A Framework for Resolving User Identities in BigQuery Using DBT & GA4

[kumartk]

Vincent gave us perhaps the most technically interesting presentation of the day. He presented a valuable approach to addressing the challenges of integrating disparate data sources, where there can be significant heterogeneity in terms of user IDs.

Using DBT, Vincent presented a solution that significantly reduces the complexity of using complex JOINs to gain a 360-degree view of users across different systems or data sources.

The goal of his model is to create temporarily grouped common IDs that summarize costa rica phone number data different user IDs. This allows this grouped ID to be used instead of complicated connections between different identifiers.

If you want to know more about it, you can find everything here: Human 37's GitHub repository.

Red Bull's Solution to Anonymize User Data in Google Analytics
A lecture by Maximilian Plötzeneder

Red Bull didn't surprise us in the least with the quality of their presentation: great storytelling and visually impressive. But beyond that, Maximilian told us about the efforts they've made to fully guarantee the privacy and anonymity of the data collected.

The issue has been high on the agenda since early 2022, when the NGO NOYB began claiming that the use of Google Analytics meant the transfer of European users' personal data to Google servers in the US, which violated European data protection laws.

Using GTM server-side, which was new to the market at the time, Red Bull sought a way to break the connection between the user and Google Analytics. The goal was to capture the data, isolate it, and anonymize it before sending it to Google for processing. All this without interfering with the proper functioning of Google Analytics.

This includes measures such as hashing the GA cookie, reducing the user agent and IP address, and minimizing data collection to the minimum necessary for business purposes.

Definitely an inspiring example of how you can continue to collect useful and actionable data without compromising the privacy of your users.