Page MenuHomePhabricator

Variant tests: C vs. D analysis followup: editing at registration
Closed, DeclinedPublic


The results of the Variant C/D experiment (T264831) are divergent. Two questions that keep coming up in our analysis are: to what extent do the Growth features fit into the user's workflow, and can we predict activation with higher accuracy? This task is for some followup work from both the variant test as well as analysis of Growth features such as the impact of the Newcomer Tasks. We're interested in understanding if we can determine if users were editing when they signed up, and if we can, what impact that has on our analyses.