Is cohort metrics possible in Adobe Analytics? | Community
Skip to main content
R
Adobe Employee
RoyyyyyAdobe Employee
Adobe Employee
April 6, 2023
Solved

Is cohort metrics possible in Adobe Analytics?

  • April 6, 2023
  • 2 replies
  • 1354 views

For example, if I want to calculate W1 2+ visit in adobe analytics, how can I build the metric to show that? My understanding is since AA filters the event with a time window, there's not way we can get fully baked data for these cohort metrics.

 

Thanks!

This post is no longer active and is closed to new replies. Need help? Start a new post to ask your question.
Best answer by Jennifer_Dungan

There is a specific Cohort Visualization available to use in Workspace:

 

 

This will do all the calculations for you. You just need to ensure that your panel is selected to an appropriate time range... since you are doing "Weekly" visits, I would suggest a "Last x Weeks" date range.

 

 

You can also enable "rolling calculation" which will mean that users must have visited in each concurrent week to be included... or don't use that setting, and users who may have skipped a week will still be included in the weeks they did have more than 2 visits (out of your "included" column of course)

2 replies

Krishna_Musku
Community Advisor
Krishna_MuskuCommunity Advisor
Community Advisor
April 6, 2023

@royyyyy If I understood your question correctly, you want to track the 2+ visits in AA right? In the cohort table viz, you can calculate 2+ visit in return criteria.

R
Adobe Employee
RoyyyyyAdobe EmployeeAuthor
Adobe Employee
April 18, 2023

Hi Krishna, 

 

Yes but it's kinda hard to use cohort table for communication purpose, e.g. create trend chart like when using metrics. Is this the only way we can calculate retention metrics?

 

Another questions is if I want to calculate things like % users did X in their first 24 hours, is that also not possible?

 

Thanks!

Jennifer_Dungan
Community Advisor and Adobe Champion
Jennifer_DunganCommunity Advisor and Adobe ChampionAccepted solution
Community Advisor and Adobe Champion
April 7, 2023

There is a specific Cohort Visualization available to use in Workspace:

 

 

This will do all the calculations for you. You just need to ensure that your panel is selected to an appropriate time range... since you are doing "Weekly" visits, I would suggest a "Last x Weeks" date range.

 

 

You can also enable "rolling calculation" which will mean that users must have visited in each concurrent week to be included... or don't use that setting, and users who may have skipped a week will still be included in the weeks they did have more than 2 visits (out of your "included" column of course)

R
Adobe Employee
RoyyyyyAdobe EmployeeAuthor
Adobe Employee
April 10, 2023

Hi Jennifer, thanks for the reply!

 

Yes I'm aware of this cohort table, but it's hard for communication purpose, e.g. create trend chart like other metrics can do (or maybe I don't know how to?). Is this the only way we can calculate retention metrics?

 

Another questions is if I want to calculate things like % users did X in their first 24 hours, is that also not possible?

 

Thanks!!

Jennifer_Dungan
Community Advisor and Adobe Champion
Jennifer_DunganCommunity Advisor and Adobe Champion
Community Advisor and Adobe Champion
April 18, 2023

Well you could try right clicking on some of the cohort cells and "create segment from cell"... but they will be initially locked to the week of that cell (you could probably replace that locked date with rolling date ranges - i.e. Last Week, 2 Weeks Ago, 3 Weeks Ago, etc)

 

When I tested it, this is the Segment that Adobe built... you can see it's locked to the Week of March 19th, but as I said, you could try replacing that with an "X Weeks Ago" date range

 

 

But as for "% users did X in their first 24 hours"... do you mean their first ever 24 hours? Or the first 24 hours of your cohort analysis? Again, if you are using defined date ranges, you could probably use that to identify actions performed at the start of the cohort.....

 

This is not really a standard cohort analysis, so I have never built anything like this.. I think this is something you are going to have to experiment with and see if the results look reasonable...

Terms & ConditionsAccessibility statement

Loading footer...