The new meme in the Open Source business blogosphere seems to revolve around an hybrid model pioneered by MySQL who (surprise!) is considering the addition of proprietary components for paid customers. Fabrizio agrees with the move, noting how a support model can only bring you this far as best customers turn into worst nightmares once they realize they have become competent enough to support their Open Source stuff. I agree on the conclusion (commercial Open Source, as we know it, won't be there in a few years' time, and a hybrid model will surface), yet I don't quite subscribe to the motivation.
Think of software as a car. People need cars. Some people actually like cars. A few are actually mad about cars. It's hard not to find value in wheels: at a very least, they bring you from A to B, but in most cases they appeal to your ego. A lot of people will save to buy a better ride and will anxiously anticipate the moment they'll get to experience that new car smell. Serious comparisons are drawn, comparative reviews are discussed on owners' forums, endless discussions arise between fans of different brands: it's a value-driven market where price matters, but it's just a row in the table, and the perceived value of a shiny iPod integration goes way beyond the price of a few cables and a plug.
The problem with commercial Open Source is that it doesn't sell cars. They sell insurance. And, guess what, people hate insurance. Every car owner needs to be covered, but I find pretty hard to imagine people doing lots of research on the fine print of insurance contracts. What really matters is price: the cheapest alternative that ticks that legal box is a great candidate for the winning spot. There is no value perceived in insurance: you need to have it, and that's it. There is no retention, either: the cheapest one wins no matter what, and people will happily jump the cliff and switch to a different supplier if that saves a buck or two.
Open Source needs to realize that the basic equation is free distribution vs. greater adoption. Adoption will bring participation, uptake, references, better software and all that. Adoption in itself, however, will provide just marginal business if the upsell proposition is based on insurance alone: there will be no drooling customers eagerly waiting at the register to buy support, as the insurance model is neither sexy and nor useful per se. The marginal value of priority updates, coverage against ambulance chasers and a throat to choke should stuff hit the fan is far from being a compelling proposition, and any baitware/crippleware tricks such as gated access to professional services for licensed customers alone are just sneaky ways to build an unstable and controversial relationship with customers who will be signing a contract, then start looking for alternatives such as aggregators, systems integrators or roll-your-own to provide them with a boot to kick the vendor backside with.
This is why most commercial Open Source as we know it, is doomed to fail. Unless vendors stop moaning about the conversion funnel and start rolling up their sleeves and provide value beyond insurance, a lot of VCs are going to be disappointed. Meanwhile, a lot of companies out there are making billions with Open Source already, as they use it to create, enter, and subvert markets, generating opportunities via community-based open development, and using Open Source as a component of a bigger picture. There is no need to look forward for the first Open Source billion dollar company: it's been there for a while already, and it's a hybrid.