Last-touch attribution
Why it matters
That simplicity is the trap. Last-touch overpays whatever sits closest to the sale, such as retargeting ads, branded search and the final reminder email. It underpays whatever created the demand weeks earlier. A team that cuts a content programme because last-touch shows few conversions can find the retargeting collapsing soon after, because it only worked by catching people the content had already brought in.
How to apply it
- Trace a sample of recent wins through every step they took, not only the last one.
- Run first-touch attribution alongside, so the channel that started the relationship is also counted.
- Track assisted conversions, which show channels that helped without getting the final click.
- Move towards a model that spreads credit, such as linear or time-decay, once the gap between last-touch and reality is clear.
- Look at the attribution window, because it decides how far back a touch can count at all.
- Hold off cutting a channel on a weak last-touch number until its earlier role has been checked.
What it is
A person rarely buys after one visit. They may read a blog post, see an ad, open two emails and then search the brand name before buying. Last-touch attribution looks at that path and gives the whole credit to the final step, the branded search in this case. Everything earlier receives nothing. Many analytics tools use it by default because it is simple: one final click means nothing needs to be stitched together.
Common mistakes
- Treating the report as the truth instead of one view of it.
- Comparing channels that sit at different stages of the buying path on the same footing.
