If the only source of truth you have for which channel “works” is your analytics platform’s last-click report, you’re probably making budget decisions based on an incomplete picture.

A note on wording: last-touch and last-click attribution mean the same thing in everyday use — all the credit goes to the final interaction before the conversion. Google Analytics calls it last-click, most media teams say last-touch. Everything below applies to both.

What last-touch attribution gets wrong

The last-touch model gives 100% of the credit for a conversion to the last touchpoint before purchase. This systematically favors channels that capture already-existing demand (like brand Search) and penalizes the channels that generated that demand in the first place (like Social or Display), even when they were the real reason the user ended up buying.

The typical symptom

If you’ve ever seen a report where “Brand Search” looks like your star channel and decided to cut budget from discovery channels to reinvest there, that’s a classic symptom of looking at the problem through last-click: you’re rewarding the channel that closes the sale and penalizing the one that generated it.

More reasonable alternatives

You don’t need a perfect attribution model (there isn’t one). Anything better than last-click helps:

  • Linear attribution: splits credit equally across every touchpoint.
  • Position-based attribution: gives more weight to the first and last touch, splitting the rest among the middle ones.
  • Data-driven models: use machine learning on your own conversion data to assign credit based on each channel’s actually observed impact.

Why it survives anyway

Last-touch persists for three reasons worth naming, because they explain why “just switch models” rarely works as advice:

  • It is the default. Most reporting comes out of the box attributing this way, so it is what ends up in the deck.
  • It is easy to explain. “This click caused this sale” survives a meeting in a way that “Display contributed 18% of assisted value” does not.
  • It is deterministic. Nobody has to defend a model. That feels safer, right up until you cut the budget that was actually generating the demand.

What matters isn’t picking the “correct” model

No attribution model is 100% accurate, and obsessing over finding the perfect one is time poorly spent. What actually matters is no longer making budget decisions based exclusively on last-click, and complementing it with at least one model that recognizes the contribution of upper-funnel channels.

Common questions

Why does last-touch attribution not work? Because it assigns 100% of the credit to the final interaction, it overstates channels that capture existing demand — brand search, retargeting, direct — and understates the channels that created that demand. Optimizing to it tends to shrink the top of the funnel until volume falls.

Is last-touch ever the right choice? For short, single-session purchase journeys with one dominant channel, the distortion is small enough not to matter. The longer and more multi-channel the journey, the more it misleads.

What should I use instead? Anything that distributes credit: position-based as a readable starting point, data-driven if you have the conversion volume for it, and incrementality testing when a decision is expensive enough to justify holding out a group.

If your attribution today is last-touch only and you need a fuller picture, see how I work with brands and we’ll figure out which model makes sense for your business.