Skip to main content

Overview calculation logic

Last updated 06/25/2026

The overview value summarizes retention over a period of time and helps you quickly understand how the data performs. This section explains how overview values are calculated. You can look up the calculation logic that matches the definition of your metric.

Logic of preset metrics​

Whether or not Also Show is turned on, the Retention Analysis model calculates the following preset metrics. Their overview values are calculated as follows:

Analysis metricsDefault calculation logicOptional calculation logic
Initial usersSum, that is, the simple sum of each date's dataAverage, that is, the arithmetic mean: Sum / time range
Retained usersSum, that is, the simple sum of each date's dataAverage, that is, the arithmetic mean: Sum / time range
Retention rateWeighted average, that is, the simple sum of retained users / the simple sum of initial users-
Churned usersSum, that is, the simple sum of each date's dataAverage, that is, the arithmetic mean: Sum / time range
Churn rateWeighted average, that is, the simple sum of churned users / the simple sum of initial users-

Logic of Also Show metrics​

For Also Show metrics with different definitions, the overview uses different default calculation logic to fit actual business scenarios.

  • Also Show metrics with return user metrics only
How the return metric is setDefault calculation logicOptional calculation logic
A formula, or not a formula and not using a per-user calculation method

Average of the return user metric across dates

  • Sum, that is, the simple sum of each date's value
  • Weighted average, which is equivalent to the arithmetic mean in this case

Per-user calculation methods

For example: Times per user, Per User, accumulative sum user average

Weighted average

For example, the overview value of times per user equals (total times on D1 + total times on D2 + ...) / (users on D1 + users on D2 + ...)

  • Sum, that is, the simple sum of each date's per-user value
  • Average, that is, the simple sum of each date's per-user value, then the arithmetic mean
  • Metric on Initial date

The overview value of a metric on the initial date uses Average by default, and Sum is optional.

  • Also Show metrics that include metrics on the initial date
How the metric calculation is setDefault calculation logicOptional calculation logic
The arithmetic operation between the return user metric and the metric on the initial date is addition, subtraction, or multiplication

Average

  • Sum, that is, the simple sum of each date's Also Show metric value
  • Weighted average, which is equivalent to the arithmetic mean in this case

The arithmetic operation is division

Weighted average

That is, the sum of each date's return user metric / the sum of each date's metric on the initial date.

  • Sum, that is, the simple sum of each date's Also Show metric value
  • Average, that is, the simple sum of each date's Also Show metric value, then the arithmetic mean

Impact of incomplete data on overview values​

Incomplete data is excluded when overview values are calculated, and only complete data is used. To avoid misunderstanding, you can select Exclude partial dates.

For example, if you view the overview value of the next-day retention rate for February 1~5 on February 6, the next day of February 5 (February 6) hasn't ended yet, so the overview value equals the weighted average of the retention rates for February 1~February 4, and neither the initial users nor the next-day retained users of February 5 are counted in the overview value.

Was this page helpful?