Hello,
I'm french so excuse my english ^^
I created a calculated field (Continent2) to group some data (If country is "France" or "Italy" then continent is "Europe" else "Other"). I know there is already a continent dimension but this is to illustrate my issue more easily.
When I use a filter with Continent, I have this result :
And when I use Continent2 (my calculated field) as a filter, I have this :
As you can see, the value of metrics is very different, especially the "Average session duration" metric with a completely wrong result.
Do you khow why ? Is there a solution ?
Thanks for your help.
Bonjour William 🙂
So:
- Are you sure the mapping is exactly the same? Can you check it by putting both Country, Continent, Continent2 in a table and see what's out?
- Do you have the same numbers in terms of Total users, sessions, ...?
As it's an AUT Average session duration, could be a weighting average issue as well, have you tried things like SUM(Total Duration)/SUM(Sessions)?
Merci,
First of all, thanks for your reply ^^
I did what you suggested and I got this :
As you can see, for "Americas", I have the same group of countries but the "Average sesssion duration" are completely different !
I don't have the "Total Duration" in my database so I can't recalculate it myself...
I hope you can help me or explain why I have this difference...
Thanks in advance ^^
Really sorry, had a look and indeed, custom grouping impacts this metric. Not sure how to fix it, as it seems the AUT aggregation is messing around 😕 Tried playing with User Engagement (Like SUM(User Engagement)/SUM(sessions)), not fixing the issue...
It's okay, thanks anyway. I've searched the internet a lot and haven't found any solutions. It's certainly impossible at this time and Google needs to fix it. Have a nice day !
Hi, I've read the attachment you provided. I think this type of question belongs to GA4 issues rather than Looker Studio. Looker Studio simply displays the data based on what you provide.
By the way, in GA4, "Average Session Duration" is calculated using the "user_engagement_msec" parameter (from the user_engagement event). So, I suggest checking the unusual traffic—where it comes from, the landing page, the first visit time, the traffic source (medium), etc. Hope this helps you gain some insights!
Hello, with what you have proposed, it may be possible to recalculate "Average Session Duration" but I have the same issue with "Total users" and I don't think we can recalculate this metric because we have no information on the user such as his name or ID.
Thanks anyway for your message.
User | Count |
---|---|
3 | |
1 | |
1 | |
1 | |
1 |