Do they get those numbers via telemetry or from the server-side? If the former, those stats may be skewed by the overlap between the users who use several extensions and those who disable all telemetry.
In a mozilla bug report where they discuss removing user.js, they point to telemetry that indicates that no one uses this functionality. I'd argue that the Venn diagram between users of user.js and those who disable telemetry approximates a single circle.