This analysis has been completed. We found no significant difference between the group of users who saw the Community Updates module and the control group. The primary reason for this is that the number of users who registered on Spanish Wikipedia during the experiment and signed up for a campaign is too low.

The Community Updates module experiment ran on four wikis: the Arabic, Czech, French, and Spanish Wikipedias. For the purposes of analysis, we consider 2024-11-01 (at midnight) as the deployment time and examine activity on those wikis during the month of November. During that time, two of the wikis ran updates that had limitations on who could see the update: the Arabic and French Wikipedia required at least 10 edits. On Czech and Spanish Wikipedia there was no limitation.

Another limiting factor in understanding the effect of the module is that the purpose of the updates were different from wiki to wiki. For example, on Spanish Wikipedia that Community Updates module linked to Wikipedia Asia Month 2024, while on Czech Wikipedia it linked to information about their weekly online wiki advice sessions.

The analysis was limited to the Spanish Wikipedia because of the time available to perform this analysis, the wiki's size, that there was no limit who could see the module, and that the module linked to a campaign that required users to sign up (meaning that we could track who signed up).

During the experiment period, the module was shown to about 3,500 users who signed up during the experiment, and about 1,000 users who had signed up previously. We limited our analysis to the former group as the Growth team focuses on the newcomer experience, and also because the impression and click-through counts for that group were substantially higher.

The per-user click-through rate of the module during the campaign was 1.7%. There was very little difference between the desktop and mobile web platforms when it comes to this rate. This rate is low but similar to baselines we gathered prior to the experiment for the "banner module", a previous feature on the Newcomer Homepage that worked similarly to the Community Updates module.

The sign-up rate for users after they clicked on the link in the Community Updates module, less than 5%, was substantially lower than baselines from other experiments. This led to an overall Campaign Sign-Up Rate, the key metric of the experiment that was also substantially lower than expected. When combined with the fact that the sign-up rate in the control group was also small but not zero, we end up with overall results that are not statistically significant.