Cohort
Last updated
Last updated
A cohort table is a way to analyze retention by tracking how groups of users (called "cohorts") stay active or engaged with a product or service over time. Retention is a key metric that shows how well a product keeps its users over the long run, and the cohort table breaks down retention by looking at specific groups who started using the product at the same time.
Create a Blank Chart and find Cohort in the Chart Configuration
Select the pool event in the first bar - the retention will be measured within this pool. You can choose to filter it as the first time taking this action or any other dimension filters.
Select the event to measure the retention within that pool in the second bar, meaning "of all that involved in the first action, how many came back and involved in the second action". The first and second action can be the same or different events.
Define the Conversion Window - every day/week/month...
If you haven't yet created your events, check out this:
You set up a cohort of all that placed an order at the first time and came back and placed orders again, with a Conversion window of Every Month (Rolling 30 Days). If customers did Place first orders on day X, but they came back and placed orders 31 days after X, then these customers were only counted in the pool size but considered Drop-off at cohort because the window required was surpassed.
Conversion Window dictates whether or not an action is considered "converted". Only when the following events were performed within this window can they be counted in your cohort.
Adjust the number as you would like your cohort to measure. Default option is Every Month (Rolling 30 Days).
You can click on the Days menu to change the unit to Weeks, Months, Quarters or Years
Every cohort needs a subject to follow through all steps. Simply select the person or thing you want to anchor on.
By default, Presight funnel will count Total Events (or the number of times each events were performed)
You can choose other measure, like Unique Tracking Subjects, or other applicable metrics already existed in your workspace.
Even more, you can also dynamically aggregate something else from the table, like COUNT DISTINCT (User Phone Number)
By simply selecting 2 events you'd like to use to indicate a "return", you can immediately gauge how you retain your users, using our Cohort chart.
The first event is the anchor point of how you want to group your users (eg. Order for the First Time cohort, Account Creation year cohort,...)
The second event indicates what to count as "return" (eg. Order Placed)
The % over time will indicate how all the users in one cohort return to do the second event every period.
Note: what period to check for return highly depends on your business. For social apps, it might be daily; for B2b, it might be annually.
With Presight, you can also explore other aspect of the "return", like the dollar amount your users spend on each and every return. This is a highly valuable insight if you want to optimize for more spending every return.
You can extract Converted and Drop-off Items (Users, Orders, Products...) by right-clicking at the chosen time:
A records table will show up for the detailed information of the Items.