I’ve set up server-side tagging for GA4 on a few websites as an addition to the client-side setup.
When comparing the GA4 properties, I consistently notice that the server-side property shows more views, sessions, and other metrics associated with the source/medium “direct/none” and “not set” compared to the client-side property.
The GA4 property settings are completely identical.
Has anyone else experienced this?
Is there an explanation for it?
Parallel use of GA4 client-side and server-side always leads to problems in both properties.
This is due to the way gtag works, here for example it describes how to configure the rollup property, it can’t just be added in parallel:
Also, parallel use usually doesn’t make any sense. GA4 server side works as a proxy, so it’s not very clear what exactly you are trying to determine in this way?
We’re using a (temporary) parallel setup to check whether the data flows “correctly” into the server-side property before fully switching to server-side after a short observation period.
Google talks a lot of things but real-life scenario is different. So IMO implement either GA4 server-side or client-side. Parallel setup will cause you a lot of data discrepancies. That’s why it’s not recommended.
I am struggling the same think and Actually i do not get why we can’t run 2 separate properties in parallel, can someone explain on that?
To be honest it seems to be good practice to avoid data loss if ex. server crashes or something like that?
I am not getting why it is not allowed?
And what roll-up properties have to do with that - aren’t roll up properties just consolidation of multiple source GA4 properties?
It’s not forbidden, it’s just not recommended in my experience with this approach.
You can use this, but you need to consider that it will not work very correctly.