is there any foreseeable use case for mixed cohorts? If not, it sounds like this would be a useful enhancement.


I await the prioritization gods to tell me what to do :)  Personally, I think dynamic cohorts and timeseries for all metrics might be more important.