This wiki page has a nice collection of stats on editor popularity. The data is up to date, but the graphs aren't. I'm not a big fan of the logarithmic scale either.
So here's one graph to tell the main story.
I focused on "the big editors" to keep the graph simple. If you want more detail, just head over to the wiki page.
You can read the graph horizontally, showing first the distribution of changesets, then number of unique contributors, then total edits. On the left, market share. On the right absolute numbers.
There's some very clear patterns there. I really like how you can connect the dots for contributors of the "default editor" at the time: first Potlatch, then Potlatch2, then iD. All three of them reached 80% market share at their peak. But iD went down in relative terms because of Maps.me. That could only happen if Maps.me editors don't use iD much. That's a good thing, as it show they are new mappers. And it's a bad thing, as it shows that we haven't (yet) succeeded in getting them more deeply involved in OSM.
To make some of that more clear, here's three more charts. Changesets per contributor show that JOSM users are quite productive. There's also a very clear growth path for JOSM users. Merkaartor has a similar pattern. Maps.me hardly shows, with just 4 changesets per contributor.
Some changesets are bigger than other. JOSM changesets are the biggest. Potlatch2 are somewhere in the middel, and iD changesets are quite small. The average Maps.me changeset has only 2 changes.
So what's the overall productivity of contributors? Here JOSM is quite extreme.
Note that this doesn't say anything about quality or amount of work. For example a JOSM changeset editing thousands of objects could have been made in minutes. Someone could have surveyed a day to collect ten POIs and map them with iD.
As one of the few remaining Potlatch users, I had to make this graph too:
As Potlatch2 lost the status of default editor, the remaining users became ever more productive. That makes sense, because "low engagement" contributors won't find the way to that editor. So the only relevant numbers are those for 2011 and 2012. And compared to that, the low numbers for iD are striking. Low numbers may mean that more people with less motivation can be pushed to make at least one edit, so you can call that a success. This is the argument to call Maps.me a editing a huge success. But it can also mean that the editor isn't as inviting to work on more stuff than just on the thing you wanted to do. Anyway, a much deeper analysis would be necessary to draw any conclusions on that. You'd have to take account of previous mapping experience, later shifts to JOSM, and possible differences between 2011 and 2016 newbies, to name just a few controls. Also: the numbers are rising every year, even as it remains the editor for new contributors.
And then there's the good old Potlatch 1 of course. There's only one reason to open that ugly duckling: go to a place where you think something was deleted, press U, and you can see and recover it. It is amazing that no other editor has a similar feature that makes this so simple.
You can download the cleaned up data here (dropbox).