Last-click attribution is a rule that gives 100% of the credit for a conversion to the final click a customer made before converting. Any earlier ads, emails or searches that introduced them to you get nothing, however much they shaped the decision.
How last-click attribution works
Take a homeowner in Bristol planning a loft conversion. They see your video ad on YouTube, later click a Facebook ad and browse the gallery, then find one of your guides through Google search. Two weeks on, they search your company name, click your brand search ad and send an enquiry. Under last click, the brand search ad receives the whole enquiry; the video, the Facebook ad and the guide receive nothing.
What counts as “the last click” depends on which tool is doing the counting, because each one only sees its own part of the journey:
- Google Ads credits the last Google Ads click it knows about. It cannot see the Facebook ad at all.
- Meta Ads credits its own ads within the attribution window you set, including people who only viewed an ad.
- GA4 offers a last-click model across every channel it records. It skips direct visits where an earlier source exists, which makes it a last non-direct click model in practice.
At the time of writing (October 2026), last click and data-driven are the only two models left in Google Ads and GA4. Google retired first click, linear, time decay and position-based from both products during 2023.
Why it matters
Last click is simple, stable and easy to explain, which is why it was the default in analytics tools for so long. Many UK small businesses have run on it for years without ever choosing it.
Its weakness is that it rewards whatever sits at the end of the journey. Brand search, remarketing and voucher code sites look outstanding, because they catch people who had already decided. Video, social prospecting, PR and content look weak, because they create demand that something else collects. Move budget on last-click evidence alone and you can starve the channels that fill the pipeline, then wonder months later why brand searches have fallen.
It also explains why your platforms disagree. Google Ads and Meta can each claim the same sale as their own last click, because neither can see the other. Added together, their figures usually exceed the orders you actually took.
Common mistakes
- Cutting prospecting campaigns because last click shows them producing few conversions.
- Treating a high return on brand search as proof of new demand, when many of those buyers would have clicked your organic listing anyway.
- Paying affiliate or voucher sites commission on last click for customers who were already at the checkout.
- Adding up the conversions each ad platform reports and calling the total your sales.
- Switching a conversion action from last click to data-driven and reading the following month’s shift as a change in performance.
How to act on it
Keep last click as one lens rather than the only one. It is good at telling you which ads close sales. For which channels start journeys, look at the conversion paths and model comparison reports in GA4’s Advertising section and set them against data-driven attribution. A channel that scores well there but poorly under last click deserves a test before a cut.
Give channels different jobs and different measures. Prospecting can be judged on new visitors who later convert, closing channels on cost per sale. For the most expensive questions, such as how much brand search really adds, run an incrementality test rather than trusting any attribution rule.
When I take on PPC management for an account, one of the first things I check is which model each conversion action uses, because the bidding follows whatever the model says. The entry on attribution models compares the alternatives.
