I have several goals that are “once per day” goals (i.e., not “daily”, per se, but I can’t count them more than once per day, and my commitments are measured in “daily completions”).
I’ve been doing this manually (by just being diligent to only enter one data point per day, at most), but I’ve recently been clued into custom aggregation methods, so I’m wondering… is there a built-in “only include one datapoint per day in the cumulative number” aggregation method?
Ah, yep! Really useful to hear you had trouble finding info about this – do you happen to remember anywhere you tried to look for it, e.g. search terms you used in the help docs, etc? It’s always useful for us to hear when something’s a bit too difficult to find, and it’s also possible for me to add more search terms if there’s a particular search term that someone tries that doesn’t pull up the relevant info, etc.
You can find all the info about various different aggregation methods in the Custom goals help doc. The whole list is there, but you’d be thinking about probably nonzero or binary, for this use-case, or possibly cap1.
(Skatesum would also be an option, but I tend to avoid using it, as it doesn’t react well if you want to change the daily rate. That said, nonzero, binary or cap1 would also not work if you ever want to change the rate, since they’d automatically derail on anything with a rate more than 1/day. I think they basically would all work in this case; my personal preference would probably be binary.)
What about “max” as the aggregation method? Then if there’s more than one datapoint for the day, all but one (the largest) will be ignored.
Thank you, Nicky!
I simply clicked on the “aka aggday” link next to the aggregation method in one of my goals, and started looking at the list of possibilities, and my eyes glazed over. 
I was also very tired yesterday afternoon, and now that you’ve clued me in, I can see how those aggregation methods you mentioned would work for me.
I tried the “binary” method on one of my goals and then realized that it counts derail datapoints as progress (which isn’t quite what I want), so I switched it to nonzero and that seems to work well enough!
Thank you again!
I think that would also work for me! Thanks!
