An attribution model is the rule that determines how much credit a conversion gives to which ad interaction. If a user encountered multiple ads before making a purchase, this is where the question of whether that credit goes to a single click or is spread across the whole path gets answered.
Choosing a model changes the report, not the sale that actually happened. The same conversion data can assign different shares to different channels depending on which model reads it. For example, a user first sees a brand awareness ad, then three days later searches for “[brand name] + product” and clicks a search ad and buys; under the last-click model, all the credit goes to the search ad, even though part of the purchase decision may have come from that first touchpoint.
Google Ads’ attribution models documentation states that first-click, linear, time-decay, and position-based models are no longer supported, and that conversion actions using those models were automatically upgraded to data-driven attribution. Currently, only two models are active: last-click and data-driven.
What Is an Attribution Model?
This model is the rule that defines how credit gets split when there’s more than one ad interaction on the path to a conversion.
A model determines an interpretation, not a fact
There’s no way to know for certain which ad actually influenced the user. A model offers a reasonable estimate and allocation rule based on the interaction data available — which is why you shouldn’t be looking for the “correct model,” but the model that fits the situation.
The Last-Click Model
This model gives all the credit for a conversion to the last-clicked ad and keyword. Touches that came before it remain invisible in the report.
Paints a simple but incomplete picture
Last-click is easy to understand. But it hides the contribution of ads at the brand-awareness or consideration stage — a contribution that’s real even if the report never shows it. For example, if a Display campaign never gets clicked but generates impressions, and the users who saw those impressions search for the brand name and buy within two weeks, the last-click report shows the Display campaign as “0 conversions,” even though it may have driven part of that search volume. Looking at this model alone and shutting down upper-funnel campaigns as a result can be misleading.
Data-Driven Attribution
This model is now the default for most conversion actions. Credit is distributed across multiple touchpoints based on the account’s historical data. See What Is Data-Driven Attribution? for more detail.
Older models were automatically upgraded
Conversion actions using first-click, linear, time-decay, and position-based models were automatically migrated by Google to data-driven attribution. This is no longer a manually selectable option — the only thing an account owner can do now is correctly read the distribution that data-driven attribution produces.
What Does the Choice of Model Change?
A model change can directly affect channel-level budget decisions.
Upper-funnel channels can look different
Brand-awareness or consideration-stage campaigns may look worthless under the last-click model, while getting a share closer to their real contribution under the data-driven model. A brand’s Display and YouTube campaigns might take up 20 percent of total budget in the last-click report while showing almost no conversions; reading the same data under the data-driven model can reveal that these campaigns were actually triggering a portion of search conversions.
Compare reports from the same period
Directly comparing two reports produced under different models can be misleading. If you’re going to compare, use the same model for the same time period.
Pre-Launch Checklist
When evaluating an attribution model:
- Which model are the account’s conversion actions currently reporting under?
- Are channel-level budget decisions relying on a single model’s report?
- Is the reporting difference after a model change being confused with an actual performance change?
A Common Mistake
The most common mistake is looking only at the last-click report and shutting down upper-funnel campaigns as “not delivering.” Those campaigns’ contribution may simply fall outside the window that model can see.
The second mistake is assuming data-driven attribution works with the same reliability in every account. In low-conversion-volume accounts, the model has to work with limited historical data, and the distribution can shift over time — this is covered further in What Is Data-Driven Attribution?.
Summary
An attribution model determines how conversion credit is split among ad interactions. Google Ads now only supports the last-click and data-driven models; older models were automatically upgraded to data-driven. Choosing a model doesn’t change the actual sale — it changes which channel that sale gets credited to, and upper-funnel campaigns are where that difference shows up most visibly.