Mixpanel’s Segmentation and Insights reports present data in totals, uniques, and averages.
Selecting Count Type
In Segmentation reports you can change the count method by selecting the dropdown in the upper right corner of the report.
In Insights reports you can change the count method by selecting the dropdown that displays "Show Total" above the event selector in the upper left corner of the report.
The Formulas report allows you to use either totals or uniques, while Funnels and Retention count using unique values.
A total count in Segmentation and Insights will sums the total amount of instances an event is fired. For example, if one person fires an event 50 times, that will count as 50 when looking at Total.
Each data point sent to Mixpanel has a distinct_id. The distinct_id is the unique identifier for the user that fired that event. When looking at unique values, Mixpanel counts how many distinct_ids have fired that event.
When comparing data from a Funnel report to a Segmentation report, you will want to use the unique count for an apples-to-apples comparison since Funnel reports always use uniques.
When you look at an Insights report for unique users, you are going to see the number of unique users per segment. Unique users can appear in more than one segment.
For Funnels, a user will only be counted in a Funnel once, no matter how many additional times they do the first step. The Funnel will only use the first instance of the first step of the Funnel along with the corresponding event property values. The segmentation of the Funnel will reflect that.
Both Insights and Segmentation reports display the average number of times a specific event is fired by a unique user. The average value is calculated by taking the total count of a specific event and dividing that value by the number of unique users who performed that event.
User with distinct_id 123 fired Event A 50 times on June 2, 2016. User with distinct_id 456 fired Event A 100 times on June 2, 2016. User with distinct_id 789 fired Event A 300 times on June 2, 2016.
In Segmentation, you’d see the following values in this example when looking at Event A on June 2, 2016: