The question every beginner inherits: their platform credits 100% of a sale to the final click before purchase. That's last-click attribution — simple, deterministic, and the default in most ad accounts. So why the constant warnings?
What the evidence says: controlled studies repeatedly show last-click systematically over-credits channels that sit at the bottom of the funnel — branded search, retargeting, coupon and affiliate touches — because those are simply closest to the click. A frequently cited pattern from incrementality experiments (where one group is exposed and a holdout is not) is that branded-search "conversions" barely move when paused: the customer was already going to buy. Last-click can't see that.
The nuance: last-click isn't wrong, it's narrow. For a pure direct-response funnel with one touch, it's basically accurate. The bias compounds as your paths get longer and your upper-funnel investment grows. A 1-touch business and a 7-touch business should not trust the same model.
What to actually do: don't rip it out — benchmark against it. Run the same period through a multi-touch model (credit spread across touches) and note which channels gain or lose. Channels that lose a lot under multi-touch are your last-click over-credit suspects. Then validate the biggest movers with a holdout test before reallocating.
Bottom line for practitioners: last-click is a fine scoreboard and a terrible strategist. Keep it for reporting consistency; never let it set budget alone.
Credit Where Due
@CreditWhereDue
What exactly is "last-click" and why does everyone warn against it?
This post was published in the Telegram channel Credit Where Due (@CreditWhereDue). All its posts are on the channel page.
Open on Telegram