Data Explained | Why isn't the Google Analytics user metric included in ChannelMix standard datasets?
Jared Sloan
Question
Why isn't Google Analytics' user metric included in the Google Analytics standard datasets?
Data Explained
Users in ChannelMix will not match what you see in Google Analytics, because we pull data one day at a time from Google Analytics.
The number of users pulled daily (i.e. daily users) added up across the whole month will be more than the number of users pulled monthly (i.e. monthly users). This is because one user may come back to the website multiple times during the month. If we pulled data at a month time frame that user would be counted once... but since we pull daily data the user would be counted once every day they visit the website.
Example
A user visits the website on 8/16, 8/20, 8/27, and 8/30.
If we looked at the entire month of August in ChannelMix there would be 4 users.
If we looked at the entire month of August in Google Analytics there would be 1 user.
Are there any reporting alternatives?
Since Google Analytics is dynamically calculating users based on the timeline you pick and the level of granularity in the report it's not going to be possible to recreate it. This is the biggest reason that we don't include it in the data sets by default. Here are some alternatives to communicate data points using users.
Average Daily Users - If you write a calculation that ignores the other dimensions then you can calculate average daily users. This metric gives you some level of uniqueness and still provides an accurate representation of the data. {FIXED [Report Date] : AVG(Users)}
Average Weekly Users - This would require creating a new data set that only provides the fields Report Week and Users. We would pull the whole week together to get an accurate representation of users for the week and you could use the data in the dashboard in a weekly SUM or in a scorecard by doing another average similar to the daily one above.
Average Monthly Users - This would be the same concept as the weekly users, but we would only pull the data once a month and pull the whole month together to get an accurate representation of monthly users. Weekly and Monthly user counts could go in the same data set, but the ChannelMix Profiles would need to be different to indicate how the data is scheduled.