tl;dr: While seeking open source audio software, I found a project which asks for a small donation before you can download the application binaries. I interviewed the leader of the project and found that he is able to support a family of five through this donation system. Awesome!
Lately, I've been recording music in my spare time. Since I try to use as much Free and Open Source software as possible, I found the free digital audio workstation Ardour. When I went to download the software, I was asked for a donation before I could download it. Intriguing!
The appeal was a familiar one to anybody who uses Wikipedia - a picture of the project leader and a short message explaining the costs of the project and how your support matters. The difference is that Ardour has a monthly goal, and unless that goal looks likely to be reached, the project doesn't provide free binary downloads of their software. You can still download the source code and compile it yourself, but I imagine this is too complicated for the average end user.
When did you start this donation model for Ardour and why?
When I originally started working on Ardour, I was relatively financially independent. This changed as my personal circumstances changed, and by 2008 I was reliant on the support I was receiving from a few companies in the audio technology world. In early 2009 this support had all come to an end, and I could either give up working on Ardour and find a "regular" job, or figure out some way to make a living from Ardour itself. I had been accepting donations for several years, and had added subscriptions a year or so earlier.
The big change was the so-called pay tunnel which was inspired partly by Radiohead's handling of their album In Rainbows in 2007. I didn't want to try to force people to pay for Ardour, but I wanted to "encourage" them as much as I could.
How much revenue does it generate? Is this enough for you to work on Ardour full time?
The target is $4500 per month. Occasionally it makes or slightly exceeds that target. Often it falls short. I work on Ardour full time based on this income, though the last couple of years have seen me do some relatively short term consulting that has supplemented my income by about 10%. My real goal is to grow the income enough to pay some of the other developers who work on Ardour with incredible dedication, and no payment at all. My family lives without health insurance or any clear retirement plans, but we can also live cheaply thanks to my participation in helping start Amazon. Having complete control over my work schedule and life is worth something - I'm not sure how much. If someone offered me $200k/year to work on something that I really wanted to do, I would undoubtedly say yes and give up that freedom. If they offered me a generic programming job paying as much as I earn at present, I'd say no. For the more realistic scenarios somewhere between these two its a difficult judgement call.
On the one hand, I am extraordinarily lucky and fortunate to work as I do on a project that love. On the other, living the US with 3 children who still need our support and doing so on a bit more than the median household income is not always a comfortable experience. In general, I prefer freedom and loving my work over comfort though, and try to be optimistic that eventually revenue from Ardour will improve or a couple of other projects that I have some involvement in may do well.
What feedback have you received about the program?
Most of the complaints are of one of two varieties. The first are people who don't realize that source code for Ardour is always available and that anyone can get Ardour at any time without paying anything for it if they are willing to deal with svn and building the software. The second are people who for one reason or another cannot use PayPal or a credit card.
As usual, one doesn't hear much positive feedback about a system like this. There are several people who try to provide ideas about improving revenue flow - unfortunately most of them do so without any awareness of the actual numbers or what has been tried in the past.
As an open source project, do you worry about the missed opportunities from being redistributed elsewhere, such as in distros and on other websites?
The fact that most Linux users of Ardour get the program from their distributions without a doubt takes away some potential revenue and tends to disconnect people from the fact that Ardour is an ongoing, live software project with its own forums, support channels and developers. It also means that we have to deal with packaging mistakes and distro-dependent problems when those users eventually do find ardour.org. These are problems, but for now they are manageable.
Our biggest issue really is that the majority of Linux users are using Ubuntu and for an apparently large number of them, this distribution (like many others) is badly setup for the kind of requirements that Ardour users have. The fact that JACK and Ardour don't work "out of the box" on a given distro is a huge stumbling block for many actual and potential Ardour users.
How do you plan to improve the donation model and Ardour in the next year?
No real plans at this point. I'd prefer to focus at some point on a subscription drive to get more people to commit to 12 months of helping support the project financially. A steady stream of releases sounds like marketing BS, but my experience is that it too drives revenue. I'm not happy about that, but it seems to be a fact of life.
Do you have any advice for other open source projects looking to adopt this model?
A "conversion rate" (i.e. the fraction of people who download the program and pay for it, versus those who choose not to pay) of more than 3% should be considered successful. Attempts to find out why people don't pay result in statistics like this:
Reasons for paying nothing:
Note that "Too lazy to even say" is the default setting on the page that asks this. Most of the "Other" category are actually one of the previous 6 reasons, but people feel the need to say things like "I can't afford to pay" or "Fuck you Paul".
Finally, I will note that open source projects seem to me to divide into 3 broad categories. One I would term "infrastructure" and includes things like the Linux kernel, Firefox, Apache, OpenOffice and so forth. These projects produce tools are used by millions of users, among them people with a financial interest in seeing them supported, developed and so forth, and so they manage to have significant developer presence and momentum.
A second category are what I would term "small app" projects - soundjuicer is one of my favorite examples. These are projects that can be meaningfully completed by one person, have defined goals that can actually be reached, and low maintainance requirements. Such projects don't need financial support in order to happen (though I'm sure their authors wouldn't mind), and there are typically many different attempts to provide given functionality.
The final category is the one I would place Ardour in - large, complex niche projects that cannot possibly be developed by single individuals, but have a limited user base without clear financial reasons to support the project. Programs like the GIMP, various video editors, and other media tools all fit in this category. Finding ways to ensure that these programs move forward at a rate that makes them viable for potential users is always going to be a challenge.
Thanks again to Paul Davis, leader of the Ardour project for talking to me about this!
What do you think? Do you think other Open Source projects should adopt this model of business? Leave your comments below!