Data discrepancies in Google Analytics 4: misunderstandings and solutions

A significant portion of the support tickets we receive is related to what users perceive as “data discrepancies” in Google Analytics 4 (GA4) data.

It’s essential to underline that in 99% of these cases, the root cause is not a genuine discrepancy but a difference in data querying.

This article aims to shed light on the most common misconceptions and solutions related to GA4 data mismatches.

Suppose a user examines a specific chart in GA4 and then attempts to retrieve the same data in Windsor.ai, only to find that the data appears incorrect. The issue here is not a data discrepancy. Instead, it’s a result of different dimension and metric requests.

For example, a user looks at this chart in Google Analytics 4:

And then tries to query this data in Windsor.ai:

Now the data looks wrong.

The issue here is not that there is a data discrepancy but that the requested dimensions and metrics are different.

GA4 has various dimension scopes, including:

• Basic
• Account
• Session
• User
• Event
• Cohort
  …and more

For instance, if you query a campaign on a session scope (see here) coupled with the metric sessions, the result will be the campaign for each session in your GA4 property.

Querying a campaign on an event scope (see here) with the metric sessions will yield a session for each campaign preceding the event.

When a campaign is queried on the user scope (see here) together with a session, the result will be a session for each unique user on your GA4 property.

If you query the same data in GA4 as you do with the Windsor.ai connector, the numbers will align.

Another common source of confusion is the apparent discrepancy in user counts when the same time window is selected in GA4 and Windsor.ai. This difference is not an error but a result of repeat user visits.

In GA4, a user equates to a person (or browser ID). If a user visits your website twice within the date range you query (excluding the date), they are counted once.

However, if you include the date, they are counted twice: once for each visit on different days. When we include the date column in GA4, the numbers begin to align.

In summary, use the explore section in GA4 and query exactly the same combination of metrics and dimensions as you do in your Windsor.ai connector. The numbers should match.

Remember that most dimensions can be queried on multiple scopes (e.g., a campaign can be queried on a session and a user scope), so it’s crucial to get it right.

This article will continually be updated with new cases as they are discovered. For now, these explanations should clarify 90% of the “issues” reported.

• Google Analytics 4 Field List Documentation: Here, you can see all the different scopes listed next to the description.
• Metrics and Dimensions Explorer by Google: This tool allows you to see which metrics and dimensions can be queried together and their respective scopes.