Prepare the V2 A/B experiment for the improved mobile account creation form, building on the existing V1 experience.
Background:
The V1 experiment for improving the mobile account creation form has shown positive results, with increased account creation. To maintain momentum and accelerate learning, we will proceed directly to a V2 experiment that builds on V1 improvements.
Instead of pausing to merge V1 into core and deploy it globally, V2 will incorporate those changes and introduce additional enhancements. This approach allows us to iterate more quickly while continuing to validate impact through experimentation.
User stories:
As a mobile web user, I want a cleaner and less verbose account creation form so that I can view all required fields at a glance without scrolling and complete registration with less friction.
As a mobile web user, I want simple and clear guidance when creating a username so that I am more likely to follow the username policy.
Proposed approach:
- Use the current production experience as the control
- Use V2 (which includes all V1 improvements) as the treatment
- Launch as an A/B test on mobile web
- Continue measuring account creation completion and related metrics
- Included changes (V2 scope):
- All V1 work: T419916: [V1 experiment release] Redesign mobile web account creation form following Codex guidelines
- All V2 work:
Rationale for approach:
- Avoids delay from communication & scaling V1
- Allows us to build upon and improve from what we learned in V1 (reduce the impact of the decrease in username policy CTR)
The downside of this approach is that if we find that the V2 version is NOT leading in a positive change in account creation rates, then we will need to revert back to V1 and scale that work instead.
Acceptance criteria:
- V2 Treatment variant includes all V1 changes plus listed enhancements
- The V2 Account Creation experiment is ready to release, which is covered by: T422295: [V2 experiment release] Mobile web account creation form improvements + username TL;DR
- The V2 experience should only show for the following criteria: DB name is enwiki, skin is Minerva and assigned group is treatment
