Acknowledged

Export All Events per User in a Specific Time Frame

  • 14 September 2021
  • 2 replies
  • 23 views

Userlevel 4

Hi everyone,

We would like to have a similar option to exporting all users’ attributes from Audience Explorer in PX - i.e. we would like to be able to export all the events (and the number of times they were triggered) per each user. I can imagine getting a table in which:

  • the first row would contain names of the (triggered) events
  • the first column would contain UserIds
  • in the rest of the table we would have numbers describing how many times User X triggered the Event Y in the specified time period

I would also suggest to implement it in a way that PX users (we) need to specify the exact module containing the events which are of interest - in order to allow us to choose the events, but at the same time to limit the number of events. In that way, you could probably avoid generating huge calculations, exports or having processing issues.

Example of a basic use case:

  • We created two cohorts of users - Users in Value Discovery phase and Users in Habit Formation phase
  • Now we want to compare which are the events that the second cohort did and the first didn’t do
  • We also want to see the differences in the amount of times that certain events were triggered by these two cohorts
  • The above two points could be done (for example) in excel, by extracting the previously described tables from PX
  • Finally, we want to hypothesize which actions (events) lead to passing from Value Discovery stage to Habit Formation stage - we’ll try to engage the users to perform these actions

Note that this use case is applicable to any two cohorts which are in stages of user journey which follow one another. Also note that I’m aware that, in theory, something similar could be done with Query Builder - but only in the case that you know the exact 3 events in which you are interested in, i.e. the exact 3 events which are making the difference. And this is basically impossible to know apriori in most cases. While we are on that topic, I’ll use the opportunity to ask about the status of the following feature request:

 

And, finally, I’d also take the opportunity to emphasize once again that it is essential to make the creation of cohorts easier in PX - it’s actually a prerequisite for this feature request, but for other important things too. One of the ways in which cohort creation could become easier is by adding a feature like described here:

 

Thank you in advance for your replies!


2 replies

Userlevel 4

I’ll add another use case here which I find very relevant and applicable for in-product analytics. This time I would like to:

  • Build a machine learning clustering algorithm (outside of PX of course - for example in R or Python)
  • Insert the data about users of a certain product and their activity inside this product (so, data which we would like to get from PX)
  • Train the algorithm on this data
  • Get the result in terms of clusters of users - which will represent different behavioral personas

In the current setup, we can’t do it with the help of Gainsight PX - because we simply don’t have a way of getting/exporting these data.

Userlevel 5

Thanks @dorian_cudina for the usecase details. We are planning on Datascience to PX as on of our key strategic investments.

Reply