Milestone Group Quarterly: April 2005

Articles

• Face to Face: Kim Polese, CEO of Spikesource
• Investment Viewpoint: Stephen J. Harrick, Managing Director & General Partner, Institutional Venture Partners (IVP)
• By Invitation: John Seely Brown & John Hagel, III - IT Platforms for Business Innovation
• Milestone POV: Kathleen Gilligan on Aligning Sales Vision and Process

Face to Face:

Kim Polese, CEO of Spikesource

Milestone: SpikeSource is a new company. Talk about what problems you are solving.

Polese: We are solving the biggest problem facing the use of software in general, which is interoperability. Software, by definition, is complex. Open source is simplifying the process of building software applications and implementing software in production environments, but it’s also creating new complexities. The complexities come from the abundance of open source components and the speed with which these new companies, big and small, are picking those components up and starting to put them into production environments.

When that happens, you can very easily and very quickly have literally hundreds of open source components that comprise business-critical applications that are increasingly running the company. Those components are constantly changing and evolving: bug fixes, new features, patches, security vulnerabilities. Every time IT administrators download a patch, they have to test it against all the other components in that software infrastructure and in the stacks of software and go through a process of sort of testing, certifying and validating that it will not break everything else in the stack. They then have to push out the patch and push out the update and doing that on an ongoing basis turns out to be a pretty big problem -- a pretty big redundant problem when you start looking at open source use in production environments.

We’ve created an automated test climate that can aggregate literally hundreds of components and run hundreds of thousands of tests nightly, across multiple language run times and multiple operating systems, and ensure that all of the moving parts work together. A good analogy is that we are creating an underwriters lab for open source, focused on that one problem of interoperability.

Milestone: What are the two or three strategic trends you are currently seeing in enterprises adopting open source stacks?

Polese: First is platform independence. This means the ability to make a choice rather than being locked into a silo. The software industry thus far has really sort of operated in a model of these vertical silos where everything had to be integrated, in a vertical sense. The whole task of the vendors was to capture customers, lock them into a platform and keep them there. And there resulted vendor wars and that’s sort of been the history to date.

Well, open source suddenly blasts through all of that and throws the doors wide open, for the first time, putting the customer in the driver’s seat and giving them the choice of what platform to use. So really what’s happening is the software industry is becoming flooded with commodities of fundamental building materials.

The second trend is cost savings, from the standpoint of license cost as well as support costs. You now have not just five people in your IT department working on something, but 50,000 people or more out there on the Internet who understand how that code works, who are experts, who are contributing bug fixes and new features and patches and you don’t have to pay their salaries.

That presents very compelling cost savings. It’s not just the license costs, it’s not just the fact the software is free; it’s the fact that the support part of it is also greatly reduced in price.

The third trend is flexibility. Flexibility means the ability to use what you need for the task at hand, but avoiding having 80% of the features being unused because it is an over-engineered product - a typical problem for the traditional software business. With Linux, you can scale up or scale down. If you are building a mail server, you use the features that you need, but you don’t need a GUI interface, you don’t need sort of the front-end piece so you don’t use that code. So the flexibility that open source offers, the fact that you can just take the part that you want and use it and not use the rest, is a compelling advantage as well.

Milestone: Open source adoption in the enterprise still seems nascent. When do you project it will really be mainstream?

Polese: It seems nascent, but actually it is more mainstream than it appears to the outside perspective. That’s because there is a lot of fear, frankly, particularly in large companies talking about their use of open source and the degree to which it has become a business critical platform for these companies.

Fear comes from concerns around legal liability, concerns around lawsuits and sort of the F.U.D. around the perceived IP risk. It also comes from the fear of exposing that company publicly. For instance, if a question is posted on a public news group about a critical technical problem in a production environment, that poses potentially some liability for that company. What we are finding as we go around and talk to Fortune 100 companies is that it is surprising, and actually quite astounding, to see the degree to which it has become mainstream and the contrast to the outside perception.

That said, there are two main elements that are slowing down the adoption of mainstream use of open source, which can be removed and are in the process of being removed. The first is the intellectual property issue. That’s a complex issue, but there are a number of solutions and approaches to it that are making some headway in addressing the fear and the risk that exists.

The second is this problem of interoperability. If you are using dozens of components, you have to keep up to date with all of those different components. And they all come from a different source, and they need to go through this manual repetitive task, which is overhead. This is where Microsoft has done a very good job of positioning open source as being potentially more expensive than using Windows. It is more expensive if, in some cases, you have to go through these repetitive tasks of certifying, validating, testing and updating and doing that continuously - but you can automate that. So the automation of that problem, solving that problem, will be a great accelerator of the mainstream adoption of open source.

Milestone: The economics of building an enterprise software company have to date been finely built on this notion of proprietary IP providing competitive advantage. The more you have in the product, the more you can charge. How do the economics of a software company shift when a large portion of that underlying code is based on open source?

Polese: The fundamental problem is the commoditization of the software industry. The market is being flooded with these commodities. A good metaphor is the construction industry. You have lumber, you have steel, you have cement, you have these core components that are commodities. And yet, there is a thriving, five-billion-dollar construction industry, despite the fact that the core components are basically free, abundant and undifferentiated. And that’s because companies have sprung up to, for example, take those products and create composite products, which are rebar or flooring, which are stronger, more flexible, provide greater features and competitive advantages to the companies that are doing that.

In addition you have got a whole ecosystem of companies and professionals like architects and plumbing supply systems and HVAC - you name it, literally hundreds of industries exist because of the abundance of these raw building materials. The same thing is happening with software. That’s what Linux is about - Linux is like wood or steel. MYSQL, the databases and the other core building blocks, like the application server and the web server, are essentially fueling what will be the growth of a whole new dimension of the software industry.

So far we have been so focused on these core infrastructure building blocks and how to extract the most license dollars out of that building material that we have not even begun to contemplate where software might ultimately go. In fact, I believe that we are about to embark on a building boom in the software business, and it’s because of commoditization, which is actually a good thing. Far more money will be made because of using open source, than lost because of it.

Milestone: An argument can be made that most technology buyers bought the open source solutions can be construed as cost-conscious buyers. We are already in a market with lots and lots of cost conscious buyers. So is commercial enterprise software based on open source fundamentally a low-margin business?

Polese: Not necessarily, no. This is a nascent market so we are just now seeing the first business models emerge. There is more than one business model. The margins don't necessarily have to be low either.

The other thing to keep in mind is that the population of ultimate end users is enormous. Software is becoming part of everyday life - it’s all around us. It’s just beginning, this penetration of software in your stereo, your refrigerator, your phone, your car, wherever it happens to be. So no, open source is not necessarily a low-margin business and we are just beginning to see some new business models emerge that will prove that out.

Milestone: There seems to be two basic revenue models for open source players like yourself; the dual license model and the support model. How does SpikeSource plan on making money?

Polese: Those are two of what we believe will be multiple business models emerging for open source. How we will make money? We are a technology company. We are not a component provider, we are not a systems integrator, we are not a professional services organization. We are a technology company. Our technology is this automated test harness and we have three main revenue streams.

One is a subscription service. We provide those aggregated stacks of open source components for free, to download on our website and a variety of different combinations and components. Customers can choose what they want. They can download the software for free but they will pay a monthly or annual subscription fee for the ongoing updates to those stacks. We will ensure those stacks continue to work, with all the latest bug fixes, patches, new features, all completely tested, certified, validated. Anytime there is a new bug that appears, we will be the first to find it, aggregate it, put it in our build and ultimately make sure that the stack is up to date with patches delivered to the customer site.

Our second revenue stream is a managed testing service. So ISVs that want to port their software and run their software in an open source environment - and don’t necessarily have Linux expertise or don’t want to staff up and build out a whole lab with all these different combinations of hardware and software components - can come to us and we will provide that service for them. We’ll test their applications against a set of open source infrastructure.

The third source of revenue is enterprise support. We offer service level agreements, 24 x 7; gold, silver, dedicated support. It’s everything a Fortune 10 company would expect from, say, an Oracle.

Milestone: What trends do you see for other software companies to port their applications over to Linux and is this a viable co-strategy for them?

Polese: We are talking to a lot of ISVs right now and what we are hearing from them is that they are getting demand from their customers to have that software, whether it is CRM or Business Intelligence or what have you, running in an open source environment. This is because their customers are moving to Linux and they are putting pressure on the ISVs to make sure that they are software certified.

This is creating a business opportunity for us and of course it’s also creating new opportunities for them. For one thing, it’s expanding the scope of platforms on which their software can run. Secondly, in many cases, it will help them save money because they can incorporate open source components in their own free software, which adds value, functionality and new features without having to spend money on software development. They can also increase the size of their margins, basically for their license revenue. As those margins are shrinking, there is opportunity for cost savings that can help maintain some of those margins.

Milestone: What do you make of the significant move of many parts of Asia towards open source, particularly government-backed initiatives in places like China?

Polese: These are very interesting for a couple of reasons. One is the green field opportunity itself. These companies are like an unpopulated planet that suddenly is becoming populated overnight and they’re crafting their IT strategies from the ground up and ultimately deploying to billions of people. That’s an enormous overnight change.

The other thing is the fact that these governments are, in fact, taking the lead to recommend software platforms. Normally governments decree that roads should be built or taxes should be levied, but they don’t normally decree that open source should be used. However, they’re doing it because they see the flexibility, the platform independence and the ability to take Linux, in the case of China, and create their own distribution on it.

Red Flag is their version and they have greater control or consistency across all the different applications that will run on that distribution of Linux. They can be part of the broader ecosystem, but then they can create their own ecosystem of startup companies, or of software companies creating new applications that will run on top of open source. They are simply seeing the same advantages enterprises are seeing: platform independence, cost savings and independence.

Milestone: What’s your take on Microsoft and open source? Do you think they can adequately respond to the threat?

Polese: Microsoft is a smart company. In the past, when they have come to the conclusion that there is change that they can’t control or that is just inevitable in the market, they adapt readily and in fact find a way to make that work to their advantage. I have every confidence that Microsoft will do that in the case of open source. They have sort of begun on more of an individual level, or a project-based level, to participate in some discussions about emerging open source projects.

That’s not something that is sort of generally known or seen yet, but there is evidence that they are participating in the discussion. I believe Microsoft will continue to innovate in their business model, which is what they have always been very good at and they will move up the stack and take advantage of that.

Milestone: How about IBM? They seem to be making great strides in their Open source strategy.

Polese: IBM has been a very important supporter of open source adoption in the enterprise and has really validated open source and Linux in very important ways. That said, they still have a pretty big $12 billion dollar proprietary software business. Some of their software is competitive with open source software; the WebSphere application server and the DB2 database being examples. So their strategy is like many other companies; they are both embracing open source but in some cases dealing with some competitive issues that open source has created.

IBM is a company that understands that ultimately, the commoditization of the software business is not a bad thing. It creates a new opportunity - specifically in the area of support and services wrapped around that software. So I think, just like Microsoft, IBM will continue to innovate in their business model and will continue to be a leader in embracing open source and finding ways to make it work for their own business model and also in their customer’s environment, in combination with proprietary software.

Milestone: What’s the best method you use to leverage your VC’s operationally? How do you get them working for SpikeSource?

Polese: We are fortunate to have a wonderful venture firm backing us, Kleiner Perkins Caufield & Byers. With them, we have a great team of professionals and also an extended network of executives who they know in the industry. They are very helpful in recruiting and building teams for companies in the early stage such as us. They also have a really good, broad view of the marketplace because they have been focused on building that ecosystem - not only companies, but also partners and networks of people in the industry. They are very instrumental just in the ongoing discussions we have about how this market is evolving and shifting and how we can ensure that we are building the right business model to take advantage of that. Ray Lane, Russ Siegelman from KPCB and now Dave Power at Fidelity are all very actively engaged. They are not just showing up at a board meeting. We are constantly talking with them on a daily basis.

Milestone: What’s the worst-case scenario for SpikeSource and what keeps you awake at night?

Polese: The good news is that the opportunity is huge. The problem that we are solving is, as I mentioned, one of the two critical problems that I believe faces the mainstream adoption of open source in the enterprise. We are out in front, specifically in this area of testing and certification of open source software.

That said, I think the enormity of the problem also create challenges, on of which is scale. We literally will be running hundreds of thousands of tests across hundreds of different components, ultimately probably thousands of different components in multiple language run times and operating systems. That is inherent in every business - you have a core IP strategy that you are executing on. So we need to ensure that we are building the most scaleable, most powerful automated test harness and doing it in a way that also engages the community.

That’s the other piece of how we get scale. Innovation and process automation is half of the puzzle. The other half is developing a thriving ecosystem of partners and developers who are contributing test results, knowledge, information and helping really build out the products and services that we are offering to the market. Tim O’Reilly calls this the architecture of participation. It’s the way that I think companies will effectively operate in the future. Every company needs to figure out how to make their customers and users part of the product and service that they are building, and contributing actively to it. That’s also a big undertaking. If anything, scale is the thing that I think about on a daily basis and it’s something that we are building into every part of our operation here at SpikeSource.

Kim Polese, CEO

As CEO of SpikeSource, Kim Polese is responsible for guiding the company’s business vision: to make open source safe for the enterprise. Prior to SpikeSource, Kim co-founded Marimba in 1996 and as President and CEO she led the company through a successful public offering and to profitability. She then served as Chairman until Marimba’s acquisition by BMC Software in April 2004. Before co-founding Marimba, Kim worked at Sun Microsystems and was the original product manager for Java, leading its launch in March 1995. Prior to Java, Kim worked in Sun’s software division on object-oriented development environments. Previously, she worked at Intellicorp Inc., helping Fortune 500 firms implement expert systems.

Kim earned a Bachelor’s degree in Biophysics from U.C. Berkeley and studied Computer Science at the University of Washington, Seattle. Kim serves on the executive council of TechNet, a bipartisan coalition of executives focused on the growth of the technology industry and economy, on the board of the Global Security Institute, and the University of California President’s Board on Science and Innovation. She is a Fellow at Carnegie Mellon University’s Center for Engineered Innovation.