Over the last several years, the Growth features have been deployed to 80% of new accounts (the treatment group) and have been withheld from 20% of new accounts (the control group). Keeping a control group enables us to analyze large experiments on the overall impact of the Growth features, but the downsides are (a) the valuable features are withheld from a fifth of newcomers, and (b) when some newcomers don't have the features it causes confusion for campaigns and events and for on-wiki discussion.
In T270786: Newcomer tasks: NEWTEA revisited, we completed a second round of analysis using this large experiment, and we now feel like we have sufficiently experimented with the overall impact of the Growth features. We want to remove the control groups for most Wikipedias, and in the future, experiment within the Growth features.
Specifically, all Wikipedias except the following should give Growth features to 100% of new accounts going forward. These Wikipedias should continue to have the control groups they currently have with no changes:
- Growth pilot wikis
- Arabic
- Czech
- Bengali
- Spanish
- Languages close to the pilot wikis:
- Portuguese
- Turkish
- Persian
- Wikipedias where the community tests the features:
- German
- English
See also: T296702: Scale: enable Growth features for existing accounts
See also: T292090: Research Spike: Enable Growth features for autocreated accounts (if the user already has Growth features)