I read a lot of business plans for payment companies, and they all look the same. Target market, tech stack, license timeline, and a revenue line that's just volume times margin. Clean. Confident. And silent on the three things that actually decide whether it works.Here's what separates the plans that go on to work: an acquirer who has already said yes, a clear list of the merchants you'll refuse, and enough capital to sit between settlement and payout for as long as the card schemes want you to.Let's get into what actually gates a launch, how to pick a segment before you pick a stack, how the payment gateway business model earns money once you strip out the parts you never keep, why underwriting is your real product, and what a realistic first twelve months looks like.What actually gates the launchThree things have to line up before starting a payment processing company: a license or a licensed sponsor, an acquiring relationship, and merchants who have agreed to move their volume to you. Founders usually plan the first, buy the third late, and underestimate the second.Licensing has a published clock. Authorized payment institution applications in the UK typically take 6 to 12 months from submission to determination, with e-money authorization longer. That timeline is knowable, so it can be planned around. Many new entrants also start under an existing licensed entity as an agent or distributor, which shortens the path considerably.Acquiring has no published clock, and this is where launches slip. When an acquirer takes you on, they're adding your entire future portfolio to their own risk numbers. So their underwriting question isn't really about you — it's about your merchants. Which verticals, which countries, what average ticket size, what refund behavior, what dispute history, who's actually behind each business.And here's the trap. To get the acquiring relationship, you need a credible merchant pipeline. To close merchants, you need the acquiring relationship. The founders who break that loop do it by building the pipeline first, in writing, before they've even picked a tech stack. Signed letters of intent, a named vertical, realistic volume forecasts per merchant, and a clear account of who you'll turn away. That document is your acquiring application. Everything else follows from it.Choose the segment before you choose the stackEvery meaningful decision in a payment business flows from one thing: which merchants you serve. Pricing power, acquirer appetite, risk cost, sales cycle, support load, required capital — they all move together once you pick a segment.Regulated high risk (iGaming, forex). Strong pricing, thin acquiring supply, heavy compliance obligation, and payouts as a core product rather than an afterthought. Merchants here switch providers readily, so approval-rate performance keeps them.Mainstream SMB e-commerce. Wide acquiring supply, transparent pricing pressure from established platforms, and margins that only work at scale or with a distribution advantage you already own.Vertical software and marketplaces. You sell payments into a customer base someone has already assembled. Sales cost drops sharply, but you take on sub-merchant onboarding, split settlement and the risk that comes with them.Local market specialism. One country's methods, one language of support, one regulator. The smallest addressable market on the list and often the most defensible, because global providers cover it badly.Pick one. Then write down what you'll refuse. That refusal list matters more than your target list. It's what keeps your portfolio inside the risk thresholds I'll get to below, and it's the part an acquirer reads closest.The payment gateway business model, priced properlyThe headline rate a merchant pays has very little to do with what you keep. Interchange goes to the issuing bank, and it has been climbing: the average combined Visa and Mastercard credit interchange rate reached 2.36% in 2025. Scheme fees go to the networks. Your revenue is the remainder, and on an interchange-plus-plus contract that remainder is a visible, negotiated line.Markups track risk more than anything else. A rough rule of thumb across providers: around 100 basis points for high-risk categories, dropping to 50 basis points or less for lower-risk, higher-volume merchants. On a blended contract, you absorb the card-mix risk instead of the merchant, which makes your quote look better upfront and eats your margin the moment premium cards show up.Six revenue lines make up a working model. These shapes reflect what we actually see in deals with payment companies, not a published tariff, so treat them as a starting frame for your own numbers rather than gospel.Run the numbers on one merchant before you run them on a market. Take a merchant processing two million euros a month at a forty-basis-point markup. That's eight thousand euros of gross margin, before platform costs, dispute handling, support, and the fraud you didn't catch. Fifteen merchants at that size make a business. Signing fifteen of them, at three to six months per deal — that's the real launch plan, not the pitch deck version.Two adjustments make the model honest. Model refunds and declines explicitly, because gross volume isn't settled volume. And model your card mix, because the gap between debit-heavy and premium-credit-heavy traffic can move your retained margin more than any pricing negotiation you'll ever win.Underwriting is your productNew providers like to describe their product as technology and their sales motion as relationships. But the thing that actually decides whether the business survives its second year is who you let in.The scheme rules make this concrete. Under Visa's Acquirer Monitoring Program, the merchant excessive threshold dropped to 1.5% on 1 April 2026, with acquirer portfolio thresholds at 0.50% and 0.70%, enforcement fees per fraud or dispute event, and a minimum of 1,500 combined events before a merchant is assessed. Mastercard runs its own program at a comparable ratio.Now read that from your acquirer's chair. Their portfolio ceiling sits far below any single merchant's own limits, and every merchant you board lives inside that ceiling. One merchant having a bad quarter can push the whole portfolio ratio, and the correction shows up as your MID getting restricted or pulled entirely. That cuts off every other merchant you signed, not just the problem one. That's the actual mechanism by which promising payment companies die. And it's entirely preventable.Four disciplines prevent it, and they need to exist before merchant number one:Board with evidence. Ultimate beneficial owners, processing history with prior providers, the website as it will actually run, refund policy, and expected volume by country and method.Monitor per merchant, daily. Dispute ratio, fraud reports, approval rate, refund rate, and volume against forecast. Monthly reporting finds problems a month after they mattered.Hold reserves that match the risk. A rolling reserve on a young merchant is not a punishment; it is the reason you can offer them terms at all. Set the percentage and the release period per merchant, not per portfolio.Write the exit criteria down. Decide in advance which numbers trigger a pricing change, a volume cap, or an offboarding, and put them in the merchant agreement so the conversation is contractual rather than personal.The cash flow question nobody asks early enoughPayment businesses fail on working capital more often than on technology. Money arrives from the acquirer on a settlement cycle, leaves to the merchant on the schedule you promised, and disputes can claw it back for months afterward. You sit in the middle of those timings with your own balance sheet.Three positions to size before launch. The reserve you hold from merchants against the reserve your acquirer holds from you, since the gap is funded by you. The settlement lag between receiving funds and paying merchants, especially where you have promised faster payouts as a selling point. And the dispute tail, where liability lands on you if the merchant has already been paid and has no reserve left.Regulatory capital adds a further layer. Beyond the minimums attached to an authorization, safeguarding rules keep customer funds separate from your own, so those balances cannot smooth your operating cash flow. Plan capital as a permanent operating requirement rather than a one-time application cost.The team the first year actually needsBuying your infrastructure instead of building it changes the shape of your hiring plan more than the size of it. Engineering headcount drops. Commercial, risk, and operations headcount doesn't.Commercial. Someone who can close merchants and someone who can onboard them properly. These are different skills, and combining them into one role is the most common early-stage mistake I see.Risk and compliance. Underwriting, AML and monitoring, in-house from the start. Outsourcing the decision about which merchants you accept means outsourcing the business.Payment operations. The person who owns approval rates, routing rules, reconciliation and provider relationships. Our analysis of 112 payment management job descriptions for The State of the Payment Manager Role 2026 found approval rate to be the most cited KPI in the role, with reporting and analytics appearing far more often than provider administration. Merchants will judge you on exactly that number.Support. Payment support is technical and time-sensitive. Merchants forgive a decline far more readily than a slow answer about where their settlement is.A realistic first 12 monthsThe sequence below reflects how the launches that go well tend to run. Several tracks overlap deliberately, because the licensing clock runs whether or not you use the time.Months 1 to 2. Choose the segment, write the acceptance and refusal criteria, and build a named merchant pipeline with volume estimates. Start the licensing route in parallel, whether direct authorization or an agent model.Months 2 to 4. Approach acquirers with the pipeline document. Expect several conversations to end in no, and treat each no as underwriting feedback on your target portfolio rather than a rejection of your company.Months 3 to 5. Select infrastructure and pricing. Decide your published rate card, your reserve policy, and your payout promise together, since each one constrains the others.Months 5 to 7. Deploy the platform, connect providers, configure routing, and run end-to-end tests including failed payments, refunds, and a full reconciliation cycle. Onboard a friendly-first merchant at limited volume.Months 7 to 12. Scale merchant by merchant, monitoring dispute ratios and approval rates per account. Add a second acquiring relationship before you need it, because the moment you need it is the moment you cannot get it. What the market expects from a new providerMerchant expectations have moved. Our State of Payment Maturity 2025 research, based on assessments from 672 businesses worldwide, found the share of companies running five or more payment providers rising from 24.6% to 37.1% in a single year. Your prospective merchants are already comparing you against several incumbents, and they will route volume by performance. A single acquirer and a single method is no longer a viable opening offer, which is why most new entrants launch on a white label payment gateway platform rather than assembling connections one at a time.That's the practical case for buying the infrastructure and spending your capital where it actually decides the outcome: distribution, underwriting, and the merchant relationship. The founders who start a payment processing company well are rarely the ones with the most capital or the cleverest tech. They're the ones who knew exactly which merchants they wanted, could prove it to an acquirer, priced with the card mix in mind, and had already decided who they'd turn away.No#PaymentProcessingOleg IshchenkoSales ManagerCorefy13 Aug, 2026