Last-click attribution

Last-click attribution means the cookie present at the moment of purchase decides who gets credit. If a visitor clicks your affiliate link, browses, then clicks a second affiliate's link before buying, the second affiliate gets credit. Your cookie was overwritten. Your introduction does not count. This is the industry-standard model and the source of one of affiliate marketing's biggest structural inequities.

The cookie-overwrite mechanic. When a network sets an attribution cookie, it stores the publisher ID. A second click from a different publisher's link writes a new value into the same cookie slot. The network honours whichever value is there at conversion time. Most networks do not track the click history; they just read the current cookie value. The visitor often does not realise they triggered the overwrite; they may click a coupon code popup at checkout without knowing it is an affiliate link.

Why last-click dominates. Three reasons. Simplicity: there is one cookie to read and no disputes. Lower fraud surface: tracking multi-touch flows opens up gaming opportunities. Network preference: last-click rewards the publisher who delivered the final-stage traffic, which networks consider higher-intent. The model is now so embedded that even networks that technically support other attribution models default to last-click in practice.

Coupon sites and browser extensions: the attribution-hijacker class. This is the practical pain point for content creators. Coupon and deals sites (RetailMeNot, Honey, CouponFollow, Slickdeals) operate at the bottom of the funnel: visitors who already decided to buy go looking for a discount code. Browser extensions (Honey before its 2024 controversy, Rakuten, Capital One Shopping, Karma, PayPal Honey successor) automatically detect checkout pages and silently inject their own affiliate links. The visitor sees a "we found you a coupon" popup; clicks; the extension sets its cookie; your cookie is overwritten; the conversion goes to the extension.

What creators lose. Estimates vary, but research-content creators (review videos, in-depth blog posts) commonly lose 10 to 30 percent of attributable conversions to last-click hijacking by browser extensions and coupon sites. The buyer made the purchase decision based on your content, but the extension captured the credit. Networks know this happens and continue to pay it out because the alternative is harder to administer.

Other attribution models that exist (but are rare in affiliate). First-click attribution credits whoever introduced the visitor first; some direct-brand programmes offer it as a premium for top publishers. Multi-touch attribution splits credit across multiple touchpoints; common in paid media analytics, almost never in affiliate. View-through attribution credits an impression without a click; used in display advertising. Custom models (J-shaped, U-shaped, time-decay) exist in marketing-mix modelling but not in production affiliate networks.

Mitigations. Direct-brand programmes increasingly offer first-click or "discount-stacking" rules that protect the originating publisher even if the buyer uses a coupon code at checkout. SaaS affiliate programmes with lifetime attribution sidestep the problem entirely because the recurring revenue belongs to whichever publisher acquired the customer, with no last-click override. For Amazon-style last-click niches, the practical defence is content positioning: get the click at the moment of decision, not weeks earlier.

The economics. Last-click attribution structurally favours bottom-of-funnel placements (coupon sites, browser extensions, checkout-page intercepts) over top-of-funnel content (reviews, education, comparison content). This is the central tension between content creators and the affiliate networks that pay them: the model rewards proximity to the transaction, not contribution to the decision.