This message was deleted.
# give-feedbackw
worried-planet-2191
10/28/2022, 9:27 PMThis message was deleted.
f
future-teacher-7046
10/28/2022, 9:29 PM
That's an interesting idea. We support metric "caps" right now where you can set a max value any single user can have. That basically stops huge outliers from screwing up results, but it still doesn't let you visualize the distribution.
🙌 1
future-teacher-7046
10/28/2022, 9:30 PM
We've talked about adding histograms to the metric overview pages, but not in the context of an experiment.
🙌 1
t
thankful-scientist-83252
10/28/2022, 9:30 PMWhat’s useful about the experiment context is it respects all the targeting condition
thankful-scientist-83252
10/28/2022, 9:32 PMthe metric caps are certainly useful but they force you to know/guess what would be an unreasonable value. Also a visualization would give you a sense of skew to answer the question of “is this driven by power users?” which is always a nagging thought for these count metrics.
139 Views