Google Analytics - High-Cardinality Dimension warning is dangerous

"This report includes a high-cardinality dimension, and some data has been grouped into (other)."

Have you ever seen a message like this in your Google Analytics reports? My congratulations and also my condolences! Congratulations because your business is attracting lots of traffic (hope you have a high conversion rate) and condolences because you have now reached the Google Analytics limit and need to deal with it.

Google Analytics High-Cardinality dimensions

 

 

 

 

Most of the reports you see in the Google Analytics Portal have been pre-calculated and pre-aggregated already on a daily basis. This saves the Google Analytics engine lots of resources and seriously speeds up report processing because there is no need to aggregate the same metrics over and over again. For those of you who know what a database is and how to use SQL (and I believe every Digital Analyst should have this knowledge): Google Analytics creates a separate table for each day with aggregated metrics per current report dimensions combinations. Opening the standard report in Google Analytics will only result in combining daily aggregated metrics, which is a thousand times faster. To save space during the pre-aggregation phase, Google Analytics limits the number of rows (number of unique dimensions combinations) per these pre-aggregated tables. The limits are below.

Some processing limits of Google Analytics

Please refer to this article for further limits: Google Analytics Limits and Quotes

  • Google Analytics free - 50,000 rows per each table (report)
  • Analytics 360 (paid version) - 75,000 rows per each table (report)

If during the day, the number of unique dimension combinations has gone over these limits, the values with lower aggregated metrics will be combined into one row using (other) as the dimension value.

Additionally, Google Analytics creates multi-day (4 days) pre-aggregated tables limited to 100,000 and 150,000 rows for free and paid versions of Google Analytics. This means you are more likely to see the high-cardinality warning when you query more than 4 days of data.

Why is this a problem?

You may well ask, if Google Analytics summarises dimensions with lower metrics into one - what's the problem? Those are very likely to be rarely appearing values and I don't need them in my report. You are right. Well, most of the time.

The thing is that the aggregation limit is applied during data processing. Once and for all. If you like to use the report filter option (see screenshot below) - you need to remember that any dimension values you are searching for might be hidden in the (other) dimension and therefore not displayed at all!

 Google Analytics Report Filter

In the screenshot above, the page /My-Rarely-Visited-Page was grouped into the (other) dimension (row). If the filter is applied, no data displays.

How to fight this

The good news is that it is pretty easy to fight the issue. All we need to do is to tell Google Analytics not to use pre-aggregated data by:

  • Applying Advanced Segment
  • Using Custom Report

Usually, I would use an Advanced Segment. Just create an Advanced Segment in Google Analytics which still includes all the data and apply it to the report. This little trick will force Google Analytics to re-calculate all metrics and re-aggregate all dimensions whilst avoiding the ugly (other) dimension. The Advanced Segment can be defined any way you like, provided it matches 100% of the Users / Sessions. In the example below, I have Advanced Segment defined to match all Users sessions where Traffic Source matches the .* Regex (empty string or any string). 

Google Analytics Catch-all Advanced Segment 

This report definition is also available for import using the Google Analytics Solution Gallery. Please use the link below to import the segment.

https://analytics.google.com/analytics/gallery/#posts/search/%3F_.viewId%3Dw55ZHd8gTYS9UCemfdlFrQ

Alternatively, you could import the Advanced Segment using this link https://analytics.google.com/analytics/web/template?uid=edw3GCRNToK9cUZaAyIECQ

Don't be scared of High Cardinality

High Cardinality dimension warnings are both exiting and not to be dreaded, as long as you know the right workaround and understand why the warning occurs.

Happy Analysis! 

Need help optimising your Google Analytics? Talk to us

Internetrix combines digital consulting with winning website design, smart website development and strong digital analytics and digital marketing skills to drive revenue or cut costs for our clients. We deliver web-based consulting, development and performance projects to customers across the Asia Pacific ranging from small business sole traders to ASX listed businesses and all levels of Australian government.

Comments (1)

Paul 05/10/2017 4:43pm

Very informative!