Skip to main content

I’m trying to figure out how to set up a chart where I’m partitioning results according to whether users purchased X before, after, at the same time, or not at all in relation to the purchase of product Y. I was trying to use purchase of Y as my event definition, and then in the user specification on right, I was thinking I should use the offset feature to try to specify that the event there has to have occurred before the event in the event definition, but I’m not sure I’m understanding the documentation fully. 

Scenarios for users who purchased Y:

  1. purchased X then Y
  2. purchased Y then X
  3. purchased X and Y (same time)
  4. purchased Y but not X

If I wanted to see unique counts of users who purchased Y for all four of the groups listed above, how would I do that? Do I need to set up cohorts for this? And would the set up be the same if I was trying to apply this type of grouping to a funnel? 

 

I would recommend you set up several funnel charts to make these counts. (1) and (2) are typical 2 step funnels, (4) you’ll also be able to see on your chart for (2), and (3) should be possible by applying several conditions to a single step of a funnel (or indeed event segmentation).

The exact setup of the charts depends on how the event properties are set up for your purchase event. Feel free to screenshot and add them here if you need guidance.

Sure. I was figuring it’d be easiest to do using custom events that already have the properties baked in, but could do it with the event spec in the actual chart definition too. It actually isn’t important that event X is the same type of event as Y. My question isn’t around setting up the properties correctly, it’s more, how do I exclude all users who did event X before funnel Y→ Z from my chart? And then, how do I also make another chart where users did X before the funnel of Y → Z? (in both cases, I don’t want event X to be a visible step in the funnel)

Would something like this in the users specification do anything for me? Do I need an offset? Would this exclude users (here chat sent is event 😵 who did event X before events in my funnel definition, or would it exclude users who did events 1 day before a year before my funnel events? Or something else?

 

Hi @kaia - you can’t make this work using segmentation, because you can’t make conditions on event times of two events relative to each other, only relative to calendar days.

 

You can make your first funnel by using the “exclude” condition in the funnel chart. The second funnel you pretty much have to make X visible, there’s no good way around that.

Reply


Terms & ConditionsCookie settings