Skip to main content

Hi community,

Is there a way to create a funnel to see how many users had a certain event as their ‘first’ event in a period.

I.e. it is currently possible to exclude users who performed certain events between other events in the funnel (see screenshot). But is it possible to exclude users who performed certain events before the first event in the funnel.

 

Specifically, I want to know how many of our daily users opened a push notification without having opened our app previously that day - are my push notifications generating daily usage that otherwise would not have taken place, or is it only users that are already active that open my push notifications. 

In the screenshot, I want to see “opened push notifications” and then exclude all ‘start sessions’ performed before the ‘push notification’. 

 

Thank you!

 

Hey @Ingmar Oemig 

I feel it’s a bit tricky to get this in a funnel.

To get to your question - want to know how many of our daily users opened a push notification without having opened our app previously that day - are my push notifications generating daily usage that otherwise would not have taken place , I would try building an event segmentation chart like this -

https://analytics.amplitude.com/demo/chart/new/wdff8ty

I’m defining 2 segments here - users who Play/Search song in a given day, but not having done a Play from notification event within each day and users who Play/Search song in a given day and also who have played from notification. If I understand your use case correctly, this demo scenario might help you build your case.

Hope this helps.

Hi Saish,

Thank you for the help here! - The above could work, however, I guess I still do not know whether the users who performed Play song from push notification were already users on that given day? 

That is the reason I dreamt of doing a funnel, is there a way of identifying only the users that ‘played song from push notification’ without having previously played a song. 

I guess the only way I can go about it is somewhat bayesian: 

  • Make a funnel: 
    • Played a song => play from notification
    • Then I know how many users play from notification after having already played
  • Then I can safely assume that the remaining ‘play from notification’ did not play before

This does sadly though become quite manual as there is no way to put this into a chart.

Thank you! 

If you have access to the cohorts feature, you can try creating a cohort of users  who “performed played song from push notification” event and “not who performed Played Song” and then apply the same analysis.
The assumption here is that the Play From notification is fired mutually exclusively wrt the Play Song or Video i.e. only one of these events get triggered in their respective cases.

 

Reply


Terms & ConditionsCookie settings