A product manager at a field-services platform opens a spreadsheet on a Tuesday morning. Column A lists 4,200 contractors who take card payments through the software. Column B shows what the platform earns from those payments, which is nothing, because payments were handed to an outside provider in year one and nobody revisited the decision.
Spreadsheets like that one send platform teams looking at processors. Five models cover the market, and the one a platform picks determines the economics, the compliance load, and how much of the money it can see before it reaches somebody else’s bank.
1. The Full-Stack Acquirer
One vendor supplies the gateway, the acquiring relationship, and the reporting. The platform integrates once, the vendor underwrites every seller, and the platform earns a referral fee or nothing at all.
This model fits a company whose product is not about money. A scheduling tool for dentists with 300 practices on it does not want an underwriting function, and buying the whole stack from one provider is the shortest path to accepting cards. The cost arrives later. The vendor sets pricing and rarely negotiates below a certain volume; the platform sees a summary rather than transaction-level economics; and moving to another provider means re-onboarding every merchant by hand. Teams that expect payments to become a revenue line outgrow this model once the referral fee is visibly smaller than the margin an owned relationship would produce.
Support costs land on the platform anyway. A dentist whose Friday deposit has not arrived calls the software company, because the software company is who they know, and the support agent has no way to see the transaction. The platform ends up staffing a queue for a product it does not control and cannot fix, which is a real operating cost charged against a referral fee that may be worth a few basis points.
2. The Embedded Payments Provider
The second model keeps the platform in the middle. A provider supplying embedded payments for SaaS platforms handles the regulated pieces underneath while the platform owns the merchant relationship, sets the pricing its customers see, and keeps the spread between that price and its own cost.
This model fits a platform whose customers already run their business inside the software, where taking money is the next thing those customers need from the software. Onboarding happens in the platform’s own interface, funds move on the platform’s schedule, and the reporting is granular enough to build pricing experiments on.
The platform takes on real operating work in exchange. It now answers support calls about deposits, holds a view on merchant risk, and needs somebody who understands disputes. Companies that treat this as a pricing decision, and staff it that way, discover the gap during their first chargeback wave.
The data usually exists already. A platform that schedules the jobs and stores the invoices knows more about its merchants than any outside underwriter does, which is what makes approval decisions faster and fraud losses lower inside a software product than outside one.
3. The Marketplace Payout Specialist
Marketplaces have a different problem.
Money arrives from a buyer and has to be split between the marketplace and one or more sellers, sometimes across borders, often with a hold in between.
Specialists in this shape handle seller onboarding, identity checks, split settlement, and the tax paperwork that follows. That last part is heavier than most founders expect. Marketplaces issue a Form 1099-K to sellers over the reporting threshold, and the 1099-K tax rules that decide who gets one have moved repeatedly, landing at $5,000 in a single year on a single platform for 2024 after years at $20,000 and 200 transactions.
The agency also postponed the new tax-reporting rule behind the lowest of those thresholds more than once, which is the argument for buying this capability rather than building it. A platform maintaining its own reporting logic rewrites it every time Congress or the agency changes position, and the rewrite lands in January.
4. Payment Facilitator Registration
The fourth model is becoming the payment company. The platform signs on as a payment facilitator with the card networks, holds a sponsorship agreement with an acquiring bank, and underwrites its own sub-merchants.
Economics are the best available here, since the platform keeps the acquiring margin instead of sharing it. The requirements are proportional. Network fees, an annual audit, reserve requirements from the sponsor bank, a compliance officer, and a fraud team are the standing costs before a single transaction settles.
Sponsor bank diligence on a first-time facilitator runs for months and asks for audited financials, written policies on onboarding and monitoring, each with a named owner inside the company. Platforms that start this process in the same quarter they promised payments to a customer discover that the bank sets the calendar.
Marketplaces taking this route inherit statutory duties as well. Federal law now makes online marketplaces verify high-volume sellers, collecting bank account, tax identification, and contact details from anyone with at least 200 unique sales and $5,000 in a year, a rule passed amid retail theft concerns about stolen goods being resold online. A platform that already collects that information for underwriting has most of the work done.
5. The Orchestration Layer
The fifth model operates above the others. An orchestration provider routes transactions across several processors and normalizes the reporting so a finance team reads one ledger, retrying a failure on a second route when the first one declines.
Platforms reach for this at scale, usually after a processor outage or an authorization rate that varies by country. Routing a declined transaction to a second acquirer recovers revenue that would otherwise be gone, and the savings are measurable on large volume.
Orchestration does nothing for a checkout that loses people before the card is entered, which is where a surprising share of platform volume disappears. Research on Website Forms Usability is consistent about the causes, including error messages that hide the actual problem and forms that discard what the user typed. A platform routing 12 million transactions a year should fix the form before it optimizes the routing.
The Contractor and the Spreadsheet
Return to the 4,200 contractors. The platform that owns its payments knows what each of them processes, prices accordingly, and adds a line of revenue that grows with customer success rather than seat count. The platform that does not know is reading a monthly summary from a vendor and hoping the numbers hold.
Whichever of the five models fits, the decision is made once and lived with for years. Column B fills in or stays empty on the strength of two weeks of work in year one, long before anybody opens that spreadsheet.
