’m still that PM, sometimes.
The problem isn’t prioritization frameworks. You’ve read them. RICE. MoSCoW. Kano. Weighted shortest job first. Every one of them assumes you have good data, clean estimates, and an organization that agrees on what “impact” means. Fintech almost never has any of those three.
Last year a Series B fintech client cut their 28-feature backlog down to 17 in a single afternoon using a matrix that fits on one page. Six weeks later they shipped. The matrix I’ll show you below isn’t revolutionary; it’s just scoped for the real constraints of fintech product.
Why standard scoring breaks in fintech
RICE assigns confidence scores. In fintech, the “confidence” variable collapses the moment legal opens its mouth. You can’t estimate confidence in a feature that depends on a license you don’t yet own.
MoSCoW assigns must-have. In fintech, every feature feels must-have because “not must-have” usually means “launches without a compliance story” — and that’s unshipable. So everything ends up in Must, and the framework degrades into a labeling exercise.
Kano maps delighters. Fintech rarely delights on features — it wins on trust. The user doesn’t care if your transfer is 2.3 seconds or 11 seconds. They care whether the money shows up and whether they understood what happened.
All three frameworks assume the product team controls the scoring dimensions. In fintech, compliance, licensing, and rail-cycle timing are exogenous. Your scoring needs to include variables you don’t fully control.
The framework: Five-column fintech scoring
One table. Five columns. Score each feature 1 to 5 on each.
Revenue impact — How much money does this drive in year 1? (PM + Finance score)
Compliance cost — Weeks of legal/license work before ship. (Compliance advisor scores)
Rail dependency — How many external payment rails does this touch? More = slower, riskier. (Engineering + PM score)
Trust visibility — Does the user see this? Does it shape perception of whether we’re “safe”? (PM + Design score)
Reversibility — If we ship this wrong, how expensive to fix? (PM + Engineering score)
Score each 1 (low / cheap / safe) to 5 (high / expensive / risky).
The MVP cut line is simple: sum the five scores. Anything above 15 is out of the MVP unless Revenue Impact alone is 4 or higher. That rule keeps high-revenue features in even when they’re complex, while cutting features that are moderately valuable but expensive on every other axis.
No weighting. No confidence multipliers. Fintech teams spend more time debating scoring formulas than actually shipping. Flat sums are faster and good enough.
The client cut — 28 features to 17 in four hours
A treasury-product Series B I’ll call “Meridian” came to me with 28 features across payments, reporting, and a first-cut AI assistant. Founders were pressing to ship all 28 within the same 90-day window they’d pitched to investors.
Monday afternoon, we put the five-column matrix on a whiteboard and went feature-by-feature. Rules of the room: only the person listed in the “Who scores it” column could score. No arguing, only asking.
What surfaced:
Fourteen features scored under 12. Easy keeps. Most were internal tooling, admin dashboards, and reporting — cheap on every axis, moderate revenue.
Three features scored 16 to 18 but had Revenue Impact 5. Two payment rails and the pricing API. Kept, despite complexity, because they were the revenue story itself.
Eleven features scored 16+ with Revenue Impact 3 or lower. Cut. Every one of them was a “while we’re at it” feature — the founders had added them over the months leading up to the roadmap lock. The AI assistant was one of them (scored 19: high compliance cost, high rail dependency, high reversibility risk, moderate revenue, high trust visibility). Cut to Phase 2.
Total cuts: 11. Remaining: 17. Meeting ended in four hours including a lunch break.
The interesting outcome: the founders were relieved. Not disappointed. Relieved. They’d been carrying the weight of a roadmap they didn’t believe in. Having a framework that gave them permission to cut, backed by people they trusted, felt like unwrapping a burden.
Six weeks later, Meridian shipped 17 features. On schedule. Two of the cut features came back in the Phase 2 roadmap; nine never did — they’d been roadmap-cruft all along.
The framework’s hidden virtue
The scoring isn’t the point. The conversation is.
When a compliance advisor scores a feature 5 on compliance cost, the PM and the engineer hear exactly how expensive that feature is, in weeks, in the same room. When an engineer scores rail dependency 4, the founder sees the feature’s cost isn’t just build effort — it’s integration fragility.
Most fintech features are killed by asymmetric information. The PM doesn’t know the license cost. The engineer doesn’t know the rail cycle. The founder doesn’t know the licensing timeline. The matrix forces each role to surface their knowledge at the same time, in the same cell.
Prioritization frameworks usually fail because they make one person decide with imperfect information. This one makes three people score with their own information, then adds the scores.
What to do Monday
Three things this week:
Print out your current roadmap. One feature per row. Add five empty columns. Put it on a wall in the room where your next product review happens.
2. Tell your compliance advisor and your lead engineer: “I’m going to ask you to score features 1 to 5 on one column each. I’m not asking you to fill in any other column.” This pre-commits them to the framework. Takes two minutes per conversation.
3. In your next product review, spend 60 minutes scoring the current roadmap. Cut everything over 15 unless Revenue Impact is 4 or 5. If your team can’t sum to the same number, that’s the real problem — it means you don’t share a definition of “expensive.” Fix that first.
If you only do one: #3.
If you’re looking at a roadmap with 25+ features and a 90-day window and wondering what to cut, I run 2-week MVP scoping sprints that produce a build-ready spec with a defended cut list. Book a strategy call at cal.com/barakazuga/strategy-call.
— Barak
The Regulated MVP is a weekly tactical brief for fintech and AI-in-finance product leaders. Every Tuesday.