The email always arrives at an awkward hour. “After a review of your account, we have determined that your business falls outside our acceptable use policy. Payouts are on hold for 90 days.” Nothing broke. Sales were fine. Chargebacks were low. What happened is that an automated review flagged a word on your product page — “peptide,” “prescription,” “prediction,” “clinical” — and an aggregator decided your business was no longer worth the tail risk.
If you operate a telehealth platform, a research-compound store, a subscription supplement brand or a market where users stake money on outcomes, this is a structural problem, not bad luck. Below is how the underwriting side actually works, what documentation gets accounts approved, and how to build a payment stack that survives both the bank and the card networks.
The difference between an aggregator and an acquirer
Mainstream platforms like Stripe, Square and PayPal are payment facilitators. They onboard millions of merchants under a shared master account, which means they cannot afford individual judgement calls. Their risk model is categorical: if a vertical has elevated regulatory exposure or historical chargeback volatility, the whole category is restricted — regardless of how clean your individual business is. You are not being evaluated; you are being classified.
A dedicated merchant account works the opposite way. You get your own merchant ID (MID) with a sponsoring acquiring bank that has actually read your business model, priced your risk and agreed to hold it. That is slower and involves more paperwork, but it replaces a policy exclusion with a relationship. It also means the bank can set conditions — a reserve, a volume cap, a required disclosure on your checkout page — instead of simply terminating you.
Why “low chargebacks” isn’t the whole argument
Merchants often lead with their dispute ratio, assuming good numbers settle the matter. Underwriters care about three things beyond that: legal exposure (can a regulator force refunds or shut you down?), delivery risk (is there a gap between the charge and fulfilment?), and reputational risk (does the acquirer want this logo in its portfolio?). Industry trade coverage on how acquirers manage high-risk merchants makes the same point: pricing and monitoring exist because the bank, not the merchant, absorbs losses when a portfolio goes wrong. Argue on all three axes, not just disputes.
What an underwriting file actually contains
Approval speed is almost entirely a function of how complete your first submission is. Expect to supply:
- Six months of processing statements, or bank statements if you are new, plus a projection of monthly volume and average ticket.
- Incorporation documents, EIN, beneficial-ownership details and a government ID for each principal.
- A live, fully populated website — not a staging site — with refund, shipping, privacy and terms pages that match what you actually do.
- Product-level substantiation: what you sell, to whom, with what label language and what claims you deliberately avoid.
- For clinical models: licensing coverage by state, physician oversight structure, pharmacy relationships (503A or 503B), and any DEA registration where it applies.
- Your HIPAA posture, including willingness to sign a Business Associate Agreement with any vendor that touches protected health information.
The fastest declines come from mismatches. A site that promises overnight shipping while your fulfilment partner takes five days, or marketing copy that implies treatment of a condition your intake form never screens for, tells an underwriter that your controls are theatre. Fix the contradictions before you apply, not after a reviewer finds them.
HIPAA, BAAs and the payment page
Telehealth operators frequently discover that their gateway has been quietly collecting data it should never see. If your checkout captures a diagnosis code, an intake answer or a prescription identifier alongside the card, the processing environment is inside your HIPAA perimeter and your vendor needs a signed BAA. Most generalist gateways will not sign one.
The practical fix is architectural: keep clinical data in your EHR, pass only a tokenised customer reference and an amount to the payment layer, and use hosted fields or a hosted checkout so raw card data never touches your servers. This narrows both your HIPAA exposure and your PCI DSS scope in one move. Where the payment step must live inside the clinician workflow, an EHR-native integration is preferable to bolting a consumer checkout onto a medical record.
Recurring refills, FSA/HSA cards and the quiet revenue leak
Subscription models are where high-risk merchants win or bleed. Two mechanics matter more than almost anything else in the stack. The first is a tokenised card vault with an Account Updater service, so expired and reissued cards refresh automatically instead of failing silently — card churn alone can quietly erode a double-digit share of recurring revenue over a year. The second is clear consent: an explicit, timestamped record of what the customer agreed to, how often they will be billed, and how to cancel in one step. That record is your primary chargeback defense exhibit.
FSA and HSA acceptance is a separate capability, not a toggle. It depends on the correct merchant category coding and the right inventory information file treatment, and it only applies to genuinely eligible products and services. Where it does apply, it reduces price sensitivity meaningfully — patients spending pre-tax funds behave differently from patients spending discretionary income.
Where a specialist provider earns its keep
At some point the decision is simply whether to keep reapplying to platforms that have already published a policy excluding you. Specialist providers exist because the alternative — being routed through a single sponsoring bank with a single risk appetite — is fragile. 2Accept, for instance, underwrites high-risk verticals and focuses on HIPAA-compliant telemedicine and online prescribing platforms that mainstream aggregators restrict, connecting merchants to a network of acquiring banks rather than a single relationship, and handling underwriting, gateway integration and chargeback defense as one workflow. The same underwriting machinery covers adjacent categories, including peptide payment processing and prediction market operators, with BAA-backed setups, tokenised recurring billing and FSA/HSA acceptance available where the model qualifies — conditions like physician oversight, 503A/503B pharmacy partnerships or DEA registration still apply, and a specialist will tell you that upfront rather than after your funds are frozen.
Build for redundancy, not perfection
One MID is a single point of failure. Mature high-risk operators run at least two, split traffic deliberately, and keep each one’s volume and dispute ratio inside its agreed limits. If you manage multiple brands — a pattern familiar to anyone maintaining a working catalogue of websites across different niches — resist the temptation to funnel everything through one descriptor. Separate MIDs per brand keep a problem on one site from contaminating the rest.
Then instrument the basics: a recognisable billing descriptor with a support number, pre-dunning emails before each refill, 24-hour refund turnaround, and alerts that catch disputes before they mature. Keep your chargeback ratio comfortably under the roughly 1% threshold the networks use as a monitoring trigger, and keep evidence — consent records, delivery confirmation, clinical documentation where permitted — ready to submit within the representment window.
The practical sequence
- Audit your site and marketing copy against what you can actually substantiate.
- Assemble the underwriting file before you talk to anyone.
- Decide where protected health information lives and keep it out of the payment path.
- Apply through a provider that underwrites your category instead of tolerating it.
- Stand up a second MID once the first is stable, and monitor both weekly.
None of this makes a high-risk business low-risk. It makes it legible — and legibility is what gets an acquirer to say yes and keep saying yes six months later.




