tfishell: I don't know how hard this would be to gather accurately, but I'd really like to know how many more people use Galaxy over installers (if Galaxy is used by far more people).
I agree though even if the real "stats" were found & posted, there would still be massive sampling bias in favour of Galaxy simply because Galaxy's telemetry is counted whilst offline installers aren't due to their nature. Example - let's say two GOG users buy the same game in 2014 (an old game that doesn't need updating much):-
Person A downloads the offline installer and backs it up then replays it annually by locally reinstalling it offline from a backup drive. Those 7x replays will all be counted as "1x offline installer download" by GOG's server / account usage telemetry. By 2030, he'd have 17x annual replays yet all will still only be counted as "1x offline download".
Person B replays it annually Galaxy / Steam-style by Download -> Play -> Delete -> Re-download next year and will appear to GOG server statistics as "7x Galaxy downloads" by 2020 and "17x Galaxy downloads" by 2030 (for one user for one game)
ie, the Officially Correct (tm) download statistics will show
"87.5% of downloads were via Galaxy vs 12.5% of Offline Installers" in 2020 and
"94.1% Galaxy v 5.9% Offline Installer" by 2030, when counting just 2x people downloading just one game, who between the two of them still have 50/50 "I use Galaxy" vs "I don't use Galaxy" usage patterns.
Same goes for those of us running a "whitelisted" firewall, whatever telemetry pings are sent out by the offline installers saying "mark +1 install for this game" (plus any in-game telemetry) won't be counted if it's blocked by a firewall (whilst Galaxy users always will). So I don't even think it's possible for GOG to measure it "honestly" without obvious sampling bias against offline installers due to their very nature and purpose of not needing the Internet = naturally lesser visibility vs those who use Galaxy like Steam (always online for each game install, constant re-downloads, etc).
Edit: Having completed that survey, I ran into the same issue where I think it was Q27 that said
"Do you use Galaxy", and if you ticked "No", it still took you to Q28 (asking the 3 things you like most about using Galaxy) instead of skipping over it to Q29. So it's not hard to claim even 100% rates of Galaxy usage based on the survey that refuses to let non-Galaxy users complete it unless they lie and say they use Galaxy... ;-)