Affiliate ROI Calculator

The arithmetic

Programme return is (incremental contribution − programme cost) ÷ programme cost, where programme cost is commissions plus network fees plus management time.

The incremental qualifier is the whole calculation. Contribution from orders that would have happened anyway is not a return on the programme.

Return = ((incremental orders × contribution) − total cost) ÷ total cost.

Numbers on it

1,400 monthly orders attributed to affiliates, $8,428 of commission and network fees, plus $1,500 of management time, $9,928 of cost.

Taking all 1,400 as incremental: $44,660 of contribution, a 350% return.

Now the honest version. If 40% would have converted anyway: typical where coupon and loyalty affiliates are a large share: incremental orders are 840, contribution is $26,796, and the return falls to 170%.

Still a good channel, and it is half of what the attributed figure claims. The 560 non-incremental orders cost $3,371 in commission for revenue the business already had.

What it does not tell you

Affiliate tracking uses last-click attribution almost universally, which credits whoever the customer touched most recently. Coupon sites intercepting a customer already at checkout are structurally advantaged by this.

Attribution also runs across channels: an affiliate credited with an order that started from a paid search click means the same order is being paid for twice.

What follows from it

Segment the programme by affiliate type: content, review, coupon, loyalty, email, and measure incrementality separately for each. The variation between them is enormous.

Then run a holdout on the coupon segment specifically, by disabling their tracking for a fortnight and measuring what happens to total orders. It is uncomfortable and it is the only way to know.

Why coupon affiliates are the recurring argument

A customer at checkout who opens a browser extension or searches for a discount code, finds one, and completes the purchase has been converted by nobody. The affiliate is credited and paid.

That is not fraud, the tracking worked as designed, but it does mean a share of the programme's cost buys nothing. Estimates of the non-incremental share in coupon-heavy programmes commonly run above half.

The usual responses are lower rates for coupon partners, excluding them from last-click credit, or removing the code field from checkout so the search never starts. Each is a legitimate choice and none of them is possible until the incrementality has been measured.

Recruitment cost belongs in the programme cost alongside commission, since finding and onboarding partners takes real time and the good ones require pursuing rather than applying.

For a programme with a small number of productive partners, that acquisition effort can exceed the network fees.

Comparing the programme against the channel it most resembles, paid search on non-brand terms; is more informative than comparing it against the business average.

Both are acquiring at the point of intent, and the cost comparison between them is the one that determines where marginal effort should go.

Measuring the programme excluding the top partner occasionally is a useful stress test, since many programmes turn out to be one relationship with a long tail attached.

Attributing the management time honestly matters for small programmes, where a part-time manager can cost more than the commissions paid.