https://www.growthbook.io/ logo
#ask-questions
Title
# ask-questions
a

adventurous-agency-77764

02/29/2024, 6:59 AM
Hello! I’m exploring the options for conversion windows and slightly confused with the documentation. Did you by chance invert the labels for None and Conversion Window in this graphical representation? Based on the surrounding text, I’d expect that None would have all rectangles blue (as it is using all data) instead of 3/4 greyed out, but wanted to check I wasn’t missing something. Thank you!
Separately, we’re running experiments for a subscription based app, so while the action (signing up for a trial, for example) would come before the experiment end, revenue (the goal metric) might come after the experiment end. Is there a way to include data that comes after the experiment end?
f

fresh-football-47124

02/29/2024, 3:38 PM
@helpful-application-7107 thoughts?
h

helpful-application-7107

02/29/2024, 3:49 PM
Oh shoot! Yes you're right. I'll fix this today. Sorry for the confusion.
As for your other question, there isn't a way we recommend to include data after an experiment ends right in the main results. However, you can create a Custom Report and manually change the end date to go beyond the actual end date to include those metric data. Does that make sense?
a

adventurous-agency-77764

02/29/2024, 3:56 PM
Thanks for both answers @helpful-application-7107!! Really appreciate it! For my second question - I think I can also adjust this by using the date that the user signed up for the trial as the date of the revenue if revenue is generated, but this does feel like a bit of a hack. So if you ever have time, I think a feature that allowed the conversion window to extend past the experiment end could be super helpful for subscription based businesses like ours! 🙌
f

fresh-football-47124

02/29/2024, 4:07 PM
@helpful-application-7107 - can you separate the experiment end date as defined in the feature rule, vs the end of the experiment report?
h

helpful-application-7107

02/29/2024, 4:16 PM
I wouldn't recommend that and I'm not sure it's always possible via the UI.
I actually just checked, and I think conversion windows should extend beyond the experiment date. So if you set a conversion window that is very long on your revenue metric, it should keep counting until each user has completed their conversion window, even after an experiment end. I will confirm later when I update the graphic.
a

adventurous-agency-77764

02/29/2024, 4:43 PM
Ah that would be great! Looking forward to hearing from you! Thanks!!
Realized I could check this myself @helpful-application-7107 , so I just compared the underlying queries for an experiment for a metric that has a 7 day conversion window set vs. a metric that does not have any conversion window set. And, I do see 7 days added to the end date in the metric CTE of the metric that should have the conversion window (and no days added to the metric without it). That’s everything I needed, thank you!!
Would be super helpful to see the blue squares extend past the end date for “conversion window” in the visual when you remake it!
👍 1
Thanks for your help and sorry if I added any confusion!
h

helpful-application-7107

02/29/2024, 4:55 PM
I think I'll make a second graphic to demonstrate how conversion windows vary by user and can extend beyond the experiment end date.
🙌 1
No worries Alice! I also really value people pointing out mistakes like the labels in our graphic. Makes everyone's experience better.
🙌 1
8 Views