David Hemmerle

Very happy to see that server-side GTM supports googleAuth now! developers.google.com/tag-pl…

Apparently, Google Signals requires also the GTM client (if used to load the GTM script from the server) to activate the region-specific settings. This prevents the error „Google signals is missing required configuration: Please enable region-specific settings in your GA4 client“

It is the first time that this message is displayed to me. Strangely enough, the setting in the GA4 client does not change anything. In my tests, this error apparently prevents Google Signals from triggering properly. Has anyone had any experience with this?

The timestamp converter template for server-side GTM makes my live so much easier 🤗 Missing a Date object in the JS Sandbox has been a pain for too long. @stape_io actually went through the hassle of using pure JS to reconstruct an ISO timestamp. THX! tagmanager.google.com/galler…

Im neusten beyond pageviews podcast von @termfrequenz geht’s neben vielen anderen spannenden Themen auch um das serverseitige Session-Handling — bzw. um die irreführende UX auf meiner Website 🫣 beyondpageviews.podigee.io/9…

I was inspired by recent contributions form @SimoAhava and @mbaersch about the sessionization of MP setups. Depending on the platform, it can be a real challenge implementing a session control client-side. I hope this approach makes things easier but let me know about any flaws.

To achieve this I deployed a combination of the Firestore and a local caching, utilizing the templateDataStorage API to store the session information temporarily on the server. With this setup I managed to optimize the query costs implied for Firestore significantly.

In my latest article I describe my approach to custom server-side session management for GA4 and MP. It enhances tracking for these use cases: ✅ Cross-Domain ✅ Multi-device ✅ Cookieless ✅ Custom session controls ✅ Session counter & engagement time david-hemmerle.de/en/blog/se…

Interesting thread about A/B test and their perceived relevance by looking at significance only and ignoring other statistics. Confirms my experience that actual effect size is looked at not often enough.