By Jimi Barkway · Published 1 September 2026 · Part of the honest manual
The short version
The published averages are a distribution, not an instruction. Work your commission from your own gross margin, churn and payback period, and you'll land on a number you can defend to a partner and afford for years. Then write down why you picked it, because in eight months you won't remember.
Ask "what commission should I pay affiliates" and the internet answers "20 to 30%". That number isn't wrong, exactly. It's an average across thousands of companies with different prices, margins and churn, which makes it about as useful to you as the average shoe size.
This module works the number out from your own economics instead. It takes twenty minutes with a calculator, and at the end you'll have something better than a percentage: a reason.
Why you can't copy the number
The biggest public dataset on SaaS affiliate commissions covers 250 programs and $68.4 million of referred revenue over twelve months (published by one of the category's tracking platforms; figures re-checked against the live report, 1 September 2026). The headline: programs earning over $1 million from affiliates pay an average commission of 24.5%. Programs under $100k average 22.1%.
Now the part the headline hides. The standard deviation on those averages runs to about eight points, and the full range in the data stretches from under 1% to over 56%. Companies paying 10% and companies paying 40% are both in there, both presumably on purpose, and the average sits between them describing neither.
An average with a spread that wide isn't an input. It's a sanity check. If your worked-out number lands somewhere between 15% and 35%, you're in the crowd. How you get to your number is the actual work, so here it is.
Start from what a customer is worth to you
Three numbers you already have, or can get tonight:
- Gross margin per customer per month. Price, minus payment processing, hosting and the support cost of one customer. A $49 plan at 80% margin keeps about $39 a month.
- How long a customer stays. Divide 1 by your monthly churn. At 5% churn the average customer stays about 20 months. At 10%, about 10.
- What a customer costs you through other channels. If ads or outbound land you a customer for $250, that's your benchmark CAC.
Multiply the first two and you get lifetime gross margin: our $49 customer at 5% churn is worth about $784. Now the commission has something to be a percentage of.
A 20% recurring commission on that customer costs $9.80 a month for as long as they stay, roughly $196 over their life. You gave up a quarter of the margin to acquire it, you paid nothing up front, and you paid nothing at all for the referrals that didn't convert. Put that next to $250 of ad spend paid in advance, win or lose, and you can see why this channel suits a bootstrapped company: a commission is a CAC you only pay on success.
Run the same numbers on a $9 product with 12% churn and the story flips. About $7 of monthly margin, an 8-month customer life, $56 of lifetime margin. A 20% recurring commission takes $11 of that $56, and the admin of managing the partner eats a chunk of the rest. The channel isn't wrong. The unit economics are, and module 0 already told you what to do about that.
Recurring, one-time, or something in between?
Recurring for the customer's lifetime is the strongest pitch to a partner: they build an income stream, not a bounty list. It's also the most dangerous default, because it's a permanent haircut on your gross margin that compounds as the program grows. And here's the bit almost nobody says out loud: a lifetime commission is close to impossible to reduce later. Ever. Cutting it is taking income away from the exact people who promote you, and they will tell everyone.
One-time (say, 100% of the first month) caps your cost and is trivial to reason about, but the partner's incentive ends at the sale. They'll send you traffic. Sticking is your problem.
The middle path most SaaS programs land on: recurring with a duration. 20 to 25% for 12 or 24 months gives the partner a real stream, keeps the haircut finite, and leaves you room to be generous later rather than stingy later. Generous later is a delightful conversation. Stingy later is a revolt.
Whichever you pick, do the compounding maths once: if affiliates ever drive 30% of your revenue and you pay 20% for life, you've committed 6% of total revenue, forever. That can be a perfectly good deal. It should also be a decision you can show your accountant, not a default you inherited from a dropdown.
The cookie window is a commercial decision
The cookie window is how long after a click a new customer still credits the affiliate. Ours defaults to 60 days, and for subscription SaaS a longer window costs you less than it looks: the commission only ever triggers when someone actually pays, so a generous window mostly changes who gets credit, not how much you spend.
Who it matters to is content affiliates. Someone reads a comparison post, thinks about it for three weeks, then buys. A 7-day window pays that writer nothing and they notice. If your program leans on bloggers, reviewers and newsletter writers, 60 to 90 days is the honest setting. If your product is an impulse purchase, shorter is fine and nobody will complain.
Pay on money, not on signups
If you have a free trial, pay the commission when the trial converts, not when it starts. Paying per started trial sounds motivating and creates an incentive for exactly the traffic you don't want: bulk, low-intent, occasionally fake. Real money changing hands is the one event nobody can game cheaply, which is why it's the event the commission should follow. (Module 7 covers what happens when people try anyway.)
Write down why
Last step, five minutes, weirdly valuable. Open a doc and record six lines: the rate, the model (recurring, one-time, duration), the cookie window, whether trials pay, the date, and two sentences of reasoning with the numbers you used. That's your commission decision record.
Eight months from now a big partner will ask for 30%, or you'll wonder whether you're overpaying, and instead of vibes you'll have the arithmetic you did today. Change the number if the numbers changed. That's the whole discipline.
Next: the commission gets people in the door, and module 2 is the six rules you write before anyone walks through it.