Hi. Some time ago we asked you to fill in a questionnaire about the DAWs you used. Well, we have finally published the results, and they are very interesting! As promised at the time I leave you an overview here, but at https://www.lim.di.unimi.it/dawsurvey/results/2023/ you can find many more juicy details (I recommend reading the paper and downloading this PDF poster)
P.S. If for some reason the results trigger you: have mercy, I am only an ambassador! (read the disclaimer at the bottom of the poster)
Iām really surprised that Bitwig scores so low in these results.
Granted there are only 613 datapoints and their origin is limited.
I know a lot of people on different OSes that have it as their preferred DAW, but this is my own anecdotal evidence
Imo having Audacity in there was a bit misleading as right now it doesnāt have any DAW capabilities (I would argue that non-destructive editing is a prime feature of DAWs), but in the future they intend to release it as a true DAW (still in development).
Still somewhat interesting results. Thnx for sharing!
Hope we can see this kind of analysis from a much wider and varied audience (say 10-100x as many respondents).
The bars indicate numerically that they sum to 100% but the scaling on them is completely off.
Why is ableton(job) (40%) bigger than Reaper(job) (50%)
Fl studio (Hobby) at 80% smaller than Logic (hobby) at only 52%
Besides the horrific display of information (sorry but its bad), im also a little concerned about the demographics the data came from.
The majority of male and italian respondants has absolutely had to have skewed the results. You mention FL Studio being more adopted in italy, but considering more than half of your data (assuming the pie chart is right) came from italy, doesnt that mean that any majority in italy would be an overall majority?? (that point might not be as robust but the concern still shows)
Now, do i intend to be mean? no. This however is statistics, its science. It looks like the thought process for display of this information was at best rushed and at worst not even considered.
I would love to give a shot at making better graphs for this if possible. Is the data also available?
EDIT:
Had a little read into the disclaimer, fair enough mentioning the flaw of having more italian data. Maybe just make it more prominent firstā¦
Iāve also downloaded the data and plan on figuring out why the graphs came out like thatā¦
Yeah, that data doesnāt align with our results at all. We get our data from all our customers all over the world, we have way more data (tens of thousands of datapoints) and itās real world usage, not self reported stuff.
Thanks for the comment, I agree there is still margin for improvements, let me explain the rationale about this graph: The inline percentages are relative to the specific daw user base, the absolute number of users is the total width of the bars, so, eg, in the case of hobby FL vs hobby Logic, Logic wins in terms of absolute number of users. Plain olā bar graphs are available at the page I linked.
ok, thats definitely super confusing. Having the bar graphs available is nice at least.
I think a better display might have been vertical bars for each DAW, split by colour based on the portion used for work / hobby. Right now its not obvious that those numbers are relative to the DAWās userbase at all.
I think that once you learn to read it itās pretty clear: the overall width of the bar of a daw tells you how many users it has, and the horizontal offset tells you if its users are more professionals or hobbyists, but yeah, at a first glance, a normal barplot is better
Thatās interesting. Can I see your data? Where are they published?
Anyway, a big portion of the poster is about sample composition, which suggests that the results should be taken with a pinch of salt. As a scientist, Iām used to that, but I can see that a more prominent warning should have been put in place for the general public. Sorry, itās our fault.
Thanks for sharing, and for the effort you have put into presenting this info.
The Italian bias makes it a little less informative to the rest of us but Iām very grateful to see it all the same!
Iāve posted this same data around a year ago. Nothing has really changed since.
Of course, weāre only one company, and our customers are mainly younger people producing EDM and adjacent genres.
Top result always fills the bar to 100%; the others are relative to that. Absolute percentages are also given.
This is from real-world data, from all over the world. Nothing customers can modify or lie about.
The āarchā section is macOS specific. It tells us that almost 80% of our Mac users own an Apple Silicon machine (71.8% Arm + 7.66% Rosetta), and only 20.5% are on Intel.
This suggests that within 3-4 years, we can likely discontinue the Intel build and ship smaller apps/plugins.
I can now understand entirely the mistrust the general public has in science, given that making a post on Facebook in Italy, asking people about their preferences, now counts as publishable science. Reminds me of another, but very similar problem: the Gell-Mann Amnesia effect.
You read or watch a news story about something you know well, and you see that itās completely wrong ā full of errors, misunderstandings, and oversimplifications.
But then, you turn the page or change the channel, and you immediately trust the next story, even though itās on a topic you donāt know well.
āThey always get it wrongā¦ except when I donāt know better.ā
OK, this is valuable information, but itās very different from what we were interested in. This is a very specific subset of DAWs used as hosts for Nexus. Youāll agree that calling it representative of the overall DAW distribution is a bit of a stretch. As you said, this data essentially excludes all non-electronic music makers (and all electronic music makers who are not your customers).
Anyway, I agree that it would be worrying if a Facebook post were considered science. Fortunately, thatās not what we did. Itās a bit sad that you assumed that so easily.
If you ask me if Iām happy with the data we got, I would say no. I had hoped for greater penetration of the forum category (e.g. KVR, Gearspace and SoS), which would have helped in terms of achieving a more global variety. Unfortunately, we were not successful there, but we plan to repeat the survey in the future. Nevertheless, we gained something from our campaign. For the sake of science, we have decided to publish the results, clearly stating all the limitations and providing the raw data to anyone who wants to conduct their own analysis. This is what the above poster is about. I donāt understand all this animosity.
I donāt understand all this animosity.
It is usual into this forum nowaday (for about anything).
I think its easy to fall into the wannabe-rockstar mindset in the synth/audio/plugin world, which automatically elevates folks entitlement levels when it comes to cutting anything down that might impact their own personal PR .. but it might also be that mix of āold, schooled, wise and seasoned greybeards being frustrated with all the dumb questions of newbiesā factor ā¦ Would be interesting to understand the JUCE dev demographic a bit better.
