The number of total activations is not enough to make informed decisions on whether a beta feature is ready to be graduated (do users think it is ready? maybe they just like the idea, or want to follow the evolution because they don't like it at all).
We could support a rating system to indicate whether a given user (a) found the feature useful, (b) liked the idea but thinks it is not ready yet, or (c) does not find the feature useful.
This can be a simple drop-down next to each beta feature with the above options (plus a neutral "Evaluate this feature" state).
Some related ideas:
- We can encourage (e.g., with notifications) users to re-evaluate a beta feature after major change, and get an idea on the effect of those changes in user perception.
- We can invite those users providing a negative rating to contribute to the talk page to add further details.
- We can do further analysis on a given group of users (are those who don't like the feature new editors?) and identify solutions for them.