Recurring Commission Affiliate Programs: Judge Churn, Not Rate

A 30% recurring commission sounds better than 20% until your referrals cancel in month two. That is the most common mistake I see affiliates make when they compare recurring commission affiliate programs. They rank by headline percentage, when the number that decides what they actually earn is how long the referred customer keeps paying. Churn matters more than the rate. Cookie windows, hold periods and clawback rules all matter too, but they come second.

Here is how to judge a program before you spend a newsletter issue, a video or a comparison post on it.

The 12-month math: 10 customers, three programs

Say you refer 10 customers to a $49/month product in January. Compare three offers over the next 12 months:

• Program A: 30% recurring on a product that is easy to drop. Eight of your ten customers cancel after two payments. Two stay all year.
• Program B: 20% recurring on a sticky product. Two customers cancel after three payments. Eight stay all year.
• Program C: a one-time $100 bounty per paying customer, nothing after.

The working for A: 8 × 2 × $14.70 = $235.20, plus 2 × 12 × $14.70 = $352.80. For B: 2 × 3 × $9.80 = $58.80, plus 8 × 12 × $9.80 = $940.80.

The "worse" rate pays 70% more in year one. B is also still paying $78.40 a month going into year two, while A has shrunk to $29.40.

Program C ties B in year one, then stops. That is a fine trade if you need cash this quarter. It is a bad one if you expect to still be publishing in three years, because B's month-13 income compounds with every new referral you add.

The case for chasing the headline rate, and why it loses

The opposing view deserves a fair hearing. The commission rate is the only number printed on the signup page. It is in writing, it is comparable across programs, and churn is invisible until you have already promoted. A generous rate can also signal that a vendor takes affiliates seriously. And if your audience is high intent, your referrals might stick better than the vendor's average customer.

All true. But churn is undisclosed, not unknowable, and the rate is the lever vendors reach for when retention is weak. In my judgment, a product that loses most customers within 60 days can comfortably advertise 40% recurring because it rarely pays it for long. A vendor with strong retention offering 30% is taking on a much larger long-term liability, which is why sticky products often sit at 20 to 25%. A high rate on a cheap, single-purpose tool is sometimes a tell rather than a gift. (If you want the vendor-side reasoning, What Is a Good Affiliate Commission Rate for SaaS? Do the Math walks through it.)

Run the numbers: 30% on a two-month customer is 0.6 months of revenue. 20% on a 12-month customer is 2.4 months. The rate would need to be four times higher to close that gap.

How to estimate churn before you have data

You rarely get a retention chart, but you can get close:

• Ask the program manager directly. "Of customers referred by affiliates last year, what share were still paying at month 6?" A well-run program knows. A dodge is a yellow flag.
• Look at what the product holds. My rule of thumb: tools that store the customer's data or sit in a daily workflow (accounting, email lists, help desks, CRMs) stick. One-job utilities, like a resume builder or a one-off file converter, get cancelled when the job is done.
• Check the pricing page for annual billing. If annual plans are prominent and discounted, more of your referrals will pay upfront and fewer will vanish in month two.
• Search 1- and 2-star reviews on G2 or Capterra for "cancel" and "billing". Complaints about surprise renewals usually mean customers want out.
• Run a cheap test. Promote once in a minor placement, then check your dashboard at day 90. If four of six referred customers have stopped paying, you learned that for the price of one mention.

The plumbing: cookie window, hold period, payout terms

Once churn looks acceptable, check whether the program will actually credit and pay you.

Cookie window. Newsletter readers click, close the tab and come back 10 to 20 days later. A 7-day cookie loses most of them. I treat 30 days as the floor for content affiliates, and 60 to 90 for B2B products with long evaluations. Also ask whether attribution is first click or last click; under last click, a coupon site can take a sale you originated. More detail in Affiliate Cookie Duration: How Long Should Your Window Be?.

Hold period. A 30-day hold plus a monthly payout run means a commission earned on March 3 might reach you in early May. That is normal. A 90-day hold is long but defensible on annual plans with refund windows.

Payout threshold. Divide it by your per-payment commission. At $9.80 a month, one referral needs six payments ($58.80) to clear a $50 threshold. A $100 threshold on a low-priced product means small affiliates wait most of a year.

Payout methods. Confirm they pay through something you can receive, such as PayPal, Wise or Paystack, and on a stated schedule rather than "when we get to it".

Refund clawbacks and server-side tracking

Recurring affiliate programs pay over many months, so refunds and disputes touch more of your income than they would with a one-time bounty. Get answers to these in writing:

1. Does a refund cancel only the commission on the refunded payment, or every commission tied to that customer?
2. Does a partial refund reduce the commission proportionally, or wipe it out?
3. If the refund happens after you were paid, is it netted against your next payout or invoiced to you?
4. Are chargebacks reversed, and restored if the vendor wins the dispute?

Then ask how sales are recorded. If a program depends on a browser pixel on the thank-you page, ad blockers and closed tabs cost you commissions, and renewals in month seven fire no pixel at all because nobody visits a page. Commissions should come from the billing system itself, as explained in Stripe Affiliate Tracking: Attribute Subscriptions Without a Pixel. Server-side tracking also prevents fake browser-reported conversions, one of several fraud controls covered in Affiliate Fraud Detection for SaaS: Six Patterns, Six Controls.

As one example of what good looks like, Anderro only creates commissions from server-side sources such as Stripe charge and invoice.paid events or HMAC-signed Paddle Billing webhooks; browser-reported payments never create one, and a refund after payout becomes a negative adjustment against the affiliate's next payout rather than a bill.

A scorecard before you write a word

When I review saas affiliate programs for my own audience, I score them in this order, and the first item carries the most weight:

1. Expected customer lifetime. Month-6 retention if they will share it, product-category signals if not.
2. Per-payment commission. Price × rate. 20% of $99 beats 30% of $19.
3. Recurring term. Lifetime vs capped. At $9.80 a month, a 12-month cap pays $117.60; if customers stay 30 months, lifetime pays $294.
4. Server-side payment tracking, especially for renewals.
5. Clawback rules that only reverse the refunded payment.
6. Cookie window of at least 30 days.
7. Time to first dollar. Hold period plus threshold plus schedule should land your first payout within about three months of one referral.

The best recurring commission affiliate programs answer questions 1 and 5 within a day. So before your next promotion, send the program manager one email asking two things: what share of affiliate-referred customers are still paying at month 6, and what happens to already-paid commissions when a customer refunds. If you get silence, the headline rate tells you nothing worth trusting.