How to choose SaaS affiliate commission rates

Start with what you can afford

Choose a rate from your customer economics and the work you want affiliates to do. A percentage on another company's website is not evidence that the same offer will work for you. Its price, service costs, customer retention and commission base may be different.

Start with monthly revenue after expected refunds. Subtract the cost of serving the customer, then the affiliate award. The remainder still has to help fund your fixed overhead and program operations. A positive remainder is not the same thing as profit.

Use the affiliate commission calculator to compare durations while holding your assumptions constant. Change one assumption at a time so you can see what drives the result.

Work through one paid month

Consider an illustrative $100 monthly subscription, 80% gross margin before commissions, no refunds and a 20% affiliate rate. The month produces $80 before the affiliate award, pays $20 in commission, and leaves $60 before fixed overhead and program costs.

If the gross margin is only 30%, the same offer leaves $10. If gross margin is 10%, it leaves a $10 loss. The commission rate did not change; the ability to support it did.

Now assume 10% of revenue is refunded and that both your commission base and modeled service contribution scale with net revenue. Net revenue is $90, the modeled gross contribution is $72, the award is $18, and $54 remains. This is a simplifying assumption. If service costs or payment fees are not recovered when you refund a customer, your remaining contribution will be lower.

Compare the earning duration, not just the rate

With the $100 price, 80% margin and 20% rate above, suppose a customer pays for 24 months with no refunds or churn:

• First invoice only: $20 total commission and $1,900 modeled contribution.
• First 12 monthly invoices: $240 total commission and $1,680 modeled contribution.
• All 24 monthly invoices: $480 total commission and $1,440 modeled contribution.

These are arithmetic scenarios, not predictions. A first-invoice percentage is also different from a fixed acquisition bounty. A $200 bounty on this customer would require its own payback calculation; the calculator's first-invoice scenario does not model that bounty.

An ongoing offer may be attractive to an affiliate willing to invest in durable educational content. A shorter offer limits how long you earn new commission obligations. Discuss the tradeoff with the people you want to recruit instead of assuming one duration always wins.

Stress-test retention and refunds

A fixed-horizon model avoids assuming a customer stays forever. With 5% monthly customer churn, the probability that an initially paying customer reaches month 2 is 95%; month 3 is 90.25%. Month 1 includes the initial paid invoice. There are no new customers added to this cohort.

This assumes the same churn probability each month, no reactivation, no price changes and no expansion revenue. Your real cohorts can behave differently. Compare your observed cohort history with the model before using it for a budget.

Try a higher-churn case, a lower-margin case and a higher-refund case. Record the point at which the offer no longer leaves enough contribution for your operating costs. Do not label a rate “safe” simply because the default calculator scenario is positive.

Make the rules clear before recruitment

Document whether the award uses invoiced or collected revenue, how discounts and tax are treated, and which refunds reverse an award. Specify when the earning period starts and what happens when a customer changes plan. Define approval and payment timing separately: an earned award need not be immediately payable.

Use test transactions to confirm the software matches those rules. For example, Rewardful's campaign documentation describes commission calculations from paid invoices and configurable campaigns. This is an example of why implementation details need checking, not a claim that every platform behaves alike.

Review your initial offer after you have evidence from recruited affiliates and retained customers. Honor existing commitments when changing terms. Start with the launch checklist to keep offer design, tracking and recruitment connected.

Sources and calculation notes

The worked examples and finite-horizon model above are original arithmetic examples from Referral Bear. They are not measured industry averages. The model rounds only displayed currency values to cents, so displayed monthly rows can differ slightly from totals calculated at full precision.

Stripe explains monthly customer churn; Stripe's subscriber definition also shows why canceling one of several paid subscriptions is not necessarily a lost subscriber. Use a consistent customer definition in your own inputs. Sources checked 13 September 2026.