Online Music and Open Source Business Models

In this part of our series on An Open Source Business, let’s take a look at our friends in the online music space and see what we can learn from them.

The Deal recently had an article about online music startups which should strike a chord with anybody who’s thinking about or trying to make a business out of open source.  Look at what they had to say:

“huge numbers, lots of hype, a surfeit of hope and a major chance of failure… some of the business models are inherently economically unfeasible… It’s completely unsettled and more and more fragmented…The rules of the industry and the economics of the industry have completely changed…Technological advances offer more and more delivery mechanisms, user options and wizardly new features…However, just who can make money off all this is almost as uncertain now as it was five years back…Everyone is gambling there will be a way to monetize distribution of recorded music, But no one has come up with the solution…Last year’s great hopes are this year’s busts.”

Sound familiar?  It should.  In a nutshell, open source business models share the same strategic problem that these online music startups have: how do you make any money when most of what you provide is available for free? Let’s look at the ways:

Free the Software, Sell the Services

Just about every commercialized open source project follows this business model.  The software is free, but the developers charge for services such as support, training, customization, and software development.  Sometimes the services are “productized” into manuals, seminars, installation CD’s, and packaged support, but the idea is the same.

This model works well…to an extent.  For example, we’re the main developers of opentaps Open Source ERP + CRM, and we’ve found that users are indeed willing to engage us for opentaps-related services because of our experience and knowledge with the system.  However, we’ve also found that users are willing to hire us mostly for customizations which are unique to their needs.  We’re still responsible for the architecture and user interface of opentaps ourselves, and that’s why since the release of opentaps 1.0 we’ve invested in everything from integrating Spring, Hibernate, and the Google Web Toolkit to building a Domain Driven Architecture.

Like the Free Version?  Please Pay Us for Even More!

Many open source software developers, and virtually all open source software companies funded by venture capitalists, engage in the “commercial/open source” model.  An open source edition is available free of charge to attract potential users, and a fancier commercial version is available for pay.

This is not an easy business model.  Let’s go back to music as an example.  I like Pink Floyd, but if you gave me The Dark Side of the Moon for free, would I pay you for Ummagumma, The Final Cut, and every other song by Pink Floyd?  No, I wouldn’t.  (Another example is travel: how many people actually pay for First Class?)

But perhaps the best evidence that this is a difficult business model comes from the commercial open source companies themselves.  Compared to a few years ago, their websites are de-emphasizing the open source version (sometimes you really have to look even to find the download page), and their “community edition” licenses are increasingly restrictive.

Nevertheless, I think this is a model which could be very successful if two conditions are met:

  1. You must have a very large open source user base.  Think MySQL.
  2. You must segment that user base carefully and identify the unique needs for your “enterprise edition” product.  The need must be fundamental — a little bit of eye candy and a few cool features alone won’t be enough.

Be careful, though: if you execute this model incorrectly, you could easily lose the goodwill of your open source users and unwittingly give away a viable commercial product for free.

The Alchemy of Open Source

There is a famous story of the Stone Soup, where many free ingredients came together to make an amazing finished product.  Lest you think it’s just a fable, Red Hat and Ubuntu do exactly that–they’ve combined major open source projects such as Linux, Gnome, Apache, Firefox, Thunderbird, OpenOffice, and MySQL and built major businesses from them.

This is the business model we’ve chosen for opentaps so far.  We’ve built opentaps from major open source projects such as Apache, Funambol, Google Web Toolkit, Jasper Reports, Pentaho, MySQL, PostgreSQL, and too many others to name here.  We’ve had to be patient at times, but over the years, we’ve grown as all those other projects have matured.  Amazingly enough, these open source projects have put us years ahead of many commercial ERP systems technically and enabled us to build opentaps sustainably, so that we now have a fully integrated ERP and CRM system with business intelligence, ecommerce, and mobility integration without any VC funding.

But this is not an easy business model to execute.  You must be willing to understand other open source projects and have the technical ability to work with them.  Most importantly, you need patience.  With this business model, you are growing with the community of open source projects.

In the End . . .  Just Make it Better

No matter what business model you choose, ultimately you’ll succeed if you make technology easier and better for your users. In the online music world, there actually has been a great success story — iTunes.  They’ve done it by making downloading music easy and fun.  So learn from them.  If you can make software easy and fun, you will be successful.  Next to a great product, the business model is just a footnote.

In the next part of An Open Source Business, we’ll take a look at marketing strategies for open source software.

13 Comments

  1. Interesting post. This kind of takes the book Appetite for Self-Destruction to the next level.

  2. Yes, it does, doesn’t it? And yet, this is exactly what a lot of technology driven industries are facing now: everybody from entertainment, news, telecommunications, to software must figure out how to compete with or even make a business out of “free.”

  3. To me, the most interesting business stories in the music industry at the moment are the ones that are creating a more open model, supporting artists without getting them to sign over their rights, helping fans form a community around them, helping artists find other artists to work with, allowing fans to support artists financially etc. Jamendo (also mentioned in that article) is a shining example of this. Many software projects are supported in a similar way e.g. Ardour. This seems to work quite well for projects/artists once they have become popular enough.

    I guess it\\\’s quite similar to corporate sponsorship of free software projects, which clearly benefits both parties.

    Another model is where companies sell functionality which doesn\\\’t yet exist in a particular open source platform. The Plone CMS is a nice example of this. The company pays directly for the development work to get what they need, the development community, and other users benefit from the new features and the company benefits again by having more people using, testing and improving the features they paid to add in the first place.

    I don\\\’t have much faith in the future of the \\"Pay for the enterprise edition\\" or the idea of selling indirectly related services e.g. hosting for a CMS as being a reliable revenue stream for free software development. I hope that the idea that you can give of your creativity freely and be supported by those who want more of it becomes the predominant model, but I also feel it\\\’s a massive change of mindset and there is a need for these other models while that mindset changes.

  4. The information which you are given in your site is very useful for us. Keep it up and give more updations

  5. This is really interesting. Thanks for telling me about jamendo. It seems they have the same thing as a dual licensing model which has been used by MySQL, JBoss, etc. and which we use for opentaps: free for users, pay for commercial resellers. The issue, though, is can Jamendo support big name artists? Or will it always be for the small artists, because big name artists will find the record labels’ deals too alluring?

    By the way, don’t be too hard on the “enterprise edition” model. I agree that most people who’ve tried it have done it badly–a few add-on features plus support is just not that compelling. But there truly are people with different needs who are willing to pay, just like they really are people who pay for First Class on an airplane. For example, somebody may be from an “Oracle shop” and want their application only with Oracle DB, Oracle app server, etc. and are happy to pay for it, because their alternative is a multimillion dollar Oracle app too. In that case, it would be a good way to fund development for the rest of us.

  6. Thank you! We’ll do our best for you.

  7. Si Chen,

    I think your comment pretty well describes it.

    There is one part you miss though: If you are playing in the b2b space (unlike music stores) you MUST have a sales channel. The small and medium enterprises out there do not buy after a thorough investigation and a tender (conducted and paid by whom?) but they are mostly led by somebody or something. And this \"somebody\" needs to be paid.

    Unfortunately there is very little or no money left to give away. At least not enough to compete with \"old\" (proprietary) vendors who often pay +40% right from the first Euro spent by the customer.

    And there is more: by looking at the commercial offering of some of the better known OSS companies there is not much of a price difference compared to commercial SW manufacturers. While this difference may add up to big bucks for enterprises it is negligible for smaller businesses. But enterprises (in Europe at least) always go for Volkswagen and Mercedes rather than for Hyundai or Lexus (or differently put: by buying big names there is no risk that the purchasing decision will ever be questioned…

    Having dealt with OSS for quite some time now I do not have a recipe either. But what I usually say: There ain\’t no such thing as a free lunch. And businesses do not expect it. Make them pay a small amount, let your partners take the bigger share and go for the large numbers. But this of course requires a good hard look at the licensing model as well.

    (Disclaimer: I am probably quite well known for my involvement with SugarCRM but I am not a partner. One reason you can read further up. Another one is that I believe that partners should (possibly) pay for training and support but not for the right to sell.)

  8. Thanks for your comment. It is right on the money and realize why a lot of open source projects are drifting in the commercial/shareware direction — because we have not created something technologically innovative enough to cut through the old ways.

    The “open source sales model” was always supposed to be one where you could download and get started using something without paying for it, and then a small fraction of those users would come back to pay for additional services. For example, I downloaded MySQL, and only after I got successful using it did I go buy a book about MySQL, go to a MySQL conference, etc. etc.

    If open source business applications still require a partner to set up and install, and open source applications companies need to recruit partners, then we need to find a way to pay the partners too, and then we’re back to selling commercial editions as well.

    The only alternative is to break down the old approach with better technology: create a way for users to get up and running with business applications easily, like the way I was able to get up and running with MySQL.

    What do you think?

  9. Hi,

    Yes I see your point. But there is another thing that I forgot to mention: There is a big difference between infrastructure (web servers, databases, firewalls) and business software. Geek software is made by geeks for geeks (no disrespect). So one geek knows exactly what the other geek wants and geek no. 3 happily contributes because he has exactly the same problem. But all 3 of them are users and creators at the same time and they LOVE technology.

    Business software is different. It may be written by geeks but only if they are ordered (paid) to do so (there may be some rare cases, though – you? – where business knowledge meets programming skills). There are no contributors either because the user the software is written for doesn’t give a damn about the software itself but cares about his VAT declaration/invoice format only. And nobody cares about technology (don’t forget, in this segment there are no CIOs, just owners).

    And free is no good. Business users do not trust free software because they themselves don’t give away anything for free. So if somebody else does it they get suspicious.

    My conclusion: OSS is great, it gives you (as in user) the freedom to choose and the security that the SW will outlive the producer. But (business) OSS doesn’t have to be free. If it’s good people will pay but they will pay less than for closed source.
    On the other hand: without visionaries like yourself we wouldn’t even have this discussion.

  10. Thanks again for sharing your thoughts. Yet again they are very deep, and I’ve thought about them for several days now before writing this…

    I have seen that when businesses need customizations for their software, they often have to invest in developing those customizations themselves, and then they are often willing to share them. Part of it is to be nice, and part of it is hoping others will help or maintain extend those customizations. These are the “seeds” of an open source project…and yet a lot more is required: a strong technical framework and a development and Q/A process which can absorb them. (See http://opensourcestrategies.blogspot.com/2007/10/limits-of-open-source.html for example) This is why it’s kind of hard for these seeds to coalesce into a good project. Also unless we had a more “universal” architecture today’s business-oriented open source projects are like small islands and don’t attract nearly enough volume to get good contributions.

    But does it have to be “free as in free beer”? Hmm….. What decisive competitive advantage does being free, as opposed to being open, convey? I will have to think about that more and write again when I have some ideas, hopefully good ones.

  11. Thanks! This makes me so much more interested in opentaps. :-)

    Not only in its business model but also its technical side.

  12. Si Chen,

    Java community has no shortage of open source ERP software.

    Let me name a few: Openbravo, Compiere, ADempiere, JFire.

    What made you decide to create/invest in opentaps?

    And also, what made you decide to use Apache OFBiz as your base framework instead of the other (maybe one of the names above)?

    And regarding OFBiz, why is opentaps a separate project than OFBiz? I’m a bit confused about “what is opentaps” vs “what is OFBiz”… so both are ERP but they’re different because opentaps is “bigger than OFBiz”. So is there a reason for opentaps to “exist” independently of OFBiz.. if opentaps is superior than “vanilla” (?) OFBiz then shouldn’t these projects merge?

    Sorry for the questions ;-)

  13. Ha ha this is probably better as a separate blog post altogether, but let me try to give a couple of answer points:

    We started opentaps because we wanted something which could meet the needs of a larger range of organizations and business needs, especially in a web-driven economy. At the time, most of the other open source projects you mention didn’t exist. We chose Open For Business (the pre-Apache name of “OFBiz”) because of its data model and because I liked its technical direction at that time.

    As for why opentaps is not a part of OFBiz, open source projects in general separate because of differences in vision for the project, and we’re no exception. OFBiz has become a project of different things built on the OFBiz framework, whereas we wanted to build the best core enterprise software we could using all the possible open source projects. Their technical direction has also drifted inwards to being more database-driven using an OFBiz-only scripting languages whereas ours has drifted outwards to being more J2EE standard and domain driven. We also have different standards on user interface, usability, and stability than OFBiz.

    If you have more questions about OFBiz and opentaps, feel free to join us on the SourceForge forum.