Someone offers you 40% commission. Is that good?
You can't tell. Not from that number. Depending on four things the offer usually doesn't mention, "40%" can be worth several times more or several times less than "20%".
This is how to work it out yourself, with the arithmetic written out. At the end we run our own programme through the same checklist, including the parts where it doesn't come off well.
1. Recurring or one-off — this is the big one
Two offers on the same $14/month subscription:
- A: 50% of the first payment. Once.
- B: 20% of every payment, for as long as the customer stays.
A sounds two and a half times better. In money:
- A pays you $7.00, one time.
- B pays you $2.80 a month, every month.
They break even at exactly 2.5 months. From the third month onward, B is ahead and never stops being ahead.
| Cumulative earnings from one referral | A (50% once) | B (20% recurring) |
|---|---|---|
| Month 1 | $7.00 | $2.80 |
| Month 3 | $7.00 | $8.40 |
| Month 12 | $7.00 | $33.60 |
| Month 16 | $7.00 | $44.80 |
Published churn benchmarks put normal consumer software at 6–8% monthly churn, which works out at roughly 12 to 17 months of average subscription length. Against that, the recurring offer is worth about five to six times the one-off — despite having a headline number less than half the size.
When the one-off actually wins: if the people you send don't stay. If your audience buys on impulse and cancels in a month, take the money up front. Recurring commission is a bet that your recommendation was a good one, and it only pays if it was.
2. A percentage of what, exactly?
Two programmes both advertise 40%. On the same $14 plan:
- 40% of the list price — $5.60.
- 40% of what the customer actually pays after a 20% discount — $4.48.
Same headline. 20% less money. A third variant pays on the amount left after payment processing fees, which takes a few percent more off the top.
So the question is: 40% of which number, exactly? A programme that answers that immediately and precisely is telling you something about how carefully it is run.
The fair counterweight: when a programme pays on the discounted amount, it is usually because the customer is getting a discount — the programme is paying out twice from the same pot. A lower base isn't automatically meanness. Which leads directly to the next question.
3. What does your audience get?
There are two kinds of affiliate code.
One gives your viewer nothing. You are asking them to do you a favour.
The other saves them money. You are giving them something.
The second gets typed. The first gets ignored — not because your audience doesn't like you, but because there is no reason to go and find it at checkout.
This matters more than it looks, because it cuts against the headline comparison: a programme that gives the customer nothing can afford to pay you a higher percentage. Compare only on commission rate and you will systematically pick the code nobody uses.
4. How do you actually get credited?
Two mechanisms, and for video creators they are not close.
A link with a cookie window. Your viewer clicks, a cookie is set, and you get credit if they buy within a set number of days. Thirty days is the industry standard; real programmes run anywhere from 24 hours to 90 days.
A code they type at checkout. No cookie, no window, no expiry.
Video is a long-tail medium. A video keeps collecting views for months or years, and a 30-day cookie set by a click in March pays nothing on a purchase in June. Cookies also get cleared and blocked, and they break the moment someone clicks on their phone and buys on a laptop.
The honest counterweight: a link is automatic and a code has to be remembered. A link converts better in the moment. A code converts across the tail — and for evergreen content, the tail is most of your audience.
If a programme uses links, ask for the window in days. If it's 24 hours, that number is the offer, not the percentage.
5. How long does the average customer stay?
This is the number that turns a percentage into an actual figure, and almost nobody publishes it.
Ask directly: what is your average subscription length? A programme that knows will tell you. A programme that won't either doesn't know or doesn't like the answer, and both are worth knowing before you put it in a video.
Rough anchors for consumer software: 6–8% monthly churn is described as normal, giving roughly 12–17 months of average lifetime. Better-retained consumer products run 4–5%, or 20–25 months.
Then do the multiplication. Monthly commission × expected months = what one referral is really worth. That figure is what you compare between programmes. The percentage on its own is not comparable to anything.
6. The two questions people forget
When do you get paid, and is there a minimum? A 40% rate with a $100 payout threshold pays you nothing at all until you've accumulated $100. For a smaller channel that can take a year, and some programmes expire the balance if you don't reach it.
What happens if you stop promoting? In some programmes the referrals you've already made keep paying for as long as they stay subscribed. In others, your commission stops when your promotion stops. That is the difference between building an asset and doing a job.
The seven questions
- Recurring or one-off? If one-off, how many months of the recurring alternative does it equal?
- A percentage of what — list price, what the customer pays, or what's left after fees?
- Does my audience get anything out of it?
- A link with a cookie window, or a code with no expiry? If a link, how many days?
- How long does the average customer stay?
- When am I paid, and is there a minimum threshold?
- If I stop promoting, do existing referrals keep paying?
Answer those seven and you can compare any two offers properly. Skip them and you are comparing font sizes.
Our own programme, against the same seven questions
We would rather you held us to this than took our word for it.
1. Recurring, on subscription payments. Every month the customer stays, with no end date. One-off credit top-ups are outside it — they don't carry your viewer's discount and they don't pay you commission. If we change that, we'll change this line.
2. A percentage of what the customer actually pays — the lower of the two bases, after their discount. On the $14 Pro plan discounted to $11.20, the entry rate of 20% is $2.24 a month, not $2.80. We'd rather write that here than let you find it on your first payout.
3. Your audience gets 20% off, permanently. Not a first month, not a trial, not a one-off code. For as long as they stay subscribed.
4. A code, typed at checkout. No cookie, no window, nothing to expire. A code you put in a description today still works on someone who finds the video in two years.
5. We don't know yet, and we're not going to invent a number. The programme is new and so are we. When we have a reliable average subscription length we'll publish it the same way we published all fifty rows of our thumbnail CTR testing, including the ones that made us look bad.
6. Monthly, manually, by PayPal or Wise, once your balance passes $10. Anything under $10 rolls into the next month and never expires — we're not using a threshold to avoid paying small affiliates, we're using it to avoid sending you a transfer that the transfer fee would eat. There is no partner dashboard and no automated payment rail — a person does it. That means you get a real reply from a real person, and it also means this isn't built for anyone who wants live analytics. If that's a dealbreaker, better to know now than after you've made a video.
7. Existing referrals keep paying for as long as they stay subscribed, whether or not you ever mention us again.
The rates are 20% from your first active referral, 30% from ten, and 40% from a hundred — on what the customer pays. For context, published guidance puts SaaS affiliate rates in a 5–30% band overall, with mature programmes clustering at 15–25% and creator-economy tools at 12–22%. The entry tier is ordinary. The top tier is not.
And the arithmetic from our side, since we've asked you to do yours: at 40%, a customer paying $11.20 leaves us $6.72 out of a $14 list price. We keep 48% of the sticker. That is the actual reason the top tier sits at a hundred active referrals rather than ten — not a negotiating position.
Where these numbers come from
Three figures in this article aren't ours. They come from published affiliate-industry and SaaS benchmark guidance:
- SaaS affiliate commission ranges: Post Affiliate Pro
- Cookie durations and the 30-day standard: Trackdesk
- Consumer software churn of 6–8% a month: Churnkey
They're widely reported, and we've used them as rough anchors rather than precise facts — the arithmetic in this article doesn't depend on any of them being exactly right.
Everything about our own programme comes from our own pricing code, and every figure in it can be checked against the prices on our pricing page.
If you want the code
You generate it yourself on the page, it works immediately, and there is no application to write and no dashboard to learn.
And if you'd rather judge how we handle inconvenient numbers before deciding whether to point your audience at us, the better read is the 50 thumbnail tests we published in full — every row, including the one that only moved 8%.