opening post: behind the venture capital cycle…

October 1, 2009

While I and Accel have enjoyed investing in social media, I have been relatively slow to embrace blogging. There is a saying that [roughly] goes, “If you stay silent, people will suspect you’re an idiot. If you open your mouth you’ll remove all doubt…” But I’m paid to take risk, so here goes.

One of the most common questions I’m asked “do you think the environment is getting better out there? Will startups ever make money again?” I’d answer yes and yes, but of course with caveats.

I’m fortunate to have partners who have been investing in technology for 40+ years. This is their fourth investing “cycle,” and each has lasted roughly 14 years (we jokingly call it the “Patterson cycle” after our partner Arthur Patterson who is the most articulate on the subject). The first cycle was the semiconductor boom of the late 60’s and early 70’s (Fairchild, Intel, etc), followed by the PC Boom of the late 70’s and early 80’s (Apple, Microsoft), and then the Networking/Internet Boom of the 90’s (Cisco, UUNet, Yahoo, Ebay, Amazon, AOL, etc.). In general, these cycles end with a 1-2 year “bubble” and are followed by 5 “bottoming” or “nuclear winter” years, 5 years of “firming”, 2 years of “acceleration” and then another bubble.

These cycles are somewhat independent of the macroeconomic cycle, but the peaks, valleys, and perhaps the periodicity can certainly be influenced by it. Why? Well, this requires a bit of a primer on business cycles.

All business cycles are essentially caused by imperfect information, lead/lag times in the supply chain, and something my old supply chain professor Hau Lee called “the bullwhip effect.” It boils down to this: when demand begins to increase, end distributors send ever larger orders back to their manufacturers to stay ahead of the curve and avoid stock-outs. The manufacturers send even larger orders to their suppliers for the same reason and so-on. As demand grows, these signals intensify further down the chain. When demand peaks or drops even a little, the process reverses, with the distributors and manufacturers cutting back dramatically to allow their inventories to be drawn down. Even if the change in end user demand is low, by the time the signal gets amplified back to the parts supplier, the order size change is dramatic–like a bullwhip. The entire supply chain software industry was created to solve this problem, but because the systems were not interconnected together they did little to actually address the problem (some argue they exacerbated it). SaaS software might someday provide a solution here, but that isn’t the topic for this blog.

So, how does this apply to the technology cycle? You can think of the end user or business customer as providing the “demand” for new technologies, and entrepreneurs providing the “manufacturing” or “supply.” The demand for new technologies is not a clear signal or “order” however (how does a consumer or business raise their hand and say “give me more technology companies?”). Entrepreneurs interpret demand through other means–usually through an unexpected explosively successful company in a certain space. In the current cycle (which I’d call the “broadband and virtualization cycle”) YouTube and Facebook arguably provided those demand signals in the consumer world, and and VMWare provided them in the enterprise world. The explosive success of these companies told entrepreneurs that “demand for [these sorts of] solutions is back” and the entrepreneurs set out to create startups to meet that demand and fill the new niches created by the pioneers. It takes several years for these “second wave” or “derivative” companies to be formed, funded, and reach interesting scale…hence the lag and long cycle. Eventually too many companies are created to overcrowd the few niches available (and worse, “third wave” companies are crated to solve problems created by the as yet still nascent “second wave”), consumers and enterprises become confused, saturated, and frustrated while still trying to digest the first wave. Remember your reaction to those sock puppet superbowl ads in 2000? Can you imagine the frustration of a CIO being pitched yet another $5M license for enterprise software in 1999 when he was still struggling with his failed $100M ERP implementation?

At that point, things crash and you hit the 5 year “nuclear winter.” A few cockroaches with the right model, persistence, and good financial managment squeak through and become the enabling “pioneers” of the next generation. Remember that VMWare,, and Google started before the burst of the last bubble…in 1998.

These “enabling pioneers” and “enabling technolgies” are incredibly important. They are the substrate upon which the next layer of innovation is built. It is hard to imagine an internet boom without Cisco, for instance, or a PC boom without Intel. Our current cycle was fueled by the ubiquity of consumer broadband infrastructure (both wired and wireless) and virtualization. YouTube and Facebook would not have been possible with dial-up modems (nor would the iPhone withouth advances in the wireless data networks). Similarly, many of the most interesting software/infrastructure companies today have been enabled by VMWare and Xen. “Cloud Computing” itself is really just an intelligently hosted and managed version of virtualized infrastructure.

So…welcome to the current “cycle” where it seems we’re about 9 years in, so things should be getting better (with plenty of bumps and false starts ahead of course).

The big question is how the major recession will impact this cycle. For me as an investor, I’d argue there are some major benefits. The biggest challenge an early stage technology investor typically has is that early in the cycle when valuations are “low”, visibility into adoption trends is difficult. The new enablers had not really emerged until, say 2004 or 2005. When visibility is better (say 2007 or 2008), valuations skyrocket as everyone is prepared to speculate on the trends that have become “obvious” to any idiot with a checkbook. I might argue the timing of this recession has provided an interesting opportunity. Poor public comparables, weak customer and advertising budgets, (though perhaps not actual demand), and the long-needed rationalization (culling) of the venture capital industry have depressed valuations substantially, though the technological trends are more pronounced than ever. The social graph is here to stay. So is the iPhone-like experience and the virtual datacenter. One doesn’t have to be a genius to see it. So, one can argue this is the single best point in the cycle to invest we’re going to see.

Hindsight might prove me an idiot and lots can go wrong, but that’s my best prognostication for the moment. Feel free to provide some comments

Hello world!

September 21, 2009

Welcome to This is your first post. Edit or delete it and start blogging!