Google Pushes Data-Driven Attribution Model, Sunsetting Most Other Models


May 9, 2023
Google to Automatically Switch Some Advertisers Attribution Models

This month Google sent an email outlining that any conversion actions that are using first click, linear, time decay, or position-based attribution models will be switched to the data-driven attribution model instead.  

They also outline that conversion actions will not be able to use first click, linear, time decay, or position-based attribution models beginning in June. As of September, any conversion actions not set to data-driven attribution will be switched automatically, but you will still have the option to use last click as an attribution model. 

What is Data-Driven Attribution?

Data-driven attribution uses machine learning that reads multiple signals to more accurately understand how each marketing technique has contributed to a conversion. It’s important to note that advertisers will only have access to data-driven and last click attribution models.

To switch off data-driven attribution, follow the instructions provided by Google here.

Why We Care

This is a bit of a nuisance, to be honest. There are certain areas where data-driven attribution is certainly an advantage (looking at you P-Max!), and if Google wants to recommend its favourite new feature, fine. But as always, verify that what Google tells you is good for you, is actually good for you. Up until very recently data-driven attribution was only available on high-volume accounts with a plethora of both click and conversion data. We are seeing recommendations to apply data-driven attribution, even in low volume accounts.

Now, the only other option you have, aside from data-driven, is last click. Last click attribution does a poor job of providing the full picture regarding a conversion action. Knowing the limitations of this attribution model is important if you don’t have the volume to properly leverage the data-driven model. 

When data-driven attribution is the best option for our clients, we are more than happy to use it. But the continued push towards automation, regardless of the data inputs, showcases an arrogance that can only come from a company that has cornered the digital advertising marketplace.

We recommend you review your affected accounts and examine if data-driven attribution makes sense using attribution modelling tools in both Google Ads and Google Analytics.  If you are not getting better information from data-driven attribution, use last click to gather as much as you can..