Is there a way to reference the timestamp that the user gets bucketed into the experiment in a metric’s SQL?
I’ve seen one can include the start/end date of the experiment with SQL template variables, though this is obviously not the same thing…
f
fresh-football-47124
05/22/2023, 10:41 AM
Hi Callum, what is the use case for this?
h
hallowed-lifeguard-32533
05/22/2023, 10:48 AM
general use case: I want to create calculation denominators that help to exclude certain users from metrics depending on their status as at the point of their bucketing. And i need to do this at a metric level, so applying segments at an experiment level is not helpful to me
hallowed-lifeguard-32533
05/22/2023, 10:49 AM
e.g. i want a denominator that only includes users whose 7th day post registration falls between their bucketing, and the current timestamp (i.e. the correct way to calculate day 7 conversion)
hallowed-lifeguard-32533
05/22/2023, 10:51 AM
moreover, in the actual D7 conversion conversion metric, in order to not exclude users who, say, convert in their first hour, I need the denominator to return a timestamp that is guaranteed to fall between their bucketing timestamp, and the earliest point that they could convert (so something like, their bucketing timestamp + 1 second)
f
future-teacher-7046
05/22/2023, 12:10 PM
To make sure I understand, for the denominator, you only want to include users who are in the experiment during their 7th post-registration day. And you want the numerator to only include metric values from those users. Is that right?