This has me baffled, and brings me here to make my first post.
I had an issue with date filtering on an existing report using GA4 data, so I recreated a minimal version to test. The issue remains.
Strangely, if I set the date range to 2025-01-01 → 2025-11-30 I get much higher numbers than if I extend the range to 2025-12-31.
A sample can be viewed here.
Can anyone help me identify what might be causing this?
Fascinating.
I could pinpoint it down to a day where the numbers drop. From Dec 7 to Dec 8, specifically, if you keep the start date as Jan 1.
On the flip side, if you keep the end date as Dec 31, the drop happens between Feb 12 and Feb 13.
This makes me think that exceeding a certain number of days or weeks changes the logic of your metric, somehow?
I would love to see the definition of the Views metric here.
Hey Sam! Thanks for your reply. Unfortunately I didn’t receive a notification about it so only just saw it.
The Views metric is literally just GA4 page_views via the default GA4 connector.
That’s super weird, then. I haven’t seen anything like this before, so I assumed that you must have some special edge case.
It could be a bug in the GA4 connector, but if that were the case, I’d expect to see more reports of such an issue.
If you recreate the chart from scratch, or even go a step further and recreate the report, does the problem persist?
Hey @sam8, yeah—that example you’re looking at is a recreation of the original report which has the same problem. I created it for that exact purpose, trying to isolate the problem—which it did.
Do you have any thoughts on how I could potentially progress resolution of this issue?
Gotcha. I’ve been looking for other instances of this coming up, and it does seem to be a persistent problem with GA4 specifically.
As far as I can tell, this has something to do with how GA4 enables data sampling when a data set exceeds a certain threshold. it goes from computing exact numbers to estimating.
It’s not something we can correct from the Looker Studio side, but I’ll see if there’s anything the GA4 team can do to address.
Thanks @sam8! Yeah, I assumed it had something to do with sampling but that’s a huge difference in numbers. Would be great if you were able to advocate for us with the GA4 team.
