It usually arrives on a Tuesday. A short message from an aggregator's risk team saying the account has been deactivated, payouts are on a 90 to 120 day reserve, and the decision is final. No phone number, no appeal, and a business that was taking card payments at nine in the morning cannot take them at ten. For an online clinic, a research-peptide retailer or an event-contracts platform, that single email is an existential event — not a billing inconvenience.
This is not bad luck, and it is rarely fraud. It is the predictable result of how flat-rate payment aggregators are built, and of a mismatch between what those platforms can underwrite and what these businesses actually sell. Understanding the mechanics is the difference between building a payment stack that survives growth and building one that collapses the week you hit scale.
Aggregators sell convenience, not underwriting
When you sign up with a mainstream aggregator in four minutes, nobody underwrote you. You were placed under a master merchant identification number shared with hundreds of thousands of other sellers, with a published prohibited-business list standing in for due diligence. That model is brilliant for a coffee shop and structurally hostile to anything unusual. The aggregator carries the liability for every account beneath its master MID, so its only practical risk control is removal — fast, automated and without conversation.
Several triggers show up again and again. A sudden jump in volume that does not match the stated business description. Refund or chargeback ratios drifting toward the card brands' monitoring thresholds. A product page that uses the word "prescription", "injectable", "dosage" or "guaranteed returns". A customer service complaint routed to the acquirer rather than to you. A card brand enquiry about merchant category code accuracy. Any one of them can move an account from "fine" to "reviewed", and review inside an aggregator usually means termination rather than remediation.
Being classified as high-risk is also not a judgement on your ethics. Acquirers use the label to describe exposure: how likely a transaction is to be disputed, how regulated the product is, how long between payment and fulfilment, and how much of the revenue is recurring. A plain-English breakdown of how credit card processing works for high-risk businesses is worth reading before any underwriting conversation, because it reframes the discussion from "am I acceptable?" to "what evidence reduces the acquirer's exposure?".
What dedicated underwriting actually asks for
A direct merchant account means a bank looks at your specific business and assigns you your own MID. The trade-off is paperwork and time, measured in days rather than minutes. Expect to supply most of the following:
- Three to six months of processing statements, or bank statements if you are new, plus a realistic monthly volume and average ticket forecast.
- Incorporation documents, ownership breakdown and ID for anyone holding a material stake.
- A live website with visible terms, refund policy, shipping timelines, contact details and accurate product claims.
- For telehealth and online prescribing: physician licensing by state, the intake and consultation workflow, pharmacy fulfilment arrangements, and clear documentation if 503A or 503B compounding partners are involved. Controlled substances bring DEA registration into scope and are handled as a separate category entirely.
- For supplement and peptide sellers: labelling, intended-use language, substantiation for any claim on the page, and fulfilment or lab documentation.
- For prediction markets: the regulatory basis for the product, geofencing and KYC controls, and how settlement is handled.
- Chargeback history with reason codes, and whatever prevention tooling you already run.
The HIPAA layer that aggregators cannot sign
If your checkout sits next to a clinical intake form, your payment flow may touch protected health information — or sit close enough to it that an auditor will ask. The practical requirement is a business associate agreement, plus a technical design that keeps diagnosis, medication names and visit notes out of transaction descriptors and receipts. Most flat-rate aggregators will not execute a BAA for this use case, which is why clinics discover the problem during a security review rather than at signup. Tokenisation helps on both fronts: card data lives in a vault, your systems hold references, and PCI scope shrinks at the same time.
Recurring refills, FSA/HSA cards and the economics underneath
Subscription revenue is the reason these businesses are valuable and the reason they attract disputes. Involuntary churn from expired or reissued cards quietly kills 5 to 10 percent of a recurring book each year unless you run an account updater against the vault. Voluntary disputes come from descriptor confusion, surprise renewals and shipments that arrive after the cardholder forgot the plan existed. The fixes are unglamorous: recognisable descriptors, pre-renewal notices, one-click pause, and dunning that retries intelligently instead of hammering a declined card.
FSA and HSA acceptance is a separate capability, not a toggle. Those cards route through benefit-eligibility rules, so your product and merchant categorisation have to line up with what the administrator will approve, and your receipts need enough itemisation to survive a substantiation request. Done properly it lifts conversion on eligible items noticeably, because the customer is spending pre-tax money.
There is a wider structural story here too. Platforms increasingly want to own the payment layer rather than rent it, and the economics of white-label payment platforms and why fintechs are buying the stack outright explain the same incentive from the other direction: margin, data and control over who gets approved. A high-risk merchant chasing a dedicated MID and a fintech acquiring its own processing capability are responding to the identical weakness in rented infrastructure.
Where a specialist provider earns its place
Once a vertical has been declined by the obvious platforms, the useful question is who has already placed similar businesses with acquiring banks and knows which file gets approved. 2Accept works in exactly that gap, underwriting verticals the big aggregators restrict — HIPAA-compliant telemedicine and online prescribing platforms, prediction markets, and peptide payment processing — by connecting merchants into a network of more than 40 acquiring banks and handling the underwriting file, gateway integration and chargeback defence rather than leaving it with the operator. The practical appeal is that the awkward conditions are treated as conditions rather than rejections: physician oversight, 503A or 503B pharmacy partners and DEA registration where relevant, BAA-backed setups, tokenised vaults with Account Updater for refill billing, FSA/HSA acceptance, and API, hosted checkout or EHR-native integration. With roughly 48-hour approval averages and a 48 to 72 hour window to go live, the timeline is closer to a normal launch than the months merchants fear.
Build for the second processor before you need it
The merchants who survive a termination are the ones who prepared for it while things were calm. Keep your underwriting pack current — statements, licences, policies, fulfilment evidence — so a new application takes hours, not weeks. Use a gateway abstraction or a vault you can port, so cards are not hostage to one processor. Watch your dispute ratio monthly against card brand thresholds instead of annually. Split volume across more than one MID once you are large enough to justify it, and make sure representment evidence is assembled automatically rather than by whoever has a free afternoon.
None of this guarantees an approval, and no provider should promise one. What it does is move you from being a line item on someone else's risk list to being a merchant a bank can actually assess — which, for a telehealth clinic or a peptide retailer with real revenue, is the whole difference between a business and a liability.


