The recent turmoil by Gartner's Brian Prentice blogpost and my upcoming trip to the Open Source Think Tank inspired me to pick up a piece I was writing long time ago. It's good to see Matt Aslett participate actively as it was his business model report "open source is not a business model" that triggered me back then.

Matt now says he couldn't write the report anymore with so many companies making money from open source. Hmmm, I don't think the world changed that much from 2008. Also back then, a lot of companies were making money of open source, albeit maybe for a large part by providing time & material services around certain open source products. Specializing in open source technologies is something companies like Optarus, Sourcesense and Smile, have been very successful at (read Gianugo's opinion on sustainable software here).
 
The big question for the venture backed open source companies is obviously this: how can we get returns to scale AKA very big margins. Jamie Dixon's example MicroStrategy IMHO is just doing a lousy job, Jamie probably agrees there :-). Simply put: from a business perspective we all like to be Microsoft and their licensing revenue for sure doesn't get consumed entirely by their marketing activities.

So what about returns to scale? Any software has returns to scale: develop once use forever. Point is, with a proprietary model you can reap the benefits of this initial investment yourself. Proprietary software has (near) zero production costs for the extra dollar of revenue. And the cool thing is: entry barriers are very high, because the initial development costs are high (natural monopoly). Really, Bill can't help he's a monopolist ;-). Of course there is a big problem: consumers are paying too much because as a monopolist you can differentiate pricing the way you like it (low cost for students, universities, high costs for businesses, different pricing in different countries, etc, etc.). But the other big problem is: monopolies don't innovate. They milk their products. Of course they make it look like they innovate, but the pace is controlled by them and therefore slow. How many innovations have we seen in Windows 95 to Windows 7 (mind you, there's more than a decade between them) other than that the product now looks like it's finally stable and performing? (I won't bother you with tacit collusion here, but rest assured MS, SAP and Oracle are silently becoming each others biggest friends, not foe).

So what about open source? This is disruptive: you can't beat free. Innovation goes at a higher pace (the standing on the shoulders of giants effect kicks in) and it's basically a route towards perfect competition in software. Good for consumers, good for companies. Of course there's numerous problems I won't touch upon here (like a unpredictable development pace, etc, etc).

Software is not manna from heaven

Software is not manna from heaven, it's development requires time consuming labour. Someone has to do the initial investment in the software, and then it requires continuous effort to make it better and stable. Also it's somewhat naive to think that there're successful pure community driven open source projects without one or more companies backing them with continuous support and resources; look at the flight Drupal has taken since Acquia. There have to be companies that think they can make money by contributing to open source. In a way I am really indifferent how the code gets there... If some mobile phone maker decides to build a open source mobile OS. Fine, they make money from billing you for hardware. If a advertising company like Google decides to build a open source mobile OS. Fine, they make money form running ads on your mobile or selling your (hopefully) anonymous profile. This is all good stuff for the customer and clients of Brian Prentice, they can re-use the investment by others and leverage that. In short: to me any business model that drives people or companies to invest and grow the open source stack is good. We all benefit. 

So what if a company creates a proprietary product and in the process contributes to the projects they are using from Apache? Fine too! But doing proprietary add-ons is obviously were this "open core "discussion is about. This is were it gets messy: say there's an open core BI engine and two companies are competing with add-ons on this product and since it's open core.. (assuming the core is distributed under OSI approved license) they both contribute new features to the core. Everybody can start it's own add-on business on this core, and the community can start providing features like these proprietary add-ons in the core itself as well. This certainly has advantages over likely any closed source model out there. But hey? Don't we have a word for this already? Open Source... But then again the companies providing add-ons are basically plain old proprietary software vendors, that contribute back to the open source stack.

The problem is indeed, when the benefits of open source for the customer are attributed (by marketing) to this proprietary vendor itself. The reason for the category/term "open core" to pop up, is that in real life it's probably not nearly like my description above. There's one vendor deciding what goes in the core and what is (a paid for) add-on, good old bait & switch. I agree with Brian there, what benefit does it bring to the customer? None. 

A open source software business model simply doesn't exist, by definition. You can't charge for the software. What we are seeing in esp. with the venture backed so called open source companies, is that they still try to catch some of the value that open source software brings and make money directly of it. I think your business model should be around something else than the software. Let it go, you are not selling software anymore. But there's still plenty of stuff you can do that does scale: support subscriptions like we are doing are like insurance policies, they scale. But yes, there's competition. Someone else can provide these services too, this keeps pricing at a realistic level. Good for the customer. In short: you have to bring added value to the software. 

Good luck everybody with finding a scalable model! 

Two final thoughts

On Nuxeo studio discussion: if they have proprietary software that enables them to run Nuxeo as SAAS it's is fine, given that a customer can take it's data out and locally install Nuxeo for free and continue working like he used to (with all features). This is absolutely a great benefit to the customer as it gives him freedom of choice and he's not locked-in, like with many other SAAS providers. I see this as one of the scalable services you can provide, other than selling software.

What about integrators using Open Source as a basis to develop custom features for you. With pink glasses on their customers are getting value, they only pay for the custom stuff. But what is the business model of integrators? Their real incentive is to do a as many hours possible.. So you end up with a potentially unmaintainable bag of proprietary code, were proprietary in this case means it's your own (if your IP lawyers did their work well), but all the problems it has are yours too, esp. when they have poked into the open source core itself and not given back. Maybe something to give some more thought later...