Larry Augustin's Weblog

Thanks to Matt Asay for a great OSBC

I've spent the past two days at OSBC in San Francisco, and this was probably the best OSBC ever.  The conference was packed with many interesting people, both speakers and attendees.  I want to make sure to thank Matt publicly for putting together such a fantastic program.  The content was incredible.  This was one of the few conferences I've attended lately where not only did I want to attend the sessions, but they lived up to their expectations.  Thanks Matt for all the hard work.

One Patent Per Developer (OPPD)

Last weekend Fortune Magazine published an article in which Microsoft General Counsel Brad Smith and licensing chief Horacio Gutierrez claimed that Free and Open Source Software infringes on no fewer than 235 Microsoft patents.  The resulting blogostorm (including my own contribution) generated a lot of sound and fury, but did it really accomplish anything?

I know that it’s an obscure reference, but I’m reminded of a favorite quote from a tongue-in-cheek game I used to play in high school and college called Junta.  Junta involves taking control of a small dictatorship by battling for such important strategic spots as the bank, television station, etc. One of the event cards in the game reads, "Students Protest: No Effect". How do we turn the uproar around Microsoft’s recent patent assertions into something useful, and make sure the result is not just "no effect?"

Microsoft did take some notice of the uproar.  Bill Hilf, general manager of platform strategy and director of Microsoft's work with Open Source projects, answered some of the criticism and clarified that Microsoft’s strategy is "to license, not litigate."  But as Matt Asay blogged, that’s not entirely a comforting clarification, and the Open Source community needs to respond.

And respond we can.  Now I know that many people have issues with the current patent system, particularly as it pertains to software.  We need to continue efforts to reform the system.  But until the system is changed we need to work with the existing patent system.  Open Source developers have now for decades been contributing code to the community.  But as we are learning contributions of code are not enough.  We must fight fire with fire. We must build an Open Source patent portfolio to rival commercial portfolios.  We have a start at that already in the Open Invention Network (OIN) and the Patent Commons. Last week OIN CEO Jerry Rosenthal said that OIN stands "ready to leverage our IP portfolio to maintain the open patent environment OIN has helped create." OIN to date has accumulated more than 100 patents.  We can do a lot better than that.

Scott Collison, CEO of Ohloh, tells me that Ohloh tracks 39,664 contributors to Open Source projects.  The vast majority have contributed over the last year or two. If every Open Source developer contributed just one patent to OIN we’d have a patent portfolio that would rival most any in the industry. Famous patent troll Intellectual Ventures only owns somewhere in the range of 1,000 to 5,000 patents.  One Patent Per Developer (OPPD) - that's all we need.

Let us then take Microsoft's patent claims as a call to action for the Open Source community: One Patent Per Developer.  It's a lofty goal, and not one that will be achived easily.  But one that we can achieve.  With every contribution of code we each must be thinking, "one patent".  That's all it takes to help keep this code free - one patent.

LinuxWorld Podcast with Don Marti

I had a chance to talk with LinuxWorld Editor Don Marti recently about Microsoft's claims of patent infringement by  by Linux and other Open Source software.  The podcast is available here.

It's Time for Microsoft to Put Up or Shut Up

An article published Sunday reveals that Microsoft General Counsel Brad Smith and licensing chief Horacio Gutierrez told Roger Parloff of Fortune Magazine that Free and Open Source Software infringes on no fewer than 235 Microsoft patents.

For some time now Microsoft has been claiming infringement of its patents by Open Source software.  Gutierrez even claims such infringement is willful:

"This is not a case of some accidental, unknowing infringement," Gutierrez asserts. "There is an overwhelming number of patents being infringed."

Setting aside questions of Gutierrez’ command of grammar, I take offense at the notion that anyone in the Open Source world is willfully violating a Microsoft patent, particularly since Microsoft refuses to disclose any potential violations:

"Gutierrez refuses to identify specific patents or explain how they're being infringed, lest FOSS advocates start filing challenges to them."

Gutierrez and Microsoft aren’t interested in intellectual property rights.  They’re not interested in allowing the Open Source world to defend itself. They’re not interested in a fair fight.  Like a bully, they refuse to face the Open Source world in a fair fight, instead hinting at willful infringement and making backhanded threats.  Why?  Like any bully they fear that when faced with a fair fight in the light of day they will be revealed for the bully they are.  Like any bully they fear that which they threaten.

I work with many companies that have partnerships with Microsoft. For many years I’ve tried to give Microsoft the benefit of the doubt and watched them slowly move from anti-Open Source to an Open Source supporter, or at least a company that recognizes the importance of Open Source ISVs to the future of the Windows platform.  But this bullying has to stop now.

If Microsoft believes that Free and Open Source Software violates any of their patents, let them put those patents forward now, in the light of day, where we can all evaluate them on their merits.  If not, then stop trying to bully customers into paying royalties to use Open Source.  It’s time for Microsoft to put up or shut up.

Open Source Venture Investment Through Q1 of 2007

Matthew Aslett over at Computer Business Review has been tracking the flow of venture capital into Open Source startups and was kind enough to share some of his data with me this week as I was preparing for the upcoming Software 2007 and OSBC conferences. The data was originally collected and published by Matt Asay and Robin Vasan.

The level of VC investment flowing into Open Source is impressive and growing. Last year $481MM was raised by 48 startups. Since 2000 Open Source startups have raised over $1.9B in financing. Here are the last three years:

2004	$298MM	36 deals
2005	$306MM	41 deals
2006	$481MM	48 deals

I plotted the amount of investment and number of deals by quarter since 2000 and came up with this graph:

For the last three years VCs have been investing in about 11 deals per quarter (trending upward slightly) with increasing average deal size. The first quarter of 2007 continued this trend with $100MM flowing into 11 deals.

So much for the historical data; what’s the Open Source venture climate likely to look like for the next few years?

I think the number of investments made per quarter, as Matthew is currently tracking them, will stay flat to decline in 2007 and 2008. There are two reasons for this. First, the venture business is a portfolio business. For most funds it doesn’t make sense to focus all of your investing on one sector. Typically a venture fund makes money off of 2 or 3 highly successful investments. So a fund makes one or more investments in multiple sectors, spreading the risk across both the companies and sectors. It’s the principle of diversification. Once a venture firm has a few Open Source investments they’re not as likely to do another one lest they risk over-weighting their portfolio too much to one sector. (Of course I personally am a counter-example, being heavily invested in Open Source; So much for diversification!) A lot of firms already have their handful of open source deals. This creates an interesting dilemma for entrepreneurs. It becomes more difficult to get funding from the experienced Open Source investors because their portfolios already have several Open Source investments.

Second, the ability to make a large return on an investment in an open source company is still largely unproven and investors are going to hesitate to plow more money into Open Source until they have more evidence of success. Yes, we have a few success stories: Red Hat, JBoss, TrollTech, and SourceFire. But that’s a pretty thin list. Further, JBoss, while a success at a $350MM acquisition, was not evidence of the $1B target venture investors would like to see. Also, SourceFire has stumbled and is now struggling with a market cap just over $300MM. Again, a success, but a modest one in the VC industry and not the evidence of the ability to reach $1B that venture investors need to see. Based on the maturity of the existing private Open Source companies, I don’t think we’re likely to see another large proof point until 2008, and multiple proof points until 2009. I think MySQL is a first quarter 2008 IPO if they can successfully transition to US GAAP accounting and demonstrate predictable growth in 2007. After MySQL there are several private Open Source companies that will do more than $20MM in bookings or revenue in 2007, positioning themselves for $40MM+ in 2008 and an early 2009 IPO. That means we’re at least a year, and maybe 2, from the more significant proof points needed to encourage more investment.

But you may have noticed that at the beginning of this posting I qualified my comments with the phrase “as Matthew is currently tracking them.” The companies tracked in Matthew’s data are those that identify themselves as primarily Open Source. Where we will see an increase is in the number of companies using Open Source as an element of their strategy. These companies won’t identify themselves as Open Source companies, but they will use a number of Open Source elements in their strategy. For example, companies like UBmatrix whose model is more traditional but also provides some core Open Source components.

In many ways that’s not bad. It means Open Source may be becoming more common. Rather than being an unusual strategy, it’s becoming something that virtually every software company will adopt at some level.

HP Open Source podcast

I was thrilled to be invited to speak recently at an HP internal seminar series on Open Source.  HP has made a podcast of the talk available on their Open Source blog.

Podcast with Tom Berquist of Ingres and MR Rangaswami of the SandHill Group

I had a great conversation with Tom Berquist, CFO of Ingres, and MR Rangaswami, co-founder of the Sand Hill Group, a few weeks ago on the subject of Open Source software.  The podcast is available at sandhill.com.  Tom and MR have been around the world of Enterprise software seemingly forever, and it was great to get their perspectives on the industry changes.  Before taking the CFO job at Ingres, Tom was previously a research analyst at Citigroup and Goldman Sachs covering software and has an incredible depth of expertise in the industry.

Conversation with Dana Blankenhorn

I recently had a conversation with Dana Blankenhorn about politics, society, personal empowerment and Open Source.  I call it a conversation because I spent as much time listening to him as I did talking.  Dana is one of the dying breed of journalists that actually understands and thinks about the topics he rights about.  He's posted some of our conversation here.

Natural Revenue Growth in Open Source Startups

Matt Asay’s comments on Open Source software sales cycles prompted me to send him a long email, some of which he blogged. Yes, I’m the person behind that email to Matt.

Stephen Walli also jumped into the fray with a response to Matt’s comments about the downside of an Open Source business model. Stephen was joking with me on Friday that it’s “pick on Matt week”.

If you read my email to Matt, you’ll see that the main point is centered on the natural revenue growth rate of a startup. All businesses have a natural organic revenue growth rate. The challenge with respect to that is twofold: 1) find the product & business model that maximizes that natural rate, and 2) execute on removing whatever are the natural limiters to that growth rate. Limiters are the scalable part of that natural growth rate. For example, once you have the right product and know the right sales pitch, the limiter may be finding the right people to hear the pitch (e.g. sales leads). Or, if you have the right product, right sales pitch, and know the target customers, the limiter may be how fast you can hire and train the sales people to deliver the right pitch. The limiters in software business are typically leads or sales people. The limiters in a services business are typically the ability to deliver on the services (hiring and training the services delivery people).

When you try to force a business to grow faster than its natural growth rate you usually get into trouble; “unnatural acts” they’re called. Investors and management in any company are going to push the company to grow. That’s normal. But the push needs to be to maximize the natural growth rate, and not grow through unnatural acts simply to meet growth rate demands. It’s a fine distinction, but an important one.

Every business is different, but returning to our theme of Open Source, I have some observations common to many Open Source businesses on revenue growth rates. First, most such businesses use Open Source as a lead generation and demand creation mechanism. They use Open Source to create customer pull for their commercial offering, whether it’s a service and support offering, or a commercial product offering. This is in contrast to a push model which relies on a sales force (or channel) to introduce (“push”) your software to customers. Push and pull models have very different operating characteristics; that is, they have very different cost structures and revenue growth profiles.

I believe that one of the benefits of Open Source models is the ability to create customer pull at low cost. (The changing environment around software development and the software marketplace that has enabled this is a topic I hope to spend some time discussing in the near future.) However, along with that pull model is a delayed (or deferred) revenue growth ramp. Pull in an Open Source model is typically created through widespread adoption and usage of your software. Once the software is widely used, natural customer segments begin to appear. For example, customers that want insurance, indemnification, support, enhanced features, etc. But all those requests from users (pull) depend on their adoption of the software. So the initial focus of an Open source model needs to be customer adoption. And what drives customer adoption of Open source software? A quality product.

Returning to our discussion of natural growth rates, if the product does not have the adoption and usage to generate pull, the model won’t work. The initial temptation is often to fix this problem by creating “push” (i.e. spending money on BMW-driving sales reps to push the product). While push may be an answer, the real answer needs to come from an understanding of why the adoption isn’t there. Why aren’t people downloading, installing and using our product? How can we better understand the needs of the user, and make our software more useful to them so they will want to use. After all, we’re making the software and source code available for free. Price is not a barrier. If people won’t use our software for free, how much good is push really going to do?

One final thought as I wrap this up. Investors and management need to be aligned on the pull model. That means alignment on when revenues ramp (after widespread adoption when pull begins to kick in), and alignment on spending leading up to the revenue ramp. The longer it takes for you to create pull (and the corresponding revenue), the more money you burn through. The company needs to tightly control spending during that initial phase. The company needs to have a laser focus on product; spending on other efforts (e.g. sales) need wait until the customers have proven their interest in the product through adoption. If investors and management push too quickly, they will break the model.

Open Source Software is About Empowering the User: A Personal Defining Moment

I often speak about the implications of Open Source for software companies probably because that's where I make my living now - investing in Open Source startups and helping them develop their business models.

However, a recent talk by Mark Shuttleworth inspired me to comment on one of the major reasons for the success of Open Source among end users:  individual empowerment.  I remember with great clarity the first time I truly understood the empowering nature of Open Source.  I've been contributing to Free Software since 1985, but it was a seemingly trivial experience I had in 1995 that caused the light bulbs to go on for me.

Although my business at the time was built on Linux, we were using QuickBooks as our accounting system.  Our printers were networked (Samba), and when we tried to use a networked printer from QuickBooks, QuickBooks would crash.  After some debugging, I discovered the reason for the crash.  QuickBooks could handle 11 character printer names (DOS legacy 8+3 device names), but longer names (as were common with networked devices) were overrunning a buffer and causing a crash.

Any of us who have written code can picture the error in the QuickBooks source.  But of course, QuickBooks is not Free or Open Source Software.  I spent half a day trying to get through technical support at Intuit to explain that not only did I have a problem, but I understood the nature of the flaw in their software.  I was motivated by the desire to help Intuit make QuickBooks a better product, and simply wanted to get my bug report in front of the right developers who could make the fix.  It was an incredibly frustrating experience that left me with a great feeling of helplessness.  I knew the problem, and had the skills to fix it.  But I couldn't get through and Intuit technical support stymied my best attempts.  QuickBooks, it seems, did not support "networked use", and any correspoding bugs were simply disregarded.  Since I couldn't re-create a printer device name longer than 11 characters in MS-DOS without pointing to a networked device, the problem was not one Intuit technical support would record.

Although in many ways that one event is just a small and trivial example, it became a defining moment for me.  Although at the time I had been contributing to Free software for 10 years (since 1985), I never understood how empowering to the user it was until that moment.  To this day I look back at moment as the instance I truly understood why Open Source will win.