introduced to the CEO of a fabulous operating company called The Genuine Parts Company. Genuine Parts is a New York Stock Exchange Fortune 250 company, but a very quiet company that you may be less familiar with. Their biggest business is NAPA Auto Parts and then they have an industrial products business and an office products business. I thought that having some real operating experience was going to be very helpful to my career. So I joined them in the mid ’90s and did a variety of things, but found myself increasingly helping our business units navigate the technological changes and the business model implications brought on by the emergence of the Internet. As a result of that, I spent a lot of time with venture capitalists and venture-backed companies that wanted to be our customer or partner or to co-invest in something with us. Through that I got to know many folks in the west coast tech and venture community. Madrona was one of those groups and one of its companies ended up being a customer of our Office Products Division. Then through that relationship, Madrona, NAPA and Amazon (where my partner Tom remains on the board) at one point explored doing a joint venture This was back in the late ’90s and ultimately we all decided the timing wasn’t right for an online auto parts store and didn’t do the deal. I got to know the Madrona team better and after they raised a fund in late ’99, they asked me to be a part of the team.

Angel Mehta: What’s the toughest experience you’ve ever had as a venture partner at Madrona, be it with one of your portfolio companies or raising capital?

Matt Mcllwain: The most challenging time? Madrona’s ’99 fund was our first outside investor fund. Like everybody, we made a fair number of investments with the 1999 fund and 2000 – 2001 brought everyone back to reality. We shut down about five companies in 2001 and that was hard. But we doubled-up on companies like Classmates Online and Sharebuilder that ended up being big winners for us. Also, we continued to make new investments in companies like Isilon, World Wide Packets and Impinj that are doing really well now. So I think the toughest thing was that process of prioritization of the portfolio and then having to come to a decision that you’re going to shut down some companies that maybe in a different era could have made it.

Angel Mehta: Tell me about one of your most exciting investments right now in the software world.

Matt Mcllwain: One of them is Isilon Systems, a company focused on clustered storage for digital content applications. We met the founders, two very bright engineers out of Real Networks, back in early 2001. Their vision then was to provide clustered storage for digital content. They built compelling products and were able to win NBC’s business for the Summer and Winter Olympics Ofoto, Comcast, Myspace and many others. The core idea was that digital content requires larger files and larger volumes of files, which fundamentally have a different need in terms of storage. Isilon built a distributed file system deployed in a clustered architecture for digital content applications so as the market grew, the storage capacity could grow with it. Fast-forward five years and that’s a company that has hundreds of customers and should generate almost $100 million in revenue in 2007 with strong margins. For us, the common thread is to invest in a company at a very early stage, roll up our sleeves, and help build the company in to a success story. My excitement for Isilon is that they stayed focused on an emerging market, executed very well on building a line of products and now have expanded in to other areas like oil and gas exploration, bioIT and digital document work flow in traditional industries worldwide.

Angel Mehta: What about some of the concerns you had prior to doing the deal and how the entrepreneurs helped you get over those fears and finally get the deal done?

Matt Mcllwain: Well, there is a general set of things that we look for in our companies. We know, because we invest at such an early stage, that the team is not going to be complete. So on the team front, we’re trying to make an assessment of the quality of a team today that is going to be able to, in the appropriate sequence, attract more quality members to the team. I tend to ask: Is there potential to build an ensemble cast? You need to start with very high-quality folks in their respective skilled areas in order to attract the rest of the ensemble cast over time, so making an assessment of the team with regard to that.

Secondly, is this a market that’s big enough to matter? Generally speaking, we’re looking out there for a market that has both the potential for an early stage upstart to get their foot in the door to have a good outcome and where there is something that can take it to the next level so that you can turn a good outcome into something that’s truly great. This approach has helped Madrona continue to generate positive exits 31 over the last ten years, in both up and down market cycles. There are companies, like an Amazon or more recently an Isilon, that have been able to execute initially on a focused strategy and then expand out into bigger markets. The market piece is always a big one that we look for and ask: What’s the early strategy that you’re going to stay focused on? What is that core kernel of differentiation that you’re going to be able to bring to market so that you can at least execute on the first part of it which is that there can be a good success here? So, you have to believe in the team, the market, and in the core differentiation.

Angel Mehta: How is the enterprise software market as a whole? With SaaS becoming the emerging model and white space evaporating, are opportunities for software start-ups few and far between…true or false in your mind?

Matt Mcllwain: Well, what I think is true is that many of the foundations of how software gets built, sold, distributed, managed and priced are all changing for enterprise software. It really makes people nervous to see so many variables in play at once. So I think it’s less about software as a service per se but pricing is going from an enterprise license in a lot of cases to a subscription model. The traditional channels for selling software in to enterprises are likely to change as a result of new software delivery, service/support and pricing models. If you’re meaning specifically the enterprise customers, there isn’t the same kind of symbiotic big consulting implementation partner going in alongside a big packaged application partner to take a solution into an enterprise. That’s not the way much of the enterprise software is getting sold today. You also have this whole question of how does that software then get physically deployed? Is it more of a Salesforce.com model, a download model or a classical deployment model? I think that’s going to depend on the type of application and also whether it’s more of an infrastructure technology or an application technology.

Then the final issue, of course, is how does open source fit into all that? Open source, in my mind, is basically saying we’re going to lower the cost of goods sold on the software component of how you deliver a valuable solution into an enterprise. Strategically, open source disaggregates the software from the service and support. There are many variables changing right now and I think that causes people to be reticent to try to invest in the enterprise software space.

Occasionally, a great team comes along with a very clear focus and iConclude is one. iConclude was a group of folks out of Mercury Interactive and some of them had worked at a successful start-up we backed called Performant that Mercury had bought. iConclude’s focus is problem resolution management for the IT data center. They built a technology that allows you to triage and solve problems in an automated way and then if you can’t solve the problem in an automated way, get it into a more automated process then in the past for problem resolution. But they’ve built this platform for visual automated repairs that they can go out and get repairs from the IT community. They partnered with Amazon’s Mechanical Turk service to get data center problem resolution experts to write these flows that solve repair problems. That’s a compelling notion of community applied to enterprise software and so here’s a strong team, big market, there’s tons of pain in trying to resolve problems quickly and effectively in the increasingly automated data center and a clever notion of how they can be successful. In fact, Opsware liked the concept so much that they decided to buy iConclude this year for a substantial gain.

Angel Mehta: I want to ask you about the opportunity offshore. When do the emerging markets like India and China become appealing for their corporate purchasing power? I’m curious to know if venture investors talk about the possibility that these businesses overseas are dramatically under-automated, probably are going to need the same kind of enterprise software that businesses here have already deployed so that would represent a big market opportunity? Is that something that ever comes up?

Matt Mcllwain: It does come up some in a macro sense…in the terms of where are you going to find customer growth over time. I don’t think those folks are going to buy enterprise package applications the way that people were buying it in the ’90s. I think customers in developing markets are much more inclined to buy a software-as-a-service type product that is priced in more of a subscription model that is web accessed in one form or another. This is especially true for applications and even to some extent certain types of infrastructure software products for the enterprise, but I guess I think about them as more of a potential customer segment over time for early-stage companies based in the US. Generally speaking, in commercial software, it’s good to stay somewhat geographically focused when you’re trying to get from zero to a couple million in revenue because you absolutely want to make sure that your first 25 customers are raving fans of what you do and that’s considerably harder if that raving fan is in India and you’re trying to service them well from Seattle.

So even companies like iConclude have some interesting international customer wins that really almost came over the transom. Somebody found out about what they were doing, liked it so much and bought the product basically through a telesales organization, which is encouraging. But I think most early-stage companies can and should be geographically focussed in the early years. When you get to a later stage, the stage Isilon is in now, then yes, you need to be intentional about expanding into Asia and Europe. In Isilon’s case, they actually chose to expand into Asia before expanding into Europe. I think the more conventional model would have been the other way around but the team had some relationships that had some success there in prior lives.