Each report dimension (e.g., Page, Browser, Screen Resolution, etc.) has a number of values that can be assigned to it. The total number of unique values for a dimension is known as its cardinality. For instance, the Mobile (or ga:isMobile) dimension has two potential values (Yes or No), so the cardinality for that dimension is two. Other dimensions can have any number of values assigned. For example, the Page dimension has a different value for every URL that appears on your site.
Dimensions with a large number of possible values are known as high-cardinality dimensions. Reports containing high-cardinality dimensions may be affected by Analytics system limits, resulting in the creation of a rolled-up (other) entry in the report to contain the data that exceeds these limits.In this article:
Reduce the number of possible values in the Page dimension by filtering out any dynamic session/customer ID variables in the Query parameter settings. You can exclude query parameters from your reports by stripping them from URLs. To strip the query parameters, edit your view settings. Under Site Search Settings, edit the Query parameter options.
You can use this spreadsheet to extract query parameters. The spreadsheet uses the Analytics API add-on for Google Sheets to query the list of pages for the last 7 days and extract all unique query parameters into an easy-to-read list. You can go through the list, mark the parameters you'd like to exclude, and then generate a comma-separated string that you paste directly into the Exclude URL Query Parameters option under your View Settings. Instructions are included in the spreadsheet.
In depth explanation
Processed tables speed up reporting
Analytics uses a visits table to store raw data about each session on your property. In addition, pre-aggregated data for common reports is stored in processed tables (referred to as aggregate tables). These processed tables allow commonly requested reports to be loaded quickly and without sampling.1
Data limits for daily processed tables
Daily processed tables store a maximum of 50,000 rows for standard Analytics and 1,000,000 rows for Google Analytics 360. This means that, for each processed table, many unique dimension-value combinations2 are stored per day. When there are more than that number of dimension-value combinations for a given table, Analytics stores the top N values3 and creates an (other) row for the remaining value combinations.
For 360 clients using Custom Tables, there is a higher daily limit on those tables of 1M rows per day. For more information on Custom Tables and their limits, read About Custom Tables.
Data limits for multi-day processed tables
To further speed up reports with longer date ranges, Analytics uses multi-day processed tables, which each contain 4 days' worth of data. Multi-day tables are created from the corresponding daily processed tables and store a maximum of 100k rows for standard Analytics and 150k rows for 360. As with daily processed tables, when there are more than that number of unique dimension value combinations for a given multi-day table, Analytics stores the top N values and creates an (other) row for the remaining value combinations.
Because dimension values (e.g., unique URLs or campaign keywords) often repeat across given days, the multi-day processed table limits typically only affect sites with a lot of unique content and/or keywords.
Report query limit
An additional system limit that may result in the creation of an aggregate (other) entry is the report query limit. For any date range, Analytics returns a maximum of 1M rows for the report. Rows in excess of 1M are rolled-up into an (other) row.
Import high-cardinality dimensions
Using Query-Time data import on user, product, or content data may result in your data being rolled into the (other) row in your reports (see below).
What does this mean for my reports?
When you view a report with a high-cardinality dimension that exceeds the above limits, you won't see all of the values for that dimension because some values are rolled-up into an (other) entry. Please note that Analytics is still tracking data normally but is simply unable to display all data uniquely in your reports.
It's also important to note that the values that are rolled-up into (other) may change. For example, let's say on March 1, the "/categories/hats" page was your 49,999th most visited URL with 3 pageviews, but on March 2, it was your 50,001st most visited URL with 1 pageview. If you request a report in Analytics covering those two days, you will see the "/categories/hats" URL appear with 3 pageviews because the pageview for March 2 would be rolled-up into (other).
1In cases where the user query cannot be satisfied by existing processed tables, Analytics goes back to the raw visits data to compute the requested information. In that situation, your report will have a maximum of 1M unique dimension value combinations included in the sample set for the query.↑
3The top value combinations are determined by sorting the table by the relevant metric (e.g., Sessions Pageviews, Transactions, etc.). Lower volume dimension value combinations are then aggregated into the (other) entry.↑