How to get a merchant account when Stripe says no

First, understand why Stripe said no

Stripe and PayPal are volume businesses. They automate approval, prohibit entire categories, and cut anything that creates manual work. If you're an exchange, a broker, a gambling operator, or a CBD brand, the rejection was decided before you applied. Nothing about your specific business was evaluated, so don't treat the rejection as a verdict on it. It just means you need a processor whose acquiring bank actually accepts your category, at prices that reflect it.

What high-risk underwriters actually look for

An underwriter is estimating two numbers: how many disputes you'll generate, and how much money is unrecoverable if you fail mid-delivery. Everything they ask maps to one of those two.

They want processing history, ideally three to six months of statements showing your real volume, dispute rate, and refund rate. They want financials proving you can absorb refunds without their money. They want to see the actual owners, because they're checking each one against the MATCH list and past processing failures. And they'll read your website like a compliance officer: terms of service, refund policy, contact details, product claims, and whether what you sell matches what your application says you sell.

New business with no history? You'll still get approved somewhere, but expect the category-average deal: higher rate, bigger reserve, lower volume cap. Terms improve after six months of clean processing. Plan for that instead of resenting it.

The documents to have ready

Get the packet together before you apply, because slow document responses read as disorganization. You'll need: certificate of incorporation and corporate documents, government ID for every beneficial owner, three to six months of bank statements, prior processing statements if any exist, a business bank account in the company's name, and any licenses your vertical requires. Gambling operators need their gaming license. Brokers need their regulatory authorization. CBD merchants need lab certificates for every product. Missing licenses aren't a negotiating point. They're an automatic no.

Red flags that kill applications

Some things end an application regardless of everything else. Lying about your category is the big one. If you're caught after approval, you get terminated for cause and MATCH-listed, which poisons every application for five years. A vague or hidden refund policy is another. So is a website with no company name, no address, and no reachable support, because that's what exit scams look like. Mismatches hurt too: a director on the application who isn't in the corporate registry, a bank account in a different name, revenue claims your statements don't support. Underwriters don't investigate discrepancies. They decline them.

Practical steps before you apply

Clean up the website first: visible terms, refund policy, company details, and accurate product descriptions. Fix your dispute rate before it's on a statement an underwriter reads, which mostly means clear billing descriptors, fast support responses, and honest delivery times. Apply to processors that already serve your vertical instead of blasting applications everywhere, because every decline creates a record the next underwriter sees. Ask about reserve terms, payout timing, and termination triggers before signing, and get the answers in writing. And apply before you're desperate. The worst terms go to merchants applying the week after a termination, with a frozen balance and no leverage.

And set up the rail that doesn't need an application

While you work through underwriting, which takes weeks, you can be taking payments today. Crypto processing has no acquiring bank, so there's no application to pass, no reserve, and no category policy. On Flint you can have a checkout live the same day, in parallel with whatever card account you land. Most merchants who go through a Stripe shutdown keep both rails permanently afterward, because the lesson wasn't that they picked the wrong processor. It was that one processor is one point of failure. The comparison between the two rails is here if you want the details.