Please see attached. This is a stripped-down excerpt of a data set with over 800k rows. I have tried EVERYTHING. All kinds of Countifs, Sumifs, Sumproducts, even some DCountas. I consider myself an intermediate to advanced user, and I can't figure this out at all. The main snag is calculating the original cohort: it needs to be only those users that transacted 2 or more times with "dog" BEFORE a certain date, and those users are only counted once regardless of the number of transactions (so if a user had 8 transactions of dog before the date, he still counts as one unique user). I'm at my wit's end, don't really know what to do from here. Any guidance much appreciated.
Bookmarks