Fewer pages/keywords with expanded date range

If your website has traffic across a large number of pages or keywords, you might notice that fewer pages/keywords are reported when you expand the date range of the report. For example, if your website contains many thousands of unique pages (URLs) and you are continually adding new ones, a wide date-range query might request many thousands of rows from the database. Under these circumstances, your report would take a very long time to load in order to display all of the unique URLs for that date range. To avoid long loading times, Google Analytics samples the data by retrieving a maximum of 1,000,000 unique URLs or campaign keywords for any request, divided by the number of days in the request.

For example, suppose your site has visits to 1,000,000 unique pages over the last 60 days:

  • A report for the past 30 days displays approximately 30,000 pages.
  • A report for the past 60 days displays a maximum of 16,000 pages.

Because unique URLs and campaign keywords repeat across given days, this threshold is reached only by sites with a lot of unique content and/or keywords.

If your website has visitors via many unique pages or keywords, try using one of the following tips:

  • Use view filters to collect sections of your website data, since the sampling occurs at the view level.
  • Reduce the date range of your request to one month at a time.
    The smaller your date range, the better. For example, a request of 20 days or fewer typically results in no sampling.

If you still need cumulative data for a large date range, consider using the Google Analytics Data Export API to make small date range requests and tally your numbers using your application and a database. This is a common mechanism that customers use to manipulate and store their data in a fast and convenient way. Read more about the Google Analytics Data Export API.

Note: Keep in mind that there is no loss of data in your reports when sampling occurs. As the illustration above demonstrates, you can decrease the date range of your query to retrieve the exact data that was sampled for the longer date-range request.

Sampling occurs by design, and we alert you when a report contains sampled data. For more information on sampling, see our Sampling Guide on Google Code.