Hi @Dan , @Alex ,
I have Stape business plan and set up google ads server-side purchase conversion for two different Shopify store. The setup is identical but one is performing great (set up first) and the other is significantly under reporting compared to it’s client-side setup. I’ve two different conversion action. One for client-side and another for server-side.
I’ve configure both custom domains in my multi-domain stape server container. For example:
In server GTM I have distinguish the two purchase tags by adding a condition based on incoming host domain URL in the trigger. Both are firing only for respective incoming request domain URL and working well. I’ve tested.
However, it’s been one month and 2nd one is performing significantly bad. What could be the reason? If I create two separate stape server container for each domain, will it then perform better?
It’s hard to say definitively, I only have a few guesses.
From the screenshots of the tags, I don’t see any problems.
GA4 events are sent in some cases not to the server container, but directly to GA. So Google ADS tags on the server do not work in such cases.
This is a common reason if GA is initialised several times in parallel (via a plugin, additional Google tag in the container or just in the code there is a gtag snippet on the thank you page for example).
If there’s consent - there’s probably a problem with it somewhere on the site. For example, web tag does not work correctly with consent, or vice versa.
Also, if it’s a UK site and you are NOT using consent it’s possible there is a restriction on the accunt because of this and conversions are not performing well because of it.
The problem is with the Google ADS account itself. I’ve encountered this a few times only, but when I changed my account to another, everything started working without any problems.
Some other problem in the tracking setup. Debugging and experiments are required here.
In any case, using 1 server container on two sites does not affect Google ADS results if data mapping is configured correctly.
In any case, the use of 1 server container on two sites does not affect Google ADS results if data mapping is configured correctly and there are no cases, for example, when Lookup does not produce results. I’ve encountered such a problem because lookup table didn’t hendle that url can start with www.example.com, but expected only example.com.
So for lookup data I recommend to use this tag, where you can use ‘contains’: Tag Manager Template Gallery
“Google & YouTube” app is installed, and GA4 is connected.
Client-side GA4 tags are set up, with server-side tracking used only for Google Ads.
Could the GA4 connection through the app be causing any issues?
Yes, this is a UK store. So, server.abc.co.uk has consent mode implemented, while server.xyz.co.uk doesn’t. Would you mind if I shared both website URLs with you for a quick review? I can also provide access to both the web and server GTM containers if you’d like to take a closer look.
Interestingly, the underperforming Google Ads account’s client-side setup is working well, with no consent-related warnings in the account. So, I doubt my client will agree to make changes there.
I’m attaching screenshots of the entire web GTM tag setups for both Shopify stores to give you a clearer picture.
Unfortunately I can’t advise something more specific, you need to go deeper and test the setting, make some changes and see how it affects.
In my experience the cause of this is always one of the 4 reasons I described above, most often it is 1 or 2 reasons.
I’ve seen Google do a ‘shadow ban’ without any notification on the account about problems with consent. Since the UK has a consent compliance requirement about as well as the EU - I would definitely start by fixing this on both accounts.
It’s also worth making sure that GA4 hits with your conversion always come to sGTM, that way you can see if there’s a problem with it. I recommend using Logger for this and comparing the number of logs to the number of conversions on the backend.
