Data Driven Attribution in Google Analytics(BETA)

You can find out which campaigns and channels improve your business performance. It is essential to have precise data on these contributions to innovate, improve and grow your marketing efforts.

Different sources and channels like subscriptions and sales which cause conversions can be determined. Contribution can be caused by more than one channel or source until the conversion is complete. Google Analytics Attribution(BETA) allows you to track users’ paths until a completed conversion.

Suppose that you have a potential customer as a mobile phone supplier.              

Customer’s Mobile Phone Purchasing Journey

• The customer makes a google search as “mobile phone recommendations “ and Google leads him to read your article on your blog.
• After a while, the customer makes a google search again for the mobile phone that you have recommended on your blog but this time sees your website on Google Ads.
• Because he is indecisive and wants to wait a little bit more and in the meantime, he notices remarking ads displayed while scrolling. He decides to buy this phone by clicking the ad.

If we wanted to analyze the channel that generated the conversion with Last non-Direct Click (Google Analytics default attribution model), we would have seen it as a Display. However, there are many channels that contribute to the conversion.

The Google Analytics Data-Driven Attribution model assigns credits to these paths. In the Attribution(beta) model, there are two types of attribution models: Rules-based Attribution Models and Data-driven Attribution Models.

Rules-based Attribution Models

Rules-based attribution models apply predefined rules regardless of conversion type and user behavior which are:

1. Last click: Gives all credit to the last channel.
2. First click: Gives all credit to the first channel.
3. Linear: Credits are evenly distributed across all channels clicked along the path.
4. Time decay: Gives more credit to clicks that happened closer in time.
5. Position-based: 40% credit is given to the first and last click, the remaining 20% is distributed equally to other clicks.
6. Ads-preferred last click: All credit gives to the last Google Ads channel the customer clicked before converting. If there is no Google Ads click, all credit is given to the last click. (This feature applies to Google Analytics 4.)

Data-driven Attribution Model

The data-driven attribution model, distributes credit using your account data with the help of machine learning algorithms.

It compares occurred and potential conversions using a counterfactual approach to determine the probability of resulting in conversion. All your marketing efforts are taken into account when crediting touch points. With Google’s Machine Learning, each channel’s value on conversion is more accurately determined.

In, GA3 data-driven attribution controls only the last four touch points, GA4’s touch points are being up to fifty. As follows all the marketing activities can be measured more precisely.

The data-driven attribution model and the Last Non-Direct Click (Google Analytics default attribution model) model can be compared as above.

In summary, the Data-Driven Attribution Model, dissimilar to other attribution models, gives credit to each touchpoint based on its impact on the realization of a conversion. The use of Google Machine learning is a more reliable approach to more accurately describe the customer journey, and provides a great deal of information on how the management of marketing campaigns can be changed and improved.

Notes:

• An account must have at least 600 conversions within 30 days for the Data-driven model to be available. (Only for Google Analytics Universal)
• All attribution models exclude direct visits from receiving attribution credit, unless the path to conversion consists entirely of direct visits.

References:

1. https://www.kristaseiden.com/
2. Data-driven attribution methodology in Attribution (Beta)
3. https://www.klickkonzept.de/online-marketing-blog/digital-analytics/customer-journey-nur-mit-attribution/