Sampling in Google Analytics

2019-08-28 14:30:49 UTC

Sampling in Analytics is the practice of selecting a subset of data from your traffic and reporting on the trends available in that sample set. Sampling is widely used in statistical analysis because analyzing a subset of data gives similar results to analyzing all of the data. In addition, sampling speeds up processing for reports when the volume of data is so large as to slow down report queries.

As a Visiolink client with a high number of users, you might be subject to sampling. You can always see when Google is using a sample instead of all the data. This is shown as a small message just below the date range picker:

When the number is lower than (100% of sessions), Google is sampling the data. The sampling will therefore produce different totals depending on the level of sampling. The data will often still be accurate but it will not be exactly the same numbers. Two reports might produce slightly different totals at different sampling levels. When comparing the same query on the different days, the result might therefore be slightly different. Be aware that sampling does not happen unless you have a large amount of data. It is difficult to locate the exact threshold but large clients tend to experience sampling. Our recommendation is to be aware of sampling and keep it in mind when creating queries for large amount of data.

Read more about sampling in Google Analytics here:
ABOUT DATA SAMPLING

There are a few ways to avoid sampling depending on your solution:

  • Reduce the data range

Ask yourself if the large data range is necessary to achieve what you are trying to achieve. Often smaller amounts of data will suffice when looking for a pattern or information within the data.

  • Query partitioning

If you want to look at data for a longer period, you can split the queries into smaller pieces. For example if you want to look at one year’s worth of data, query data for one month at a time and export it.

  • Invest in Google Analytics 360

Your sampling threshold when using Google Analytics 360 will be much higher, and you will most likely avoid sampling all together. Furthermore, you will gain access to new tools such as "Scheduled Unsampled Reports" and "Custom Tables". We recommend upgrading to Google Analytics 360, if you have a large dataset and are actively using and relying on data in your daily operations. If you have Google Analytics 360, you have some more options for avoiding sampling. These are described below.

________________________________________________________________________________

One solution to the sampling problem in GA 360, is to create "Custom Tables" containing the same metrics and dimensions you need on a daily basis. Then Google Analytics will aggregate data from these metrics/dimensions daily, giving you faster and unsampled access to these specific data. When the "Custom Tables" have been created, all future custom reports containing the same metrics/dimensions as these custom tables, will not contain any sampling.

You can read more about custom tables and how to create them here:
ABOUT CUSTOM TABLES

Another solution for GA 360 customers is to create "Unsampled Reports". To do this, you need to hover over the yellow checkmark visible in the custom report output. Then press "Greater Precision" and press "Schedule downloadable unsampled report". In here you select your desired settings and press "Request Unsampled". You can watch the progress of the request in the tab "Unsampled Reports". Most requests takes about 5-15 minutes for Google to process. The final result is an unsampled report showing exact numbers from the chosen date range.

You can read about unsampled reports in here:
ABOUT UNSAMPLED REPORTS

 

Feel free to contact us with any questions you might have regarding tracking and Google Analytics.

Var denne artikel en hjælp?
0 ud af 0 fandt dette nyttigt
Har du flere spørgsmål? Indsend en anmodning

Kommentarer