Data-driven attribution is a way of deciding which ads get credit for a conversion by looking at your own account’s data, rather than following a fixed rule. If someone clicked three of your ads before making an enquiry, it splits the credit between them according to how much each one appears to have changed the chance of that enquiry happening.
How data-driven attribution works
An attribution model is the rule for sharing credit. The simplest, last-click attribution, gives all of it to the final ad clicked. Data-driven attribution, often shortened to DDA, compares the paths of people who converted with the paths of people who did not. If people who clicked a particular generic search ad early on convert noticeably more often than similar people who did not, that ad earns a share of the credit, even though it was not the last click.
Credit is fractional. A conversion might be recorded as 0.6 to a brand search campaign and 0.4 to a generic one, which is why conversion columns in Google Ads sometimes show decimals. In Google Ads, the model covers interactions with Google ads only: search and Shopping clicks, YouTube engagements and other Google placements. It does not see your email, organic or Meta traffic. Google Analytics 4 also offers data-driven attribution, but there it shares credit across all channels, so the two will not agree.
Data-driven attribution is the default for new conversion actions in Google Ads. Google removed the older rule-based options, such as first click, linear, time decay and position-based, in 2023, leaving data-driven and last click as the main choices.
Why it matters
The model decides which campaigns look successful, and Smart Bidding uses the same credit to set bids. Under last click, brand campaigns tend to look excellent because people search your name just before buying, while the generic searches that introduced them look weak. A UK furniture retailer judging on last click might cut a generic “oak dining table” campaign that was actually starting most journeys. DDA gives that earlier campaign some of the credit, which usually leads to a fairer budget split.
It is still a model, not proof. It estimates contribution from patterns in your data and cannot see offline influences such as word of mouth or a mention in the local paper.
Common mistakes
- Comparing Google Ads conversions with GA4 conversions and assuming one is wrong, when they use different attribution scopes.
- Switching attribution model mid-month and then comparing performance across the change.
- Treating fractional credit as a precise measure of each campaign’s value.
- Leaving an old conversion action on last click while new ones use DDA, so campaigns are judged by mixed rules.
- Expecting DDA to account for channels outside Google Ads.
How to act on it
Check which model each conversion action uses under Goals, then Conversions, and settle on one for the actions you bid on. For most accounts that is data-driven. When you change a model, note the date and allow a few weeks before judging the effect, because reported conversions shift between campaigns.
Use the model comparison report to see which campaigns gain or lose credit compared with last click. That shows you which campaigns start journeys rather than finish them. For a fuller picture across Google, Meta and email, judge spend against total business results as well. Working out which channels genuinely drive enquiries is the core of performance marketing across paid channels.