[John]

Unfortunately, due to security and privacy considerations, Google does not provide a direct way to fetch or suggest a user’s Google Analytics property ID from any site other than the Google Analytics dashboard itself. This helps to prevent unauthorized access to a user’s Google Analytics details. Therefore, your users will need to retrieve the property ID from their Google Analytics account and provide it to your site for the data integration.

[John]

Your issue may be with the timing of when your data is put into the dataLayer and when the tags are firing. The tags may be firing before the necessary data is available. The dataLayer works in a stack-like format where information is only available from the point it was added onwards. So, you may need to adjust when the purchasing data is pushed to the dataLayer.

[John]

Your SQL query isn’t inherently flawed, but there seems to be a mismatch between your actual data distribution and how the query processes the data. One possible reason for sky-high values from Day 1 onward might be multiple records per user in the ‘table’ you are using for the analysis. Ensure that the ‘first_open’ event is unique per user – ideally, there should only be one occurrence of ‘first_open’ per user. If there are multiple ‘first_open’ actions per user, that might create a redundancy in the ‘user_pseudo_id’ leading to inflated numbers after Day 0.

You may want to check your data source or tracking setup for any setting that might log multiple ‘first_open’ events for the same user. Another possibility could be that the ‘user_pseudo_id’ isn’t persistent for each user across different days, causing inaccurate counts of ‘unique’ users in your output.

Remember that even though the structure of your SQL query is okay, the input data quality and data tracking setup play a significant role. The query can only analyze based on the data it is given, and if that data is flawed in some way, it reflects in your analysis.

[John]

Alright, so here’s what we can do, mate:

1. First, we need to get these categories into your GTM. You’ve got two options for this: either parse every article page to grab the categories or ask the front-end to push them to the dataLayer on pageviews. The latter would probably be more robust.

2. Now, once you’ve got the list pulling into GTM, you probably want to sort it, maybe alphabetically? You can then use this in your event property.

3. Don’t forget about Google Analytics! Make sure to register the property in there.

Keep in mind though, when you’re analyzing this, it could get a bit tricky. You’ll mostly need to use ‘contains’ on this dimension. It’s not perfect, but at least it’s something. If you want more options in the future, you could look into doing ETL, or maybe even using a more advanced analytics solution. GA is pretty bare-bones because it’s generally free, right? So, it’s got its limitations.

[John]

Sure, I’d be happy to help. The issue you’re experiencing sounds like it could be related to the authentication process. When you use API clients like the Google Analytics Data API Python client, they require specific credentials to authenticate your request. In your case, it seems like the root of the issue might be that your developer account credentials – or the process you’re using to authenticate them – aren’t being accepted by the client. This could be due to a number of reasons, such as expired tokens, incorrect account scopes, or issues with your account’s access to the data you’re trying to retrieve. It’s also possible that there’s an issue with how the API client or your code manages these credentials.

I recommend checking for all these potential problems. Make sure your tokens are up-to-date and your account has the correct permissions and scopes to access the data you need. Try to isolate and identify any issues in how your code or the API client is handling credentials. If you continue to have trouble, for a more detailed analysis and maybe even workarounds, consider reaching out to Google’s support or to the broader developer community through forums, as they may have experience with similar issues. Good luck!

[John]

After the switch to GA4, if you’ve not updated your code and keep using the old window.ga method for tracking, it won’t throw Javascript errors in the browser. The function would still technically be valid, but the custom events you’re trying to track wouldn’t be sent to your GA account, as the UA property would no longer be operational. To consistently track events, you need to migrate to the GA4’s event model structure and change your tracking setup accordingly.