Incrementality or Illusion

Most affiliate programmes report impressive revenue numbers. Far fewer can demonstrate that the revenue is actually new. The difference is incrementality testing.

Affiliate attribution has a structural problem that the industry has been comfortable ignoring for most of its existence. Last-click attribution — still the dominant model in most networks — assigns full revenue credit to the last affiliate touchpoint before a conversion. In practice, this means that a user who had already decided to purchase and searched for a discount code in the final moment before checkout generates a commission for a coupon publisher who contributed nothing to the sale.

What cannibalisation looks like

Cannibalisation in an affiliate programme is rarely visible in standard reporting. Your affiliate dashboard shows strong click volumes, healthy conversion rates, and growing revenue. What it does not show is how much of that revenue would have converted anyway through direct, branded search, or email — without the affiliate touchpoint. When the affiliate commission is paid on revenue that would have been earned regardless, the programme is not generating returns; it is generating costs.

The holdout methodology

Incrementality testing requires temporarily withholding affiliate exposure from a controlled segment of your audience and comparing conversion rates between the exposed and unexposed groups. The difference is the true incremental lift the programme is generating. This methodology is more operationally complex than standard attribution, but it is the only way to determine whether your programme is creating value or redistributing it from other channels.

Structuring for incrementality from the outset

The cleanest solution is to design the programme architecture around incrementality from the beginning: prioritise publishers that reach genuinely new audiences rather than retargeting your existing customer base; structure commission tiers to reward verified new-customer acquisition rather than revenue volume; and run ongoing holdout tests to monitor programme health over time. A smaller, well-structured programme will outperform a large, poorly-structured one in almost every scenario.

The affiliate programmes that survive scrutiny — from CFOs, from boards, and from increasingly sophisticated attribution tools — are the ones built on verifiable incrementality. They are also the programmes that attract and retain the highest-quality publishers, because they pay on results that both parties can trust.