This doesn’t sound so much as a GA issue, but rather the challenges posed by anonymity on the…
Ben Easley

You’re right about the fact that these challenges are global and aren’t an issue specific to GA.

My main critic lays more on the data perspective than on the attribution issue. What I’m trying to show is that GA solves a problem by modifying the data before showing it to you. Changing the attribution of a session from “unknown” to “last known” is not a conservative change at all. It severely impacts your metrics. You can be fine with it, but if you’re not, there’s no going back. You can’t access the unmodified data to check it or to alter it.

Personally, and I insist on “personally”, I’d rather have a tool saying “I don’t know, do whatever you want with that unknown stuff” than a tool telling me “this is organic” when it’s not. To me, it’s just a lie. But it’s just me.

That’s why I’m talking about “data philosophy”. I think that if you really care about data accuracy, it’s hard to roll with something that alters your original data without you knowing it.

To talk a bit about attribution, it’s totally valid to roll with GA and to be really happy with how they handle attribution. It’s just really important to understand what’s happening because … a lot is happening.

I don’t know about how GA 360 handles this attribution challenge. The events are stored in BigQuery but … are they stored before the application of GA’s attribution logic or after? Can you alter/clean them?

By the way, thanks a lot for reading me and taking the time to comment!