What is Cohort Explorer?
Cohort Explorer lets you analyze what your top groups (or cohorts) of users do within your application, and compare their usage trends and behaviors against other cohorts of customers. It's a highly flexible tool and offers a variety of filters you can use to zero in on the users that matter most to your business. You can select a cohort based on a user's date of install, first monetization, spending, engagement, sessions, ad source, age, gender, location and more. Once you've made your selection, Cohort Explorer will give you an in-depth breakdown of that cohort's in-app activity.
How can I find this?
Cohort Explorer is located in the Reports tab in the Upsight Analytics dashboard. Follow the steps below to access it:
- Click on "Reports" in the menu bar at the top of the dashboard.
- Select "Cohort Explorer" from the navigation table on the left.
Cohort Explorer Video Overview
Create Cohort - Discovery Mode
Use Case: Assume we ran a campaign on Nov 30th, 2012 and acquired some new users to install the application. For those users who installed the application on Nov 30th, 2012, we want to find out top 10 countries whose users spent most during the first week after the install.
- Select "Discovery" mode.
- Application Name: The application you want to test. Here we have chosen Cohort Demo - iOS.
- Cohort Event: available cohort events are "install" and "1st monetization."
- Cohort Event Date Range: The date range of the cohort, available options are one week, one month, or any date range you want to test. Here we set the date range to November 30th, 2012, since we want to only select users who installed the application on that day.
- Rank By: Metrics of cohort analysis, available metrics are number of users, top spenders, most engaged, number of transactions, page views, and sessions.
Within the First: Days, weeks, or months of data to be examined since the cohort event ("install" or "1st monetization"). Here we set the value to one week, so that the results will include one week’s data since November 30th, 2012.
- Within the Past: Choose how long ago to filter the values by.
- Group By: Filter cohort by install date, install source, age, gender, and location.
- Click "Run" to generate preview of cohort groups.
Create Cohort - Custom Mode
Use Case: We want to compare activities between male and female users who installed the application in the first week of October 2012.
- Select "Custom" mode.
- Application Name: The application you want to test.
- Cohort Event: Available cohort events are "install" and "1st monetization."
- Cohort Event Date Range: The date range of the cohort, available options are one week, one month, or any date range you want to test. Here we set October 1st, 2012 as the start date and seven (7) days as duration.
- Filter By: You can filter by gender, age group and country in this field.
- Total in the First: Days, weeks, or months of data to be examined since the cohort event (‘install’ or ‘first monetization’). Here we set the value to one week, so that the results will include one week’s data since October 1, 2012.
- Click "Run" to generate a preview of cohort groups.
Workspace is the place where users can manage their cohort groups. Users can add cohort groups which are created in the ‘Create Cohort’ section to Workspace by clicking the ‘add’ button, .
Once cohort groups are added into Workspace, users can:
- Save As: save selected cohort groups so that they can be reused later
- Load: load saved cohort groups to work with
- Clear: remove all cohort groups in the current workspace
- Export: export workspace table to CSV file
- Total/Avg.Per User: this function allows users to switch between total and average per user easily
What is a cohort?
A cohort is a group of users with similar user characteristics who installed an application on the same date. Therefore, it will yield greater analytical results because it has less individual differences than a typical market segment.
In terms of the Cohort Explorer, it is a group of users defined by an install date (or first monetization) and a number of different user characteristics which can range from demographics (age, gender, location) to behaviors (including custom events) to a source of an install.
A cohort can be a group of users from Zimbabwe who installed the app in the first week of November 2012. It can also be a group of users who made their first purchases within one week of the install.
In other words, for a group to be classified as a cohort, they must satisfy these requirements:
- Installed within a defined period
- Date of first monetization
And one of the below:
- Share the same demographic (e.g., age, gender, location)
- Performed the same behavior (e.g., purchased an item)
- Installed from the same source (e.g., Facebook invite, clicked an ad)
Result section plots metrics in cohort groups into daily charts. Available metrics are "total spending," "number of transaction," "sessions," and "time on site" (more metrics will be available in the future).
Cummulative charts can be view by "calendar date" or "number of days since event."
- Calendar Date: user can specify the date range of the data they want to look at. It can be a period of time within the cohort selection period
- Day Since Event: If user chooses this option, then the start date of the graph will be the start date of the cohort group and will continue user specified number of days since then
Either option allows user to view data beyond the date range of the cohort group. For example, although the date range of cohort groups in our use case is one week, we can plot two weeks of data (or longer) since the start date of the cohort group
Charts can also be viewed in ‘Cumulative’ or ‘Non-Cumulative’ way.
- Cumulative: the total value since the start date
- Non-Cumulative: the actual daily value
To plot metrics in charts, user needs to click the column headings; available metrics are "total spending," "number of transaction," "sessions," and "time on site" (more metrics will be available in the future).
Move cursor to the row of data to highlight the data series in the graph.
You can also hide any data series by clicking the colored dot next to it.
Who can access this data?
Cohort Explorer is available to all accounts, but you can restrict access for certain users via permission classes.
How much does Cohort Explorer cost?
Cohort Explorer comes standard with your Upsight Analytics service.
How often is data updated?
Data is updated daily.
When a certain user attributes' change, will the Cohort Explorer update and reflect the changes?
No, a user is permanently bucketed into a cohort after they have been processed.