How Do VC’s Think About Open Source?

Question: How does the VC community deal with open source? I am working on an embedded device that can be built either way, open or closed. This decision is complicated since our group (due to past successes) is capable of funding the first few rounds internally. We won’t need external VC until the last round in order to enable full production. Which product would the VC rather see, SlimDevices (completely open) or the Zune (completely closed)?  Actual product is neither of those devices. Is the creation of a user/hacker community worth the exposure of having the source code open? What does a VC perceive as the downsides of open source?

Before I take a crack at this, I want to point you to an excellent post that Will Price at Hummer Winblad posted yesterday – The Three Most Important Letters in Open Source: CYA.

Like most things in software – some VCs get it and some VCs don’t.  Now that open source has become mainstream – initially through the commercialization of Linux but also the success of VC-backed companies like RedHat, JBoss, and MySQL – there is real history and precedent around how to create a successful open source company.  In addition, there are a set of VCs (think “individual partners, not firms”) that have a positive track record helping fund and build open source companies.

There are several different layers to the issue.  The first is the most basic – many companies use open source components as part of their proprietary products.  Several years ago VCs and entrepreneurs got tuned into the risks associated with this due to several absurd lawsuits – such as the SCO / IBM one – that created an uncertain potential future liability for these companies, their customers, and any larger company that acquired them.  As a result, there was an immediate irrational buyer backlash against open source – we sold several companies to large public company acquirers that had very tortured negotiations around their open source IP – usually resulting in unnecessarily complex IP representations in the purchase agreement, actually code modification as a condition to closing, or fundamental challenges in getting a deal done.  As people began to understand the liability dynamics better and products for evaluating source code for open source emerged (including several that are VC backed), this started to settle down and today is in a much more rational zone.

Next is the actual creation of commercial company around an open source product.  Many successful open source products have a small number of key architects (and often one “king of the project.”)  A relatively small set of people have figured out how to effectively commercialize the companies and – as a result – you’ve seen some impressive businesses built around open source projects.  The key phrase here is “a relatively small set of people” – the vast majority of software VCs don’t really understand open source in any great depth – especially when you get into the intersection of community, legal, and commercialization issues.  As an entrepreneur – this is yet another case where you should make sure you are filtering your potential universe of funders carefully and trying to evaluate early whether or not someone will get it.

The question asks very directly “what’s the downside” and “is the creation of a user / hacker community worth the exposure of having the source code open.”  These are tough questions to answer because they are circumstantial in nature – most commercialized open source projects either emerge from (a) an existing open source project that becomes popular and reaches some critical mass or (b) a specific product that has a software component that can be enhanced by active user involvement in the build / development of it.  Companies in category (b) are particularly interesting, as they have open source software as a strategy component of their business, usually driving commercialization theme that is attached to the software (e.g. a hardware device that is user configurable / programmable.)

Open source is undeniably here to stay and a critical part of the software ecosystem.  However, as the question forshadows, even though open source (and its father – “free software”) has been around for a while, the mainstream software industry is still wrestling with many issues surrounding it and we expect they will continue to for a while.  Herein lies great opportunities for entrepreneurs and investors.

  • Very good question and analysis. As the senior parter at an open source management consulting firm and fairly embedded in the VC community we have conducted due diligence on a number of open source investments over the last couple of years and work with dozens of VC’s portfolio companies. We’ve seen a definite positive trend in the open source knowledge of investors in our experience starting in the later half od 2006. No longer do most VCs look for the old “eyeballs” like measurements of volume of software downloads and equate that to “investibility”. Today they understand a more accurate measure is conversion of those downlaod numbers to registered community participants or direct revenue of some kind. But most importantly, we think, is the examination of a company on it’s own merits without open source in the equation. Could this company stand on its own without an open source model per se and use open source to enhance an already strong team and product? A good example of this is Cleversafe, they have a good team, a good product that could change the storage market, and oh by the way, a wonderful opportunity to utilize open source to leverage and enhance a fundamentally good opportunity.

  • I agree — good post. I also echo Andrew’s last comment. It is increasingly unlikely that any company (in a technology industry or otherwise) can avoid open source entirely. As a result, at least on some level, nearly all technology companies are using and benefiting from open source in some way, whether they know it or not. As this reality continues to sink in, the focus is shifting toward evaluating and understanding both the real risks and real values presented by open source in much the same way any other aspect of that company’s business would be evaluated.
    From my vantage point as an attorney involved with open source licensing and compliance issues, I think part of this has to do with the fact that strategies are increasingly available for dealing with the more thorny open source legal issues. Likewise, on both a legal and practical level the open source ecosystem now offers much of the same infrastructure available in the proprietary world (compliance, certification, service, support, indemnification, insurance, etc.). As a result, legal and practical risks continue to decrease and I see more clients seriously looking at how open source can benefit their value proposition and overall success. This can mean leveraging open source in the design of a proprietary product/system or actually implementing an open source business model around a product. It has also started to include the adoption of certain aspects of open source licensing and open source communities in hybrid proprietary-open source business models. But, in all cases, the focus is on what real value can be delivered by open source, not the mere fact that open source is involved.

  • A friend just dropped me a note saying that Microsoft has recently articulated their view on how they approach acquiring companies that have open source in their code. The summary of their current rule is “we will acquire companies using open source if the open source is not the mission critical aspect of their product AND if they have thought through a migration path away from that usage, that we could employ if we bought them.” Microsoft’s (and other large companies that are active acquirers) will have “rules” and these “rules” will change regularly – it’s useful if you are on the selling side to explore and understand them early in the process if you are heavy users of open source.

  • Jon Smirl

    There are two fairly independent worlds using open source. First there is open source software, like JBoss and Redhat. Second there is the embedded device world. In the embedded world there are basically two choices: Micrsosoft WinCE or Linux.
    The hardware for embedded devices is proprietary and can be patented. The software is generally useless without the hardware. So the danger from being open is that someone will clone the hardware and then escape the software development costs. But that’s what Linux is all about, escaping the software development costs. It is risky as a vendor to clone a someone’s complex piece of hardware then run software on it that your internal developers don’t know how to fix. In the market this doesn’t happen very often, look at all of the network routers and MP3 players – all of their software is somewhat different.
    As a developer there are major downsides to WinCE. You are trying to get it to run on a new piece of hardware. When you have problems you are forced into calling Microsoft and praying that will fix problem on time, if at all. In my experiences with them this is a 50/50 crap shoot. If you are buying $100M in licenses all your problems get fixed. If you are buying $1M they won’t return your phone calls.
    With Linux you are totally in control of the user experience. If something doesn’t work right all of the software needed to fix it is available for modification. I only need to point out the famous case of the Thai ambassador that got locked inside his car and couldn’t get out because of a WinCE crash. The auto maker didn’t have the source and couldn’t look for problems.
    But choosing Linux for an embedded device means that most of your code is going to be open. It is possible to hide some key parts, but the bulk of it will be open. Since your device is open, the thought is to take advantage of this and create a community around it. A recent example of this is the Nokia 770/800 and

  • The secret sauce is no longer the software anymore. The open source software projects can easily beat the crap out of proprietary software anyday. And for entrepreneurs like us, it makes perfect sense to use Open Source software as much as possible. And most of the software out there is dual licensed or LGPL’ed or under a BSD style license. Its amazingly simple nowadays to build a great product.

  • Good points were raised in this post. I thought readers of this post might be interested in one of the first open source community software projects for an embedded device – all the way back in 1981. The results were impressive: