Cohort Reporting

MoPub cohort reporting MVP offers insights that are tied to a point of reference from the day of “first ad request”. In other words, the “first monetizable cohort” provides a set of cumulative and daily metrics with key breakdowns that shows granular insights when changes occur. Please make sure to check cohort reporting daily and particularly pay attention if you made any change to ads monetization strategy, user acquisition activities or app logic.

Overview

Data Glossary

Section Name of data Definition
Cohort Type First monetizable cohort The cohort type from cohort reporting MVP is “first monetizable cohort”, which is a group of devices that send ad requests to an app for the first time. This signal is being used as a proxy for our system to recognize the “new device” of an app.
Filter Cohort date range The date range of the cohorts. It is a required filter. Example - If the date range is 1/1-1/5 , the two-dimensional heatmap will show 5 cohorts of data based on the date range for comparison. For the single-dimensional line chart/tab, it will show the oldest date of cohort data (1/1’s cohort in this example) with in-depth metrics and various breakdowns.
  App The app that cohorts belong to. It is a required filter
  Country Country filter. This is also available as line chart breakdown.
  Format Ad format filter. This is also available as line chart breakdown.
  Ad Partner Ad partner, including networks or MoPub marketplace. This is also available as line chart breakdown.
Tab/Line Chart Ad revenue per cohort (cumulative) When the “cumulative toggle” is enabled for the first tab, this metric shows the cumulative ad revenue of this cohort over the cohort size, which is the proxy of LTV insight.
Formula: (cumulative ad server revenue) / (# of cohort size).

The tab value is the most recent DX cumulative revenue per cohort.
  Ad revenue per user (daily) When the “cumulative toggle” is disabled, this metric shows the average daily ad revenue generated by active users from this cohort over daily active users from this cohort.
Formula: (daily ad server revenue of this cohort) / (# of daily active user in this cohort).

The tab value is the average of each day’s revenue per user.
  Retention This is the average percentage of users that return to the app in this cohort.
Formula: (# of active users in this cohort on DX) / (# of cohort size).

The tab value is the most recent DX retention.
  Impression per cohort (cumulative) When the “cumulative toggle” is enabled for the third tab, this metric shows the cumulative impression of this cohort over the cohort size.
Formula: (cumulative impressions) / (# of cohort size).

The tab value is the most recent DX cumulative impressions per cohort.
  Impression per user (daily) The daily impressions generated by active users from this cohort.The tab value is the total impression from this cohort within the date range, which is the aggregated value of daily impressions.
  eCPM The daily eCPM generated by this cohort.

The tab value is the Total rev (aggregated daily rev) /Total imp (aggregated daily imp)
Heatmap Ad revenue per cohort (cumulative)

Ad revenue per user (daily)
The heatmap shows multiple cohorts’ data based on selected date range with colored background. Each row represents one cohort’s data.
The metrics from the heatmap including
* Cohort size: total number of devices in this cohort
* Ad revenue per cohort (Cumulative)
* Retention
Overall, the darker shade of the cell color indicates the value of the metric is higher than others.

FAQ

  1. Is there any data limitation of cohort reporting?

Cohort reporting data comes from MoPub’s in-house data calculation. As we only store the underlying event-level data for a limited time frame, the minor data inaccuracy will occur when a device stops using the app for more than 90 days. For example, if a device originally belongs to 1/1 cohort, but stopped using the app for more than 90 days. When the device uses the app again on 4/1, we will label this device into 4/1’s cohort as we can no longer reference the 1/1 event-level data of this device. This will occur when a device is inactive/idle for more than 90 days for a given app.

  1. What will happen if an app is newly added to MoPub?

As cohort reporting data is calculated based on the device event-level data received by MoPub system, we will recognize all the app traffic coming into MoPub as the new device signal when the app is newly added to MoPub. This will produce a potential data spike for the first few days of cohort data after the app is live with MoPub. The data spike will organically go away after 1-7 days after MoPub collects a reasonable amount of event-level data and can differentiate first-monetizable devices from the old devices.

  1. Will MoPub Analytics have the same dataset?

Due to the complexity of viewing two dimensional dataset of cohort reporting, we will only display cohort reporting on publisher UI as a dedicated reporting page with its customized data visualization.

  1. Is the eCPM value in cohort reporting related to the autoCPM setting?

Yes. The eCPM data from cohort reporting is calculated based on MoPub ad server revenue and impressions. The MoPub ad server revenue comes from autoCPM value if the autoCPM (original & enhanced) feature is enabled for a given network.


Last updated November 01, 2021

TWITTER, MOPUB, and the Bird logo are trademarks of Twitter, Inc. or its affiliates. All third party logos and trademarks included are the property of their respective owners.

© 2021 MoPub (a division of Twitter, Inc.)