[dimtion]

I would be very surprised that the average number of steps across their userbase has uniformly decreased for all users.

For example we could imagine the scenario where some people were able to keep their habit and did not have a great decrease in step numbers, while some other sub-groups have become sedentary because of the confinement. This would show up as an increase in variance in the number of steps.

Stats are hard and it is very easy to draw a wrong conclusion (that sound right) from aggregated data.