I have a question around metrics and possibility for 'templating' metrics. Say I have a set of high level metrics like Bookings/Gross Revenue/Net Revenue that can also be filtered down by an attribute like city. I'd like to create a templated metric like 'Bookings {{ city }}' That I can then define when I add the metric to an experiment vs defining up front. That way if I want to run an experiment across 10 cities I don't have to create the Bookings metric 10 times filtered for each city (and 10 times for every other metric). I don't think that's possible currently but I'd like to either know I'm wrong and it is possible or that I'm thinking about the problem wrong and there is a better way to do this

you can use fact tables to get very close to this

Hmm in this scenario don't we still end up with a separate metric for every thing we want to filter by

I see

I guess really what I would like is for there to be a templated/filtered metric that gets configured with the template variable/filter at the time it is added to an experiment but otherwise doesn't pollute my list of overall metrics. I don't mind having a a 'Bookings {{cityFilter}}' metric in my list of metrics but having 50 of them along with N other metrics that need 50 versions of that metric seems unwieldy. Some part of this discussion is a theoretical problem not an actual one for us right now. We are running 4 experiments right now that fit this description and may end up tracking 6-7 metrics for each so its not a huge number but its highly likely we will run similar experiments in the future. I think our plan right now is to run the experiments and then archive the filtered metrics so they don't crowd our metric list.