If you’ve landed here, you’ve probably checked two platforms providing website analytics and wondered why there is a discrepancy. The good news is this is a very common question asked by anyone who’s reviewed data from another platform against Google Analytics. There are a few great resources on the topic we’ve compiled, which we hope will add clarity to this age old question.
10 Reasons Your Google Analytics Data May Not Match Other Sources | Seer Interactive Insights
Understanding Data Discrepancies in GA4 and Looker Studio | Seer Interactive
Ironically, many users report that website metrics do not match up between two of Google’s own tools.
A few major reasons that can cause these discrepancies.
When you’re sending traffic through tags on landing pages, the GA4 tracking code must be available on those pages. Otherwise, you will see clicks in one platform, but sessions will be missing from GA4, or vice versa.
This can also happen if the tracking code is wrong or didn’t fire for any technical reason. So, it’s important to do some quality assurance before launching any campaigns.
Different platforms use different types of attribution models, which can cause discrepancies in their data.
For instance, by default, some platforms may use a 7-day click or 1-day view model. However, GA4 uses a cross-channel data-driven attribution model. These settings can even be modified in GA which can create further discrepancies.
You may see discrepancy because one platform would give credit to tracked events, whereas GA4’s algorithm will attribute conversions using their own proprietary methodology.
Google Analytics 4 can suffer from sampling which occurs after a 10 million events quota is met.
Thresholding can also limit the data you can view in GA4. It happens when there’s a possibility of knowing the user’s demographics, interests, or any other info that could lead to revealing information about their identity or personal preferences.
It simply replaces the dimension with ‘Unknown’ so you can see the numbers, but cannot get any more info. These thresholds are defined by Google and are not adjustable.
Unlike sampling, it can happen in standard reports and explorations. When thresholding is applied, you will see an orange triangle at the top.
All reports in GA4 come with some dimensions, each with a specific set of values you can see. For instance, the device category dimension can have up to 4 values, i.e., Desktop, Mobile, Tablet, and Smart TV. Thus its cardinality is 4.
The number of unique values under a dimension is its cardinality. Normally, the dimensions with a few unique values won’t have high cardinality, but a dimension like Page location can have thousands of different values.
As per Google, if a dimension has more than 500 unique values in one day, it is considered high-cardinality. This results in hitting the row limit for that report. You will see another row with (other) when high-cardinality kicks in.
Usually, high-cardinality hits in the standard reports and you won’t see them in your explorations.
To deal with these GA4 issues, try using shorter date ranges, predefined dimensions, and only use high-cardinality dimensions when it’s necessary. Simply changing the date range in GA can produce variations in how they sample data.
Different tools can have different definitions for the same metrics, layer that onto the complex nature of how each user’s device interacts with the open web, data sampling applied by Google, how different vendors handle bot traffic, and more, it’s safe to say that expecting GA to match other platforms is an impossible expectation. What’s important is to leverage each platform for its intended purpose and benefits. The Octane11 tag is optimized to the B2B use case and provides account signal tied channels driving visits to the website, whereas Google Analytics is not optimized for B2B and does not provide account signal.
