Last year I put forth the idea that commercial, closed source software could not only succeed on Linux Desktops but, in fact, be directly beneficial to the Desktop Linux world as a whole.
Back in February of this year I started by talking about the direction I wanted to go (as an independent “shareware” type developer) and followed by releasing the first of three Linux desktop applications (RadicalCodex, Radical Comic Designer and DoThisNow).
Since then I’ve been carefully tracking the results. The downloads. The page views. The sales. The support load. Where the press attention is. Etc.
The goal being (if this little experiment of mine succeeded) : To have enough data to construct a detailed case for how some of the larger, and more prominent, software vendors could feel confident that they would be able to recoup initial investments, and make a profit, by selling software to Linux desktop users.
Now, 3 1/2 months later, I’d like to take a few minutes to talk about the results.
Let’s get the big question out of the way right up front:
Is it possible to make a living as a “mom and pop” type software (shareware) company focusing on consumer oriented Linux Desktop software?
Yes. But only just barely. And it is considerably harder than it is for other platforms.
I’m still working on pulling together the full numbers in a way that’s digestible (and worthwhile), but here’s three pieces of data that I find incredibly interesting:
- The conversion ratio (ie “How many times does the application get downloaded for every sale?”) is off the charts good. Looks like the average was 60:1. (For every 60 downloads… one sale.)
- Download numbers of Linux applications from RadicalBreeze.com have been almost exactly the same as Mac downloads for the same application (RadicalCodex). Windows downloads have been roughly three times higher than Mac or Linux.
- Ubuntu (32bit and 64bit combined) accounted for roughly 75 - 80% of all Linux downloads for RadicalCodex each month. The non-distro-specific tar.gz archive accounted for roughly 10%, followed (very closely) by the OpenSuse package. Fedora barely even has been a blip on the radar.
A few things worth noting when looking through the results:
- I spent almost no time promoting the Mac or Windows versions of these applications (outside of one or two press releases).
- I only made packages available for Ubuntu, Fedora and OpenSuse.
- At various points there was a fair amount of press (both negative and positive) about the Linux versions.
These results left me with a few conclusions:
- There is a segment of Linux users that, almost desperately, want to support commercial software endeavors.
- It is incredibly difficult to get the word out about commercial Linux software. Most Linux news sites don’t typically carry such announcements. And most Linux users are not accustomed to searching “download sites” for new software (for many good reasons).
- Unless that (item 2) is changed somehow, selling consumer oriented, commercial software on Linux is just not going to work long term or large-scale.
- From a software distribution perspective, there is very little reason to spend time supporting non-Ubuntu distros (unless it takes very little investment).
I am somewhat bummed.
A direct application of the existing “shareware” model to the Linux marketplace is, technically, doable right now. But not overly worth while.
Which makes it difficult to show a strong business case for larger commercial efforts (outside of niches such as virtualization, etc.).
I believe it is entirely possible that this will change over the coming years. That it will be attractive and profitable for companies (large and small) to bring their software to the Linux Desktop. But, for this very moment, it’s just not happening.
So. Where now then?
Luckily there are additional revenue models that have been proposed by many people. Donations. Paid support. Free with paid “plugins”. Etc.
And, being as the Radical Breeze applications are already in a good state…
I’ll shortly be announcing a change…
(If you have any questions on this topic for me, just ask in the comments below. I’ll do my best to answer everything I can.)
May 26th, 2009 - 9:11 pm
You need something big for users to focus around and look at, like an app store, but in Linux’s own unique way (an app repository with a trustable payment method?).
Debian packages are clearly the way to go, with the number of Ubuntu users out there. Other distro users can at least pick up the savvy to pick up a debian unpacking tool (or in the parlance of Gentoo, have a “fetch restricted” ebuild).
I really don’t know if commercial software on Linux is viable. There is a THRIVING Floss ecosystem for *some* types of apps, and I don’t know if Linux users (including myself) are capable of making the leap to give credit card info for other types of apps. It’s a really tough bridge to cross. But make it easy, and make it commonplace, and make it clear that it’s a big push that big software vendors are behind, and MAYBE you’ll drag some of us kicking and screaming into a more modern commercial Linux era.
May 26th, 2009 - 9:33 pm
I would like to support commercial software on Linux, but to be honest I don’t believe there’s much chance anyone will get anywhere on this endeavour. I am ready to jump the ship, as soon as something more viable than Linux comes along. Can someone please make a MacOS that is not evil as heck? I would run it.
I love Linux, but I sort of hate it too, you know? I used to be a complete freetard (now I am using that word too, not awesome at all), but I just want shit to work sometimes.
It’s sad in a way, that PC-BSD didn’t come along before Linux, becasue their model seems to be a bit more balanced.
Another thing Bryan: I think you charged too much for your software. I know it might be worth it, but you’ll get a lot more sales with really low prices. Valve experimented with cutting prices in half and had an increase in sales that was in the ranges of 200-300%.
May 26th, 2009 - 10:00 pm
Excellent post (and unusually good, well thought out comments as well).
Do you think if you had maybe 20 apps already done that you simply maintain would have enough revenue flowing in to continue?
May 26th, 2009 - 10:33 pm
Bryan, in your opinion, do you think a “Steam/Impulse/Click-and-run”-like centralized service for commercial applications would have any success at reaching a mass consumer market on Linux, given the current user culture? In the feedback you have received during the experiment, how would you rate the willingness to pay for software?
Ubuntu, amongst others, have brought a new wave of end users, do you feel like their behavior differ greatly from the traditional GNU-camp (i.e less ideological and more “practical”)? What, if you would speculate, is their primary driving force to move to Linux (costs/freedom/openness/etc…).
Do you have any figures (or gut feeling) on the willingness for businesses, as opposed to end-users, to spend money on Linux software or support (and with desktop, not server, in mind)?
May 27th, 2009 - 3:01 am
Hi Bryan,
thanks for conducting your experiment, I thought you get results like that, actually you did better than I thought.
The problem is that you still have to be “tech savvy” to use Linux as you are usually required to install it yourself, and “tech savvy” people know how to automate their system and get around problems with eBook readers. So we have a chicken and an egg situation, and until you can reliably get Linux as a pre-install option in regular outlets there won’t be folk who need the “shareware” style of software, but we need it available for when the non “tech savvy” folks come along.
I’m sorry that you weren’t able to make a living out of your project, I think that it has its place.
Darrin Ritter
May 27th, 2009 - 8:42 am
Just thinking out loud, but if somebody were to run a Linux app store with some kind of profit sharing program, it should be somebody who has:
1. Street cred among practically-minded Linux users
2. Street cred among Linux or cross-platform developers
3. Reach to a broad audience, to promote new store when it launches (to get a critical mass)
Just saying.
May 27th, 2009 - 10:20 am
@Johan I wouldn’t use the word freetard, it is a word that was coined by a bitter person in order to ‘demonize’ certain people.
I agree with you, I want things to just work too and admittedly it can be frustrating occasionally but on the whole things are getting better by degrees of magnitude ( I’m remembering back to when USB didn’t work on Linux now its faultless).
Now days I spend my days doing stuff on my computer and having fun rather fixing stuff that doesn’t work.
I think we all need to come to a point of balance when it comes to these things some folk can be a little ‘over zealous’ and some seem to swing from one extreme to the other and become very Anti Linux, we must try and aim to be somewhere in the middle.
Darrin
May 27th, 2009 - 1:03 pm
Two things to note about Linux.
First, when I bought commercial Linux software in the past, I have been disappointed because the commercial applications broke for various reasons - a change in the library, a change in file location etc. and the vendor provided no way to fix it. Whatever the cause, the commercial application broke and I was no longer able to use it, even though I had paid for it. This never happens with an open source application. After that experience, I try to avoid commercial applications whenever possible. The moral of that story is that commercial application vendors should either write applications in a way that can’t be broken - bearing in mind that Linux changes more rapidly than Windows or OS/X, or maintain updates in order to compete with FOSS applications.
The second thing to remember is that the distribution channel for Linux and open source is the free distribution and repository maintainers. Why would anyone go looking anywhere else when you have the huge range of software installable at a click on your package manager. This means that in order to distribute commercial software most effectively, you need to get it into the open source distribution channel - ie. it must be open source. Hence the easiest to market commercial model on Linux is the customised embedded appliance, or the web based application based on open source client components. Alternatively, a rather more tightly defined environment like Android or Google gears might be a good idea. Another option is a common Windows/Linux application based on WINE.
May 27th, 2009 - 3:52 pm
It is just so convenient to install and keep maintained open source applications that are supported by the Linux distribution - you don’t even have to buy shrink wrapped software in a box or fill in your credit card details online, and you can try before you commit. It is the same as preloaded software on Windows (except you have at least three thousand applications to choose from) - you are going to need something special to persuade the customer to replace the default FOSS distro supported application.
To avoid application breaking problems, I would suggest the following:
For complex commercial applications, I would suggest a virtual machine with the app installed (same hosted VM can also be used under Windows).
For commercial applications of low to medium complexity, I would suggest WINE, or Android (same app can be used for Windows), or statically linked Linux application.
May 27th, 2009 - 4:06 pm
Just a thought that for most Linux users, using Linux is about freedom of speech, so buying closed source commercial software kinda defeats the purpose.
However, I do believe there are certain areas of software that even the Linux users of freedom of speech would use. Keep in mind that a lot of current Linux users (this is changing slowly but surely) are fairly tech savvy and in all likely hood play computer games.
Yes I know that there is software out there from commercial vendors like Adobe that people would want to see come to the Linux desktop natively (Like Photoshop). But in my opinion (for what it’s worth) you’d probably see the greatest rate of return if commercial software gaming companies would make a port for Linux for there games. Blizzard is making a small leap in some ways by making MAC clients for all there games moving forward. Also, yes I am aware that Eve-online tried supporting a Linux client for a while and recently has since canceled that client. For those of you that tried that client however it used Cedega as the client and as much as I respect any company or software that tries to close the gap on software between Windows and Linux, Cedega has never worked right for me and I always end up back with Wine. Cedega is not a “native” Linux port, had Eve-online made a “native” port for Linux like a tar.gz, .deb or .rpm, I think it would have had a heck of a lot more success.
May 27th, 2009 - 4:15 pm
Hey Bryan…
My comment about your experiment would be way too long, so I posted it as a blog post on my own blog: http://pharaohtechblog.blogspot.com/2009/05/linux-commercial-suckage.html
May 27th, 2009 - 4:49 pm
I think the distribution issue is the biggest issue here.
Most folk are used to their own distro’s repos and typically don’t stray outside of those. They want to open up Synaptic and grab an app, and if it isn’t there, they aren’t going to bother with it.
What would be great would be an app like Steam that can be run on all the major distros (Ubuntu, openSUSE, Fedora) which points to a central app store. In the app store, when a dev submits an app, they can submit packages for the different distros, (debs, rpms, etc) so that when you open up the app sore application, no matter what distro you are running, you can install the app. And if the dev only submitted the app as a deb, then it won’t show up as available on a Fedora box. So the app store application on sees what is available for the distro on which it is currently installed.
I would actually go so far as to say that this could replace all other distro repos. Put all the applications, free and paid, into this store. That way, you don’t have to go to Synaptic or YaST for some apps and to a different application for other apps. Have this app store application store your credit card info that you enter into it when you purchase a paid app and going forward, you can purchase with one click (a la Amazon MP3 and iTunes). The app store could function in the same way as the iTunes App Store, so that it is easier for small devs to get paid for apps and not have to worry about setting up ecommerce sites and dealing with all the transactions - the app store handles all of it and just sends a check to the dev.
And of course when I talk about similarities to the iTunes App Store, I mean only in structure. This app store would not curate or vet any application nor would it prevent any application from being listed in the store.
And perhaps in time, the app store application could be configured to install apps on any distro from just tarball or something, so that devs wouldn’t have to build 3 or 4 packages to get it on to multiple distros, the application would simply know how to install the app on the distro on which is were installed.
May 27th, 2009 - 4:53 pm
I never heard of your software until LinuxToday linked to this blog. I’m not interested in comic books. Maybe the response wasn’t so good, because the software is focused toward a limited subset of users/interests?
The DoItNow program looks more interesting. BTW, I use openSuse 11.1.
May 27th, 2009 - 7:19 pm
Do you think ideas like AppStore as propose in Ubuntu Developer Summit might encourage developers more?
May 27th, 2009 - 7:41 pm
[...] have been some excellent questions and comments from the last two posts (Commercial Software Experiment: The Results and Radical Breeze Software Now Free), so I want to take a few moments and address a few of them [...]
May 28th, 2009 - 4:42 pm
[...] Lunduke has been writing a lot about his attempts to make money from writing proprietary commercial software lately. I had initially decided not to ever mention any of his posts because I believe the man to [...]
May 30th, 2009 - 5:13 pm
Thanks for your postings. I’ve pretty much reasoned through the similar arguments and observations before; hard data is nice.
Maybe with more windows people coming over this might change, but I’m not holding it high.
July 20th, 2009 - 5:25 am
Is a month really long enough to go from one business model to the next? I understand why you did it but it’s still disappointing to see someone try and only give it a month before saying open model doesn’t work. You work hard on your product so you deserve compensation. Damn, only a month?
August 13th, 2009 - 2:17 am
I think one month is definately too short a period to judge what a product can do. Also, I think there should be something said for small applications and big ones. I can think of many larger applications that Linux users can benefit from, such as a good audio editor application like Cool Edit by Syntrillium (now owned by Adobe) or a good Quicken style application for point of sale and small business money management - applications of this scale are what commercial purchasers are going to go for. Then, they also don’t necessarily need to be a “try-before-you-buy” style app. Your app could, for example be part of a touch screen point of sale cash register (just as an example - mind you). They still need to be affordable, or packaged with something else.
August 26th, 2009 - 10:50 am
I think there’s a bit of a cultural aspect here. Plenty of freeware OSX apps ask for donations, with NeoOffice a good example. Mac users seem to expect to pay out money to get decent software. On the other hand if you launch a linux app and get nagged for donations it’s like, “but it’s linux - should be free??”.
Since the distributions manage most / all linux software, I think they should take the lead to embrace commercial linux software, preferably without alienating the purists in the community. The Ubuntu commercial repository is ideal for this - make it easy to find, secure to use and integrated into Ubuntu. Win win.