Hi there, I would like to draw the attention of the team to the issue described in https://growthbookusers.slack.com/archives/C01T6Q1SVFV/p1641828963004700 & https://github.com/growthbook/growthbook/issues/405. Our use case of Growthbook implies that a user can see an experimental version of the feed multiple times (each time with new content) & convert at any time with attribution not only to the first exposure but to the experimental version of the feed in general. There are two possible ways to address this problem: either set a nearly-infinite conversion window, or enable attribution to cohorts for further exposure cases (not just the first one for a specific experiment). However, if we use the former, conversions happening after the experiment is stopped still contribute to the statistics. It would be really nice to have the latter mechanism as an option. What is the opinion of the team? Do you plan to implement that?

Our current thinking is outlined in that GitHub issue you linked. Would that solve your use case or do you think something else is needed?

I believe that options 2 and 3 mentioned in the GitHub issue would be good ways of handling our use case.