CPA vs CPL vs revshare vs hybrid: choosing a payout model
The payout model decides who carries which risk — and whether the deal survives contact with real traffic. This guide compares CPA, CPL, FTD, sale, COD, revshare and hybrid honestly: what each pays for, who bears the risk, and how to choose for your funnel.
The payout model map
Every CPA-family deal pays for a different moment of truth. CPA proper pays for a defined action — whatever you agree it to be. CPL pays for a lead: a registration or contact that passes qualification. FTD pays for a first deposit. Sale or CPS pays for a completed purchase. COD pays for a confirmed cash-on-delivery order. Revshare pays a percentage of the revenue the user generates over time. Hybrid splits the difference: a smaller fixed payment plus a revenue share. Choosing among them is not about which is "best" — it is about which one aligns the buyer's risk with the moment your business actually earns.
| Model | You pay for | Risk sits with | Typical fit |
|---|---|---|---|
| CPA | Any defined action | Shared, per definition | General performance deals |
| CPL | Qualified lead | Buyer carries volume risk | Dating, sweeps, education, finance leads |
| FTD | First deposit | Buyer carries most media risk | iGaming, forex, crypto |
| Sale / CPS | Completed purchase | Buyer; refunds shared per terms | E-commerce, subscriptions |
| COD | Confirmed order | Shared; delivery risk on you | Nutra in COD markets |
| Revshare | % of user revenue | You carry volume, buyer carries patience | High-LTV products, established brands |
| Hybrid | Fixed + revenue share | Split per structure | Scaling proven offers |
CPA and CPL: paying for actions and leads
CPL is the lightest commitment: you pay for a lead that meets your criteria — valid contact, target GEO, perhaps a verification step. It produces volume fast, which makes it the natural choice for funnels whose value is proven downstream and for advertisers with strong internal conversion teams. Its weakness is symmetric: because the event is shallow, the buyer's incentive stops at the lead. If your monetization depends on what the lead does next, you must measure that per source and gate scaling on it, or you will buy impressive lead counts that never pay.
CPA in the broader sense simply means the billable event is whatever you define — which is why the definition work matters more than the label. The model's strength is precision: pay exactly for the moment value starts. Its weakness is operational: the deeper the event, the slower validation, and slow validation degrades the optimization loop. Choose the deepest event you can validate quickly, not the deepest one you can imagine.
FTD and sale: paying at the money moment
FTD and sale deals pay at the moment money first moves — a deposit, a purchase. This aligns incentives tightly: the buyer earns only when your revenue starts, so media risk sits mostly with the buyer, and payouts are correspondingly higher per event. These models dominate iGaming, trading and crypto (FTD) and e-commerce (sale), where the first transaction is a clean, verifiable milestone.
The trade-offs live in the qualifiers. A bare FTD invites bonus hunters — minimum deposit, trigger the payout, disappear. Qualifiers like wager requirements or second deposits filter them, but every qualifier suppresses the buyer's conversion rate and raises the required payout. For sales, the equivalent issue is refunds: define whether the payout survives a refund and within what window. In both models, the health of the deal is measured downstream — repeat deposits, repeat purchases — so share cohort data per source and let the buyer optimize for users who stay.
COD: the approve-rate business
COD — cash on delivery — is the dominant e-commerce and nutra model in markets where card penetration and online-payment trust are low: much of Eastern Europe, Asia, LatAm and MENA. The user orders online and pays the courier; you pay for the confirmed order, typically after your call center verifies it by phone. The model produces large order volume, but its economics hinge entirely on the approve rate: the share of orders that confirm, deliver and get paid.
What makes COD different is how much of the outcome sits on your side. Call-center speed is the biggest single driver — leads contacted within minutes confirm at multiples of leads contacted hours later. Delivery coverage, courier reliability and product-expectation match follow. The buyer controls source quality and pre-lander honesty; the rest is your operation. This is why COD deals are negotiated against your historical approve rate, and why improving confirmation speed raises the payout you can sustain more reliably than any traffic change. Expect terms — payout, caps, hold — to be rebuilt around that number.
Revshare and hybrid: partnering on lifetime value
Revshare pays the traffic partner a percentage of the revenue the user generates — for months or for the user's lifetime. It is the model of high-LTV verticals: iGaming, trading, subscription products. For the advertiser it eliminates media risk almost entirely — you pay from revenue, not budget. For the partner it converts traffic into an asset, which is why serious buyers accept lower immediate returns for revshare on products they trust. What it demands is trust infrastructure: transparent revenue reporting, agreed definitions (net gaming revenue, negative carryover, duration), and a partner you expect to work with for years.
Hybrid deals split the difference: a reduced fixed payment per action plus a smaller revenue share. They are the natural evolution of a proven CPA relationship — the pilot validated the economics, both sides want scale, and hybrid shares the risk of growing volume. Hybrid also smooths the incentive problem of pure CPA: the fixed payment covers the buyer's media risk, while the revenue share keeps them invested in cohort quality. If a CPA relationship is working and both sides want more of it, hybrid is usually the honest next conversation.
How to choose: a decision framework
Four questions pick the model. First: where does your revenue actually start? Pay at that moment or as close to it as your validation speed allows. Second: how fast can you validate? Deep events you confirm in days beat shallow events you confirm in weeks — the optimization loop needs feedback speed more than it needs depth. Third: how much volume risk can you carry? Revshare transfers it to you; CPL transfers it to the buyer; hybrids split it. Fourth: how proven is the funnel? Unproven funnels belong in capped CPA pilots, where the buyer prices the unknown; proven ones earn revshare and hybrid conversations.
And the meta-rule: models can change as trust builds. Most durable partnerships start with a capped CPA or FTD pilot — limited risk, fast learning — and evolve toward hybrid or revshare as the data proves the economics. What does not work is starting with revshare on an unproven funnel, or demanding FTD payouts sized for a funnel you have not measured. Bring your unit economics to the breakeven calculator, and bring the offer to the brief — the model discussion starts from your numbers, and terms are agreed in writing before traffic runs.