Solved

Retention Analysis Starting Event Question

  • 16 July 2021
  • 2 replies
  • 39 views

Badge

Hi, I have a (hopefully) simple question regarding setting up a retention analysis.

 

We have an event called ‘user_login’, where this event could trigger many times for the same user over their lifespan on the product, but the first time it triggers represents a new user. We’d like to use this event as our starting event for retention analysis, but I’m a little unclear about what happens on each subsequent trigger of ‘user_login’ for the same user...

 

Is it possible to setup the analysis so that it will only look at the firstuser_login’ trigger per user? Or any other tips on how we can set something like this up?

 

Thanks in advance!

icon

Best answer by MikkoKarvonen 19 July 2021, 14:49

Hi, I have a (hopefully) simple question regarding setting up a retention analysis.

 

We have an event called ‘user_login’, where this event could trigger many times for the same user over their lifespan on the product, but the first time it triggers represents a new user. We’d like to use this event as our starting event for retention analysis, but I’m a little unclear about what happens on each subsequent trigger of ‘user_login’ for the same user...

 

Is it possible to setup the analysis so that it will only look at the firstuser_login’ trigger per user? Or any other tips on how we can set something like this up?

 

Thanks in advance!


If you are on a Growth or Enterprise plan, you have a simple answer in the built-in Historical Count filter: https://help.amplitude.com/hc/en-us/articles/360040957432-Historical-count-Track-user-behavior-for-different-instances-of-each-user-action

 

If not, the best solution is to make sure your data is set up in a way that allows you to answer the questions you want to be answered. In this specific case, I would either create a separate event for the sign-ups (which I assume user login for a new user is) or, if you have some reason to avoid this, add an event property to user_login event that allows you to differentiate these cases. I’d recommend a separate event, as your analysis and ability to fully utilize Amplitude features will benefit from having dedicated events to all your key actions (which sign-up definitely is), but the event property approach will also allow you to answer this quetsion.

View original

2 replies

Userlevel 4
Badge +2

Hi, I have a (hopefully) simple question regarding setting up a retention analysis.

 

We have an event called ‘user_login’, where this event could trigger many times for the same user over their lifespan on the product, but the first time it triggers represents a new user. We’d like to use this event as our starting event for retention analysis, but I’m a little unclear about what happens on each subsequent trigger of ‘user_login’ for the same user...

 

Is it possible to setup the analysis so that it will only look at the firstuser_login’ trigger per user? Or any other tips on how we can set something like this up?

 

Thanks in advance!


If you are on a Growth or Enterprise plan, you have a simple answer in the built-in Historical Count filter: https://help.amplitude.com/hc/en-us/articles/360040957432-Historical-count-Track-user-behavior-for-different-instances-of-each-user-action

 

If not, the best solution is to make sure your data is set up in a way that allows you to answer the questions you want to be answered. In this specific case, I would either create a separate event for the sign-ups (which I assume user login for a new user is) or, if you have some reason to avoid this, add an event property to user_login event that allows you to differentiate these cases. I’d recommend a separate event, as your analysis and ability to fully utilize Amplitude features will benefit from having dedicated events to all your key actions (which sign-up definitely is), but the event property approach will also allow you to answer this quetsion.

Badge

Ah perfect - I’m on the Growth plan, so will go with the Historical Count solution! Thanks so much!

Reply