An affiliate programme for a digital product
It is the cheapest distribution there is (you pay only on a sale), and it is the one channel that punishes an unproven product hardest.
The short version
- 30–50% is the working range. Under 20% is ignored; 50% is normal for a product with no unit cost.
- It does nothing for an unproven product. Affiliates send traffic once and leave if it does not convert.
- Ten good partners beat two hundred signups. Recruit by hand, one at a time.
- Your buyers are the best affiliates. They already know it works and can say so honestly.
- Gumroad and Payhip run it for you. Do not build tracking; it is a solved problem.
What an affiliate programme actually is
A standing offer: anyone who sends you a buyer keeps a share of the price. For a digital product that share is normally 30–50%, because there is no unit cost and the commission is simply what you are willing to pay to acquire a customer. You pay nothing until a sale happens, which is why it sounds free and why it is so often disappointing.
The sounds-free part is real. The disappointing part is that an affiliate programme is a multiplier, and multiplying nothing gives nothing.
Why most of them produce nothing
The common sequence: launch a product, get a few sales, open an affiliate programme hoping it will do the distribution the product has not earned. Thirty people sign up. Two send traffic. Nobody buys.
What went wrong is nearly always one of three things.
| Symptom | Cause | Fix |
|---|---|---|
| Signups, no traffic | They joined and forgot | Recruit fewer, individually |
| Traffic, no sales | The page does not convert | Fix the page first |
| One burst, then silence | It converted badly once | Same: the page |
| Nobody signs up | Commission too low, or no audience overlap | Raise it; change who you ask |
The second and third rows are the same problem and the most important thing on this page: an affiliate gives you one attempt. They mention you once to an audience they spent years building. If the conversion is poor, they do not debug your funnel. They move on, permanently.
So the order is: prove the product converts on your own traffic, then recruit. How to write a sales page for a digital product is the work that has to come first.
What commission to offer
For a digital product with no unit cost, the honest frame is: what is a customer worth, and what would you pay an ad platform for one?
- Under 20%: ignored. Not worth an email to a list somebody spent years building.
- 30%: the floor that gets attention.
- 40%: the common landing point, and a fair one.
- 50%: normal for digital, and correct if repeat purchases or a list signup follow the first sale.
- Over 50%: for a launch window or a genuinely large partner, not as a standing rate.
The instinct to protect margin is the wrong instinct here, because the alternative to paying 40% is usually not keeping 100%. It is not making the sale. What the number has to respect is your pricing: a product priced at the low end leaves a commission too small to motivate anyone, which is one of several reasons cheap products are harder to sell, not easier.
Who to actually recruit
Ten partners who send real traffic beat two hundred signups. That means recruiting by hand.
- Your buyers. The best affiliates on the list. They paid, it worked, and they can say so without pretending. Email everyone who bought and offer it directly.
- People who teach the adjacent thing. Someone who teaches the skill your product operationalises has exactly your audience and is not a competitor.
- Newsletter writers in the niche. Small lists convert far better than large social accounts, because the reader chose to receive it.
- People who already mention you. Search your product name. Somebody is already recommending it for nothing.
Where not to bother: affiliate directories, "join my programme" posts to your own followers, and anyone with a large audience and no relationship to the subject. All three produce signups and no sales, which is the metric that flatters and does nothing.
The message should be specific: what the product is, who it is for, the rate, and why you are asking them. A generic invitation gets a generic response, which is none. This is the same recruiting problem as getting your first SaaS customers, and it is solved the same way: individually, at first.
Give them something to work with
The single biggest predictor of whether an affiliate sends anything is whether promoting you is easy on the day they think of it.
- A link that works, from the platform, tracked automatically.
- Three or four pre-written posts they can edit. Most will use them nearly as-is.
- Images at the right sizes, so they do not have to make any.
- One honest paragraph on who the product is not for. Affiliates who set expectations get fewer refunds, and the good ones will ask.
- A copy of the product. Give it to them. Somebody promoting a thing they have not seen writes vague copy and knows it.
Assets are also the difference between a partner who promotes once and one who keeps you in rotation, because the second mention costs them nothing to make.
The plumbing
Do not build this. Gumroad, Payhip and Lemon Squeezy all include affiliate tracking: you set a rate, they generate links, attribute sales and split the payout automatically. Where to sell digital products covers choosing between them, and affiliate support is a reason to stay on a hosted storefront rather than roll your own checkout.
Two settings worth deciding deliberately:
- Cookie window. Thirty days is standard. Longer is friendlier to affiliates and costs you nothing real.
- Whether affiliates apply or are invited. Invite-only, for everything said above.
What it is worth
Set the expectation properly: for most small digital products, affiliates are a modest and welcome share of sales rather than a channel that changes the business. The ones where it does change the business have two things: a product that converts well on cold traffic, and a price high enough that 40% is worth somebody's morning.
It is also the only distribution channel that costs nothing to keep open. Once it is set up, an unused affiliate programme costs zero, and occasionally someone sends a month's revenue because it was there when they thought of you. That asymmetry is the real argument for having one (after, and only after, the launch has shown the thing sells).
Frequently asked questions
What commission should I pay affiliates for a digital product?
Between 30% and 50%. A digital product has no unit cost, so the commission is pure customer acquisition, and 40% is the number most people settle on. Below 20% is not worth an affiliate's audience.
When should I start an affiliate programme?
After the product converts on your own traffic, not before. Affiliates send a burst once; if it does not convert, they stop and they rarely come back for a second attempt.
Where do I find affiliates?
Your own buyers first, then people who already teach the thing your product solves. Recruit individually with a specific message. Public affiliate directories produce volume and almost no sales.
Do I need special software?
No. Gumroad, Payhip and Lemon Squeezy include affiliate tracking and pay out automatically. Building your own link tracking is a way of avoiding the harder job of recruiting.
Is an affiliate programme worth it for a small product?
It can be, and it is worth less than people expect. It is a multiplier on a product that already sells, not a way of starting one.
The system behind this, written down
Everything above is the map. The Income Loop is the work inside it: modules 0–6 from the problem you solve to the offer that pays for it, plus ten traffic paths: the deeper post banks, the content sales systems and the full software build sequence, in one place.
Get The Income LoopNot ready to pay for anything? The Basic Income Loop is free and includes a complete seven-day starter for Threads, Instagram or software, enough to find out which one suits you before spending anything.