It's free, now pay me.

March 19, 2009 5:50 AM

How can I make money through open source development?

I will soon be open sourcing a web app that I've developed. It's currently in use by two different clients. They love it and think I should spread the word. It's important to me to release it as open source. I'd also like to make money.

I'm aware of several models for doing so:
1) Offering for-hire customization/development to support specific implementations.
2) Offering support.
3) Selling "premium" versions with extra features.
4) Hosted services.

Can you suggest any good resources for thinking about these and other models and other details that might be helpful to me?

posted by rbs to Work & Money (8 answers total)

For me, two of the main benefits of using open source stuff are not having to think about keeping track of how many seats are using it, and feeling free to contribute improvements. So I feel kinda queasy about (3). Also, (1) is really just a specific instance of (2).

To build a strong product and a strong community around that product, it would be best if any generally-useful improvements made in the course of (1) were available to the entire user base. So, consider using a bounty model to charge for (1) - if a customer really wants a feature that's not already available and is willing to pay to have it implemented, you should cheerfully accept such payment while making it perfectly clear that they are paying for the product to be improved, not for an exclusive right to use an improved version. I believe this is the way Namesys got most of its funding for ReiserFS before the lead developer's unrelated personal issues put that project down the toilet.

So, offer (1a) customization for a fee, (1b) ongoing development for bounties, (2) for a fee, (4) for a fee, and steer clear of (3).
posted by flabdablet's sock puppet at 6:41 AM on March 19, 2009 [1 favorite]

I work for a commercial vendor supporting an open source library catalog system, and we offer support, migration services, training, hosting, and development. With the exception of minor cosmetic tweaks, all of the code that our customers sponsor is contributed to the project. *Not* contributing changes back has disadvantages: you can get stuck maintaining a lot of slightly different versions of the same software and it can be harder to integrate changes made by other members of the community.

If you offer hosting services, the key is to automate everything. You should be able to set up, run, back up, restore, and maintain instances of the application with as little manual work as possible. If your app doesn't have a good installer, make one. It will make life easier for you and will make it more attractive to potential contributors.
posted by metaquarry at 7:00 AM on March 19, 2009

Also, one of the blogs I read, 451 CAOS Theory, discusses the commercialization of open source and business models for open source vendors.
posted by metaquarry at 7:11 AM on March 19, 2009

To use Seth Godin's terminology, you want to develop a Tribe. Here's a video of Seth talking about it.

That said, in addition to making money directly from that piece of open source software (your items 1 through 4), there are ancillary reputational benefits. If people know you as the wizard who created WizbangWebWidgets it will help you get other work, even work that isn't related to WizbangWebWidgets.
posted by alms at 7:22 AM on March 19, 2009

There's a mailing list called Free Software Business you might find useful.
posted by willem at 9:12 AM on March 19, 2009

In addition to the above stated avenues, you can also write a book and sell it. I don't know the economics of books, but if O'Reilly can host crazy conferences, he's got money from somewhere at least. That and training are nice ones.

I'd go with 2) over 4), if it's an either-or situation. Hosted solutions generally cost more than basic support contracts, and many corporate customers already run their own hosting.

And steer clear of bounties. What happens at best is one or two well funded ideas are posted and completed, leaving you with a bunch of 50 dollar requests that are basically an insult to the people who might take bounties. You're probably better off with a consulting system where you build a corporate image to take paid requests and contract support out behind it. Canonical does some stuff this way.

Other details: the size of your userbase matters. Assuming it's a Linux based product, publish Redhat RPMs, and get it into Debian (and Ubuntu by extension). Look into the requirements for those projects; but any blocker with them will likely be a roadbump in adoption everywhere else. You want a smooth ride. Also look into Gentoo and so on. In fact, your first community effort will probably be the packaging systems that go the final mile.

I also appreciate projects that use revision control, even when there's only one or two contributers. It helps me evaluate the pace and quality of development on a project with tools like Ohloh.
posted by pwnguin at 10:35 AM on March 19, 2009

Response by poster: These are all great. Thanks!
posted by rbs at 11:42 AM on March 19, 2009

I don't know the economics of books, but if O'Reilly can host crazy conferences, he's got money from somewhere at least.

A previous employer of mine used to write books. After all was said and done he would regularly make less than minimum wage ($9 - 10 around here) for his time. For the large majority of authors, it's going to either be because they like doing it or want a resume line (that might help sell consulting services, etc).

O'Reilly is a publisher, not an author. Additionally, conferences are often extremely profitable -- one of the magazines I used to do work for LOST money on all their publishing but made it back on the conference associated with the mag.
posted by fishfucker at 1:09 PM on March 19, 2009