Countly Analytics Data not adding up

We are seeing some strange data results from various reports here on our Countly.

Example 1;

When inside an App on the Dashboard, selecting "2017" gives;
4861 Total Sessions, 452 Total Users, 466 New Users, 6.4 days time spent, 1.9 min avg time spent and 66.6 avg requests received.

When Selecting the custom timeframe from January 1st to May 5th (today)

We Receive ;
4861 Total Sessions, 474 Total Users, 466 New Users, 6.4 days time spent, 1.9 min avg time spent and 34.7 avg requests received.

Why does the Total Users value and AVG Requests Time value change?

Example 2

On the same App, under Analytics > Platforms

If we add up Total Users from iOS and Android
346 iOS and 125 Android = 471 Total Users

This doesn't match the total users under Analytics > Users....



  • Hi Bryan,

    What is the Countly version you are using currently?

    Comment actions Permalink
  • I am one of the team members investigating this. We found it on v16.06 but I also see it on v16.12.2 (the version we are updating to).

    Comment actions Permalink
  • For Community Edition, on the main dashboard, other than full time (this year, month or day), total user (unique) count is estimated and corrected using the biggest time buckets from available daily, weekly and monthly stats.

    For Enterprise Edition, this data can be calculated using Drill, so this correction/abbreviation doesnt happen.

    Hope this helps.

    Comment actions Permalink
  • Hey gorkem!

    The 16.12.3 update seemed to fix the main issue we were seeing. Now we are seeing a different analytic issue.

    In Overview Dashboard for one specific Application we have 479 Total Users and 474 New Users.

    If we then go to Engagement > User Loyalty; and add the number of users, that total number of users is 494.

    How is the Total Number of Users in Engagement > User loyalty being calculated?

    Comment actions Permalink
  • Hello,
    since loyalty is precalculated metric, it is possible that same users (if achieved in different sub periods) can fall into multiple buckets, and that's why you will see more users than expected.
    We have an internal issue of improving those metrics, but currently no estimation when that could happen.

    Comment actions Permalink

Please sign in to leave a comment.