Commercial Open Source — Or Just a Free Demo?

August 12th, 2009 by Eric Barroca Leave a reply »

I’m reading a lot recently about “Commercial Open Source” being the next great thing in the software industry. I’ve just read the presentation “Talk Slides: The Commercial Open Source Business Model” by SAP Labs's Dirk Riehle. It’s a great presentation that really captures the vision and strategy of some high-profile companies in the so-called “Commercial Open Source” arena.

In summary, the “commercial open source” business model is based on the 3 pillars:

  1. a GPL (or GPL-like) software tagged “community”
  2. a proprietary version of the GPL software with some “proprietary extensions” sold using a traditional license
  3. a serious dose of communication efforts to explain how open source magically creates cheap great software for everyone (and that in fact you’re not really selling it) and generates a ton of leads allowing you to get to market faster and cheaper

Core myths of the “Commercial Open Source”

Here is how I analyse the core speech of the current "Commercial Open Source" model. I would be happy to hear about your opinion.

“Community Editions” you mean… “Free Demo”?

I don’t get the fundamental difference between offering a “low-end” open source version of your software and offering a free but proprietary one. Especially when I read that the “Community version” is for “developers, hobbyists and small deployments ‘cause it’s cool and fun tech” while the “Enterprise version” is for “enterprise deployment ‘cause it’s full of tests and stable code.”

Hmmm… how do you turn an untested, unstable software into an enterprise-grade, rock-solid software with “some extension”? Well, you don’t. :-) Either your software is well-tested and rock-solid, or it’s not.

If you just want to create a product line with different features depending on varying customer size, that’s fine but let’s call it this way.

Open source software generates leads, right?

Wrong. Freely downloadable apps generate leads. Free trials generate leads. Smart marketing efforts generate leads. Having access to source code does not generate leads, at least not when you are offering applications (it might be different for middleware or dev tools).

Want to generate leads? Create great software and offer free trial and downloadable software.

“It’s not proprietary software, it’s giving reason to buy when people use the software”

I love this one! Seriously. Of course customers need a reason to buy. That’s why they buy. :-)

The reason to buy is called “license fee for usage right”. It’s been around for 20 years and if you want to give people a “reason to buy” your software, just use a “license fee” for usage and maintenance. That is what it's designed for — it will save you a bunch of marketing dollars.

Open Source is good for Communities

It helps but it’s not enough. And it’s not limited to Open Source.

Openness, honesty, good software and good marketing create community. Ask Atlassian, Google, Twitter, Salesforce or even Microsoft. Not open source, but great communities and vibrant ecosystems.

“Commercial Open Source” or “Ashamed proprietary software”?

There is nothing wrong with selling “usage right to use binary software” (or more often called “license fees”…), which is what all “Commercial Open Source” vendors are doing. It doesn’t prevent you from creating great software, building a community and be nice. It just requires a bit more effort.

It's time to go public and add some clarity to all this. There is nothing wrong selling proprietary software, especially when you're contributing a lot of open source code (I’m a great fan of Atlassian and Day, in this respect). It is nothing to be ashamed of. Just be clear and focus on your software's competitive advantage rather than its open source "nature."

“Commercial Open Source” is not the business model of open source

There is no such thing as a business model of open source, by the way. There are many reason companies are producing open source software (from Microsoft to Google, from Oracle to RedHat). The only common fact: it’s a tsunami in the industry. Everybody’s using it, software vendors being the firsts. And many are producing some.

There are a lot of reasons to produce open source and a lot of ways to make money leveraging it, as some brilliant analysts and bloggers already said (two notable reading: “On open source business strategies (again)” by 451 Group’s Matthew Aslett or “Making Billions with Open Source, Revisited” by Redmonk’s Coté).

Here is how I would summarize it:

  1. Proprietary software (!): Build and distribute proprietary software leveraging open source ones (be it complete apps or just extensions). Take Day Software, quietly producing tons of good open source infrastructure components, they sell a great proprietary app. Or IBM with Geronimo / Websphere. Or Oracle. SpringSource and most “Commercial Open Source” companies fall into this category too. I think it’s the easiest way to make money out of open source.
  2. Support & Packaged Services: Sell support as subscription and high-value packaged services (monitoring, inventory, etc.) for open source software you’re producing. JBoss was the flagship in this business with quite a success making money with it. This is Nuxeo’s business too.
  3. Proprietary distribution: assemble open source software into a proprietary stack. It’s all open source software, but the recipe to assemble the different components together and deliver a coherent and supported stack is kept secret. This can also include some “proprietary services” such as automated updates or monitoring. This is RedHat’s business. Sun seems to look toward this way too (see Solaris and the recent WebStack).
  4. Proprietary tooling: sell proprietary tools that help running / operating / managing open source products. These tools are usually development tools, administration tools or deployment tools.
  5. SaaS: package open source software to deliver apps as a service. This is the business of managed apps hosting (to make apps run) and packaged services (to deliver great customer support and business domain knowledge). This is also Nuxeo’s business.

So in the end, what’s the key point? Is it doing open source no matter what for the sake of hype or is it solving problems by delivering great software and/or services to customers?

I wish people of the Commercial Open Source arena would focus more on the later…; cause it does not diminish their contribution to open source overall nor does it diminish their company’s greatness and value. For the best of the open source industry. These times are about transparency and openness after all…

What do you think? I have spent 10 years in the open source software industry, building a company, living through short-term hypes and various business models. And I'm still learning. :-) Would be really happy to discuss more about all this.

Cheers,

EB.

  • http://jonontech.com/ Jon Marks

    Perfectly said. To some Commercial Open Source vendors, the word “Open Source” is interchangeable with “No license fee for the core product”. The fact that those vendors are the only ones that implement the product and that every line of code in the product has been written by employees of Vendor PLC doesn’t seem to bother them. Great post.

  • http://gregdoesit.com Gergely Orosz

    Great article, I agree on most. You have to admit though that there’s a great marketing value in being commercial open source because of the “open source” word. @Jom Marks: would you buy from a vendor who has source code they haven’t gone through or ship with modules with no guarantee though? In my opinion it’s unenviable to have the vendor write the core code itself or at least fully have it tested.

  • http://www.contentprise.com/ Lokesh

    True. I liked the fact, how can you turn untested, unstable software into an enterprise-grade, rock-solid software with “some extension”? It’s not new, it’s the traditionally old way of making money by just adding a tag of “Enterprise” Version. The only good thing that comes with the enterprise version is the “product support”, which is more about giving confidence to the customer. From a developer’s or a hobbyists perspective , wouldn’t then the trend to contribute and share knowledge, be discouraged if the trend of “commercializing open source” or “so called o-s business model” or acquisitions of smaller open sourced cos by the big giants continues? Way back in 2006, these were my views – http://www.contentprise.com/2006/02/enterprising-open-source.html

  • http://www.knowledgetree.com/ Daniel Chalef

    Eric, Great post. I do agree with you that some of the messaging coming out of commercial open source vendors in the recent past has been FUD around proprietary software. This may be / have been some reciprocity for similar FUD directed by very large mostly “closed source” vendors. As to can adding “some extension” to a Community Edition (CE) really warrant calling it “enterprise-ready”: many vendors follow a “release early, release often” policy around their CE. The CE is definitely not stabilized, and potentially not feature complete. Once fully stabilized the product is released commercially. This doesn’t happen that often – maximum 2-3 times a year (which is actually preferred by most customers). The product management benefits of the above are numerous – early feedback about features, stability and the like are very helpful. And yes, bug fixes are helpful too. This approach is also shared by vendors that fit into your category #2. For example, Canonical releasing Ubuntu LTS far less regularly than other Ubuntu dists. So this isn’t about just adding an additional feature or two, it’s providing stabilized application + features + support. Best regards, Daniel

  • http://profile.typepad.com/6p0120a4e9f376970b Alain Risbourg

    Eric, I agree with you! Indeed, I think that Commercial Opensource is only a business model variant of Software business models . It’s mainly a marketing tactic that goes along business requirements : openness, trust and transparency… Commercial businesses (Proprietary as well as Opensource ones) are (mainly) profit-motivated, but for the developers/contributors of the Community-based Opensource organizations, ie the original Opensource communities/projects (such as Linux, Apache, Tomcat, BSD OS… and few others), the drivers are those of giving (or sharing), and (mainly) notoriety. The drift within the software realm is between Community-based projects and Commercial Opensource businesses! In addition to this, I think that there is another major drift, which splits the software realm, this time between Opensource organizations (both Community-based and Commercial Open Source) and Proprietary software businesses, and it’s Innovation! It would be too long to explain why, but you can know more about my thinking in my comment to John Newton’s interview in LeMagIT : http://www.lemagit.fr/article/open-source-emc-ged-opentext-sharepoint-alfresco/1542/2/john-newton-alfresco-laquo-nous-sommes-039- I wrote it in french a while ago. Sorry for that… ;-)

  • http://blogs.nuxeo.com/ebarroca/ Eric Barroca

    @Jon: agree with your comment and thanks. @Gergely Orosz: your point relates to software engineering, not licensing. I don’t trust proprietary software more than open source ones on this level. I trust well-tested software, and they can be either of those. Look at Apache HTTPD or Lucene, don’t think think they are @Daniel: I would argue that if your community version is not stable, there is chances the enterprise one won’t be too (but it’s related to software engineering, not to the business model directly). If releasing an enterprise version means enforcing a code freeze + focus on fixes period before the release, and that you do it on the main (public) code base, I think your community edition is probably also enterprise-ready. In that case you’re doing standard Q&A and release process to reduce the number of versions you have to support (which is totally right and understandable). You can’t make good enterprise software out of crappy open source one, and you get crappy software if the “open source” is not also “stabilized” at regular pace. Again, I have nothing about making money selling proprietary extension or doing proprietary distribution, or whatever as long as it’s clear. @Lokesh, @Alain: don’t read me wrong :-) I totally support make money and profit leveraging open source. This is what I do for a leaving for the past 10 years. Open source software needs businesses as well as the reverse is true. Being profit-oriented is sane, especially when it benefits to customers. Our company is profit oriented. And we are producing open source because we think it’s the most efficient development model and licensing model for us. Not because of any ideology. And I see the success of the open source licensing and development model as market driven: it succeed because it’s more efficient than the traditional models. I just think some clarity and transparency about what customers are paying for and what company are actually selling would help the business as a whole.

  • http://www.sensenet.hu BiroTom

    Wow. Great arguments. But there is a difference between commercial open source offerings, and the devil is in the details. Some offer an open version, usually with GPL and some closed source modules for money. Some offer a free version with GPL, and only charge for support, and/or legal protection. I do not like the first, but the second I think is fair. This is what our company does, too. We give away the source with GPL, and only charge for services. But the code you get in the free and the “enterprise” version is the same. No hidden modules that only enterprise customers get. They get the same source, same binaries, except they get it only when it is very well tested. The free community edition contains all the fresh code, that needs more testing. I can also assure you, that befor we went open source, we did not have an international presence, now, as we are open, we have thousands of downloads and 4 sales opportunities this week! And those, who cannot afford the commercial license, can also use the GPL and look for free support in the forum.

  • http://profile.typepad.com/6p0120a4e9f376970b Alain Risbourg

    @Eric, I don’t read you wrong : you wouldn’t work with Stéfane at Nuxeo… Indeed, I think that this makes your post even more clever : you have the correct attitude! BTW, I think that we disagree on another point … I think that, for a given (and identical) IP, the proprietary software business model is better than the commercial opensource business model as I’ve written about on my blog, after a discussion with Ludovic, from XWiki. Good luck Best Regards

  • http://blogs.nuxeo.com/ebarroca/ Eric Barroca

    @all: thanks a lot for the nice comments and interesting reactions. @Alain: completely disagree. If for any given software the proprietary model is better than the open source one, there wouldn’t be any open source software. The open source development model is clearly more efficient from a macro-economic perspective: it reduces waste by allowing different actors to work together on the same core (non-differentiator) components. And I believe it’s much more efficient in most of the cases, not only for pure infrastrucure. I also don’t believe there is such thing as “the open source business model”. There are business models leveraging the open source licensing scheme, development model and distribution model. @BiroTom: from what you say, I would think that your company is in the “support as a subscription” provider category. Your selling high-value / packaged services for a given open source software platform.

  • http://blogs.sun.com/alexismp Alexis MP

    I think Sun’s a big enough Open Source player to have different strategies for different maret segments and I would argue that it’s mostly #2 (support + add-ons) with some #3 (distro). Where do you position JBoss with their new “fedora” release model? Same as RedHat (i.e. distro)?

  • http://blogs.nuxeo.com/ebarroca/ Eric Barroca

    @Alexis thanks for stopping by. ;-) on Sun: agree it’s #2 and #3. I think I like Sun’s open source strategy. And the scale is impressive. on JBoss: as far as I know it’s proprietary distribution (and a pretty smart one). And there also maybe doing some proprietary tooling, but I can’t confirm. Overall I think the model is clever.

  • http://www.InformationArchitected.com/ Dan Keldsen

    Eric – bravo on raising some of the hairier and “unspoken” problems or at least misunderstandings of both open source and proprietary solutions, as well as avoiding the demonizing of any approach. My main problem with open source is in helping clients to wrap their heads around exactly the issues you’d described – when is free not free, when is paid not paid, what can you do or not do according to the license, who’s supporting the solution, etc.. Not that these aren’t problems in the “commerical world” as well – but people are already used to the unpleasant dealings of the big platform players already. ;) Even after 15 years, I still can’t make sense of all that Documentum sells and the million combinations and pricing schemes. Anyone else? Simplicity all around folks!

  • http://profile.typepad.com/6p0120a4e9f376970b Alain Risbourg

    @Eric, You wrote : “If for any given software the proprietary model is better than the open source one, there wouldn’t be any open source software.” I can’t agree with you on this! What about GNU ? Indeed there is (mainly) 2 drivers - money, profit, self-interest… (whatever you call it),and - altruism, gift, community-based interest… (whatever you name it). Remember the original open source contributions, GNU, Richard Stallmann and the FSF : “Stallman wanted computer users to be free, as most were in the 1960s and 1970s — free to study the source code of the software they use, free to share the software with other people, free to modify the behaviour of the software, and free to publish their modified versions of the software.” (source http://en.wikipedia.org/wiki/GNU). This is what I would call “THE Open Source business model”. Everything else (and you might agree with me on this ;-) ) is software, and …there is no such thing as “the software business model”. Best Regards

  • http://www.knowledgetree.com/ Daniel Chalef

    @eric: This isn’t necessarily about releasing buggy software: keeping with the “release early, release often” ethos, some vendors (with vendor driven open source projects) release community editions to test the water for new functionality, and often release software that may not be feature complete. Once they’re comfortable that they’ve got the product management and technology right, they product a commercial edition.

  • http://stephanecroisier.jahia.com/ scroisier@jahia.com

    Nice blog post. I posted some other questions related to the same topic on my blog: http://stephanecroisier.jahia.com/what-are-the-best-oss-business-models-for-cci Stephane Croisier

  • http://www.nuxeo.com/ Stefane Fermigier

    Alain, as I wrote multiple times in the past, there is no single “open source business model”, and the GNU “business model” (which, if you don’t remember, was to sell tapes to cover software development costs) is certainly not the most sustainable, scalable, or efficient we can think of. S.

  • http://betterfasterbigger.blogspot.com Boris

    I think you look at it from the wrong perspective. Let me show you what you are saying: Scenario 1: Company A releases software a as open-source and charges for support Your opinion: cool Scenario 2 Company A releases software a as open-source; company B takes it and enhances it with whatever (SAAS, support, additional functionality, …) Your opinion: cool Scenario 3: Company A releases software a as open-source; AND the same company A takes it and enhances it with anything (SAAS, support, additional functionality,…) and sells it as “Enterprise” Your opinion: not cool (except if it is SAAS)

    Ahem. Notice anything strange (and I don’t mean the fact that enhancing it one way is ok, another way its not). The company in 3) is actually giving something away that you are allowed to use as much as you like, no strings attached. Why does it bother anybody if they are also doing something else?

    For me, open source is not about price. It is about access to source code, and the ability to work together without constraints. Proprietary software doesn’t provide that option. Try taking a “Free Demo” and fixing its bugs, embedding it into your solution or adding the functionality you need. There is a fundamental difference between a open-source “Community Edition” and a “Free Demo”. Just a question to those offering SAAS: do you open-source your complete SAAS infrastructure? If not, what makes your model morally superior to a straightforward dual license model? You just use a different channel, thats all. I believe in dual licensing. It is a win for everybody – the OS community gets something free software they can tweak as needed, EE clients get someone to yell at and the company that builds it gets the financing it needs.

  • http://www.mcmains.net Sean McMains

    I’m actually one of Boris’ customers, and agree with his points here strongly. Using open-source, we’ve been able to search the code find issues, and, if the company doesn’t prioritize the fix as highly as we’d like, fix it ourselves and submit it back to the community. We’ve had enough success with this model that Commercial Open Source is the first approach we look for when evaluating software these days.

  • http://robertogaloppini.net Roberto Galoppini

    Hi Eric, I joined the conversation by saying that commercial open source is not just about single vendor projects, as Riehle puts it. As you say a lot of so-called open source characteristics belong also to the proprietary world. So said, I went deep analyzing the importance of participating to open source projects, in terms of benefits and costs, “even” if you are a proprietary software vendor. Too often people talk about open source business models just mentioning licensing or sources of revenues. I believe that a lot of business models’ aspects, like the production of code, matter too (http://is.gd/2s263). You raised so many interesting issues!