PPC glossary

Attribution model

An attribution model is the rule that decides how credit for a conversion is shared between the ads, clicks and channels a customer interacted with before converting. Google Ads and GA4 now use two approaches: data-driven attribution, which shares credit using machine learning, and last click, which gives all of it to the final interaction.

Reviewed by Gareth Hoyle, Managing Director
Last reviewed 31 October 2026

Which models do Google Ads and GA4 offer?

Google Ads offers data-driven attribution, which Google says is the default for most conversion actions, and last click. It also accepts external attribution imported from another tool. GA4 offers data-driven attribution, paid and organic last click, and Google paid channels last click. In GA4, direct visits receive no credit unless the whole path is direct.

How does data-driven attribution work?

It uses the account's own data to estimate how much each interaction contributed. Machine learning compares converting and non-converting paths to see what happened against what could have happened, taking in factors such as time before the conversion, device, the number of ad interactions, the order of exposure and the type of creative. Each model is specific to one advertiser and one conversion.

Which models did Google retire?

On 6 April 2023 Google announced it was removing four rule-based models from Google Ads and GA4. From June 2023 they could no longer be chosen for new conversion actions in Google Ads, and in September 2023 the remaining ones were switched to data-driven attribution and dropped from reporting, including the model comparison report.

First click: all credit to the first interaction
Linear: equal credit to every interaction
Time decay: more credit to interactions closer to the conversion
Position-based: typically 40% each to the first and last interactions, with 20% shared across the middle

What attribution cannot tell you

Every model shares out credit for conversions that were tracked. None says whether the sale would have happened without the ads, which is the question an incrementality test answers.

Example

A customer clicks a Search ad for 'running shoes for flat feet', returns a week later from a remarketing ad, then searches the brand name and buys for £120. Last click gives the full £120 to the brand keyword, which makes generic search look unprofitable. Data-driven attribution shares the value across all three clicks, based on how often similar paths convert without each step.