==== Background
Recently, we created the magic word `{{#mentor}}`, which returns a current mentor assigned to a given user. However, since at most wikis, our features run in an A/B testing (set to 80/20 right now), there will be some new users without any mentor at all. The magic word will return an empty string for those users.
To make the magic word an easy way for communities to manage mentorship, independently on whether an user has the features enable, we need to provide a solution for the 20 % control group.
One of the solutions that comes to my mind is to assign mentors to 100% of new users, regardless on whether they will have mentorship features. That way, the `{{#mentor}}` magic word will always return something.
==== Proposal
Assign a mentor to 100% of new users, so `{{#mentor}}` magic word works for all new users, irrespective of which group (experiment/control) they are in.
==== Potential issues
I am sure this will not be an issue in terms of storage space. The row in `user_properties` table is tiny, and can't hurt at all.
I'm however unsure whether this will cause any issues when analyzing the features. @nettrom_WMF Can you check whether assigning a mentor to users outside of the experiment is okay? If it's not, we'll have to figure out another solution how to help communities to serve the 20 % of control group.