Making money with open source

If you’re going to do something, do it well

I’ve been thinking a lot lately about open source business models. The project I’m most often associated with, Elgg, developed a business some two years after the project itself was founded. Now that I’m considering a new project, I’m exploring what it might mean to create an open source model that has a business model baked in from day one. Not as a loss leader or the engine for a consultancy company, but as the core of a scalable business that turns a profit proportionally to the popularity of its end-user product.

An image problem

Particularly on the web, open source isn’t cool anymore. For a while, in around 2004-2006, the success of projects like WordPress and Firefox meant that new, web-based open source projects were coming out fast and strong. The former arguably gained its market share after Movable Type, previously the market leader in blogging, imposed an ill-conceived commercial license. Five years later, although new projects like StatusNet and Diaspora are getting headlines, and WordPress is going strong as the basis of at least 12% of all websites, most new open source efforts are on behind-the-scenes technical infrastructure like CouchDB and node.js.

Furthermore, open source in general faces a demographic problem. Only 1.5% of open source contributors are women, for example, compared to the already questionable 28% in mainstream software.

Yet off the web, end-user open source projects are growing, often in conjunction with a strong business model: think Ubuntu, Android, Boxee et al. However, on the web, WordPress’s parent company Automattic is the only company I can think of that’s significantly bringing in revenue (estimated to be $30-40 million annually, and breaking even).

I believe in open source, but I also believe that for open source projects to be sustainable, they need to be able to fund their developers and be emotionally rewarding. I think there’s a place for a friendly open source project that’s both accessible to new users, and to the kinds of people who would not ordinarily contribute, while both turning a profit and having a good time.

Routes to success

There are a bunch of different commercial open source models out there. Here are a few, in increasing order of commercial viability:

1. Donationware. Some projects solicit donations from their users, charity-style. Although this may result in a few hundred dollars here and there, it’s not going to make a significant contribution to payroll; people mostly won’t pay for something unless they absolutely have to. Asking for donations bases your business on goodwill alone.
2. Advertising. Think Firefox, which makes around a dollar per user per year from its integration with Google search. Products like Vuze also contain advertising. For stand-alone apps, products like OpenCandy can bring in real cash. Open source web apps, however, are very easy to edit and customize. While that’s a strength and a net positive, it means that adverts – perhaps embedded into an admin system, for example – are very easy to remove. Once again, you’re reliant on goodwill, unless the advertising is present on the open source project’s community site. Elgg has a page where users can find third-party hosting, and makes money from affiliate links. Meanwhile, many have forgotten Matt Mullenweg’s foray into search engine spam before WordPress settled on a business model, which speaks more to the difficulty of making money with open source than Matt’s ethics. (Automattic is a great company, and he’s proven himself to be a great guy.)
3. Consultancy services. Many projects provide tailored customization features to individual customers. This can be profitable, but isn’t particularly scalable: because each customization or advisory report is bespoke, your potential profit is capped by the number of human-hours your team can put in. Effectively, you become a digital agency, with your open source product acting as a way to draw attention to yourself. Think of your software as a fridge. If GE had to design a new fridge for each customer they had, they’d never make any money, and fridges would cost $100,000. Instead, they design a fridge once and sell hundreds of thousands of them for $500. Fridges scale; bespoke consultancy does not.
4. Freemium hosting. This is the model used by WordPress, StatusNet and others: allow anyone to create a free, hosted account, and charge for professional extras like analytics features and support. Determined, technically savvy users can still download and install the software themselves (this site runs WordPress on my own server), but using the commercial offering is often a simpler, more sustainable way to go. This is a slow burn, but WordPress has shown that if you create a product with enough critically reliant users, it can work.
5. Physical, commercial products. Boxee sells an actual set-top box in conjunction with D-Link, which is doing well. Android is a red herring here: while it certainly sells phones, with the exception of Google’s Nexus range, where the company presumably takes a cut, the operating system is provided for free to handset manufacturers. The real money comes through search advertising.

What’s the right path?

I’m not sure there is a correct solution for open source web projects – except to avoid #1 and, preferably, #3. Here’s what I’m thinking at the moment:

Open source projects are continually bombarded with feature requests. Fundry, the crowdfunding site for software projects, is very interesting to me as a platform for these, and as a contributory business model for open source development. The site allows users to add new feature requests and back them with money. The development team can then choose which ones to undertake. The core team would probably have to seed the Fundry page with a number of features to begin with, but with a little momentum I think it holds some promise. It’s also a great way for community members to practically contribute without creating code, documents or designs.

However, that alone is unlikely to allow anyone to eat (particularly looking at the current level of funding pledges – at the time of writing, Fundry itself has only managed to raise $128).

Another model might be to outsource the hosted freemium service, in a way that’s tightly integrated with the open source community site. A potential user might visit the project site and see a great big “create your site” button; on clicking that, they are led to a third party (or given a choice between third parties). An affiliate commission would be provided by the chosen service provider.

Commercial support services are a viable option, and can be deeply embedded both into the project site and the software itself (as a clear “get support” button). Commercially hosted value-added services, like Automattic’s Akismet anti-spam service, are another.

Most likely, a commercially successful open source project will use a combination of these. But what do you think? Is there a glaringly obvious open source business model that most projects have missed? And does profitability have a place in the open source movement to begin with? Let me know in the comments.