Why I’m Building a Stripe Risk Tool From a Country Stripe Doesn’t Support

--

Zimbabwe is not in Stripe’s list of supported countries. I’m building a Stripe risk monitoring product from Zimbabwe. That’s not the ironic part. The ironic part is that I understand Stripe’s risk system better than most of their merchants do, and I’ve never processed a single payment through it.

Here’s how that happened.

200 Frozen Account Stories

Six months ago I started collecting frozen Stripe account posts from Reddit. Not skimming. Collecting. Cataloguing the trigger, the merchant’s response, the outcome. I wanted to know if there was a pattern.

There is. And it’s not the one most people assume.

Most merchants think Stripe freezes accounts because of high risk. That’s close, but it misses the point. Stripe freezes accounts because of *unexpected* risk. The distinction is the entire game.

The Surprise Principle

A business processing $50,000 a month with a steady 0.4% dispute rate is fine. That same business with a 0.4% rate that was 0.1% three weeks ago is a problem. Same numbers. Same rate. Completely different signal.

Stripe isn’t asking “is this merchant risky?” They’re asking “is this merchant behaving differently than they were last month?”

I started calling this the Surprise Principle, because every frozen account story I catalogued had the same shape: the merchant’s numbers changed faster than the system expected. Not worse. Just different.

Launched a product. Ran a promotion. Got featured somewhere. Revenue spiked. The merchant sees a success story. Stripe’s pattern detector sees something that looks, to an automated system, almost identical to a compromised account ramping up.

There’s no “heads up, we’re about to grow” button. So the system flags the anomaly and freezes first.

Three Signals, One Blind Spot

From the 200+ cases I catalogued, three metrics consistently sit at the center:

Dispute rate trajectory. The Visa VDMP program kicks in at 0.9%. Mastercard ECM at 1.5%. Stripe’s own threshold sits lower than both. But the number that triggers review isn’t the rate itself. It’s the rate of change. A steady 0.6% is safer than a 0.2% that’s climbing.

Refund acceleration. A steady 5% refund rate is a business model. A refund rate that jumped from 2% to 8% in two weeks is a signal that something broke. Stripe doesn’t know if it’s a product quality issue or a shipping delay. They just see the graph change shape.

Transaction velocity spikes. Not just dollar volume. Transaction count matters independently. Going from 200 transactions per month to 2,000 triggers pattern detection regardless of what you’re selling.

And here’s the blind spot: you can see all three numbers in your Stripe dashboard. You cannot see how Stripe’s risk model interprets those numbers relative to your own history. You don’t know if you’re 30 days from a freeze or 300.

Why I’m Building This From the Outside

I didn’t plan on doing so but I had to register a UK company from Zimbabwe because Stripe doesn’t operate here. That part took an afternoon. The part that took weeks was everything after.

To accept payments globally, you need a business bank account. I applied to Tide. Rejected. Non-resident director. Revolut Business. Rejected. Requires EEA residency. Starling. UK residents only. Wise Business. Same. Four applications, four rejections, each one taking days of document uploads and verification steps before the “no” arrived.

Payoneer eventually said yes. One out of five.

What should have been a routine admin task, the kind of thing a founder in London or San Francisco does between lunch and their first meeting, took me the better part of a month. Not because the paperwork was complex. Because the system wasn’t built for someone with my passport.

The global payments infrastructure assumes you live in one of about 40 countries. If you do, none of this registers as a problem. If you don’t, you spend your first month as a founder proving you’re allowed to participate before you write a single line of code.

The irony isn’t lost on me. The problem I’m solving, merchants losing their businesses to a platform they depend on but don’t understand, is a problem that exists everywhere Stripe operates. But the infrastructure to build the solution assumes I’m already inside the system I’m trying to fix.

I’m not inside it. I built PayCanary anyway.

In a way, being on the outside is an advantage. I have no legacy Stripe integration to protect. No existing merchant relationship clouding my judgement about what the platform actually does versus what it says it does. I just have the data from 200 merchants who learned the hard way.

Stripe is managing their downside. Most merchants aren’t managing theirs. The merchants who survive aren’t the ones with the best numbers. They’re the ones whose numbers are predictable.

Check Your Own Numbers

Before you close this tab, do one thing. Open your Stripe dashboard. Go to Payments > Disputes. Look at your dispute rate for the last 3 months. Now compare it to the 3 months before that.

If the number went up, it doesn’t matter that it’s still below 0.9%. What matters is the direction.

Now do the same for refunds. Payments > All payments, filter by refunded. Is the refund rate climbing? Flat? Dropping?

If both are climbing, even slowly, you’re moving toward a review trigger and your dashboard won’t warn you until it’s too late.

If you want to see how all three metrics (disputes, refunds, velocity) interact to produce your actual risk level, I built a free diagnostic that runs the same analysis I used on those 200 Reddit cases. Takes 60 seconds: paycanary.io/diagnostic

The Takeaway

Stripe doesn’t hate risk. They hate surprises.

A steady “okay” is safer than an unstable “great.” The merchants who get frozen aren’t the riskiest ones. They’re the ones whose numbers changed shape in a way that a pattern detector couldn’t predict.

Monitor the trend, not just the number.

I’m curious: if you’ve had a Stripe freeze or a close call, what triggered it? Was it a volume spike, a dispute cluster, or something else entirely? I’m still cataloguing cases and every story sharpens the model.

I write about Stripe risk, payment infrastructure from places it wasn’t designed for, and what it’s actually like to build a SaaS from southern Africa. @gumagatsby everywhere.