Definition
What is data-driven attribution?
Data-driven attribution is the default model in Google Ads and Google Analytics 4 for distributing the credit for a conversion across the advertising interactions that preceded it. Rather than a fixed rule, it calculates a variable share for each touchpoint based on the account's history. The calculation differs from one advertiser to the next and is not published: it is more faithful in theory and unverifiable in practice.
The models that remain
In 2023, Google retired the first click, linear, time decay and position-based models. Conversion actions that used them were switched to data-driven attribution. Two options remain in Google Ads: this model or last click.
How the model distributes credit
The model compares the paths that lead to a conversion with the paths that do not. A touchpoint that appears more often in converting paths receives a larger share of the credit. Google does not publish the formula applied to your account: you can neither redo the calculation nor challenge it with figures.
What the model cannot see
The model only distributes the interactions it was able to observe. Word of mouth, a referral or a billboard do not appear in it. A sale closed by phone only counts if it is sent back to the platform. Finally, the model distributes credit without proving that the advertising caused the sale: that question belongs to incrementality.
A customer sees a video ad. Later she clicks on a generic search ad, then comes back through a search for the company's name before buying. Last click gives all the credit to the brand search. Data-driven attribution can split the credit across the three interactions, in shares that Google does not detail (fictional example).
We choose this model when several channels work together and the volume is sufficient; last click is enough for a short cycle on one or two channels. Either way, we treat the result as an estimate, not a measurement. We set the model before the evaluation period, note the date and then leave it alone: a report comparing two months measured with two models only measures the change of model.
To decide on a budget, we look first at the cost per customer acquired across all channels: total monthly marketing spend divided by the number of new customers, tracked over twelve months. That figure is calculated without attribution.
Not to be confused with
- Last click
- Last click gives all the credit to the final interaction before the conversion. It is simple and reproducible, but it is wrong in a predictable way.
- Multi-touch attribution
- Multi-touch attribution refers to any method that shares credit across several interactions. In Google's tools, data-driven attribution is the only form that remains.
- Advertising incrementality
- Attribution distributes credit across observed touchpoints. Incrementality measures the share of sales that would not have happened without advertising, by comparing an exposed group with an unexposed one.
- Marketing mix modeling
- Marketing mix modeling estimates each channel's contribution from data aggregated over time. Data-driven attribution works on individual paths.
Related concepts
- Advertising incrementality
- Marketing mix modeling
- Conversion
- Customer acquisition cost
- Google Ads
- Executive dashboard
Further reading
- Multi-touch attribution: what actually remains possible
- Executives: proving your advertising caused the sale
- Conversion tracking: the three gaps that skew everything else
- Marketing metrics: the ones that change a decision
- Google Ads ROI: from the reported ROAS to real profit
Related services
Frequently asked questions
Can I switch back to last click?
Yes. Google Ads still lets you choose last click for a conversion action. Make the change before an evaluation period, never in the middle of one.
Does the attribution model change bidding?
Yes. Google Ads automated bidding learns from conversions as the model distributes them. Changing the model therefore changes the data it learns from.
