Market Jitters

What's an example of how you change your expectations?

Discounted cash flow has been the Holy Grail of traditional valuation. Anyone can understand that logic. You project a company's cash flows out into the future, discount them at the appropriate rate, and come up with a valuation. The DCF model makes intuitive sense. It's the application of that model that becomes difficult. And when you're talking about Internet companies, implementing the DCF model becomes virtually impossible. I like to say that Internet valuations are based on discounted cash flows, but in a warped sort of way.

One of my colleagues did an analysis last fall that applied DCF logic to the valuations of a bunch of tech companies. She found that for a fast-growing company such as Nextel, 63% of its current value comes from cash flows that are forecast to emerge after 2008. For priceline.com, an Internet pure play, that figure was 96%! In other words, 96% of priceline's current market value is based on cash flows that are supposed to materialize after 2008. That is absolutely amazing. We can't figure out what's going to happen 2 years from now with these companies, so when we try to look as far as 8 or 10 years from now, it's just massive guesswork.

That is a crucial driver of volatility. Tech companies live and die by big, incremental margins. A small top-line shortfall causes tremendous damage down through the bottom line, as each element of the business model starts to rock like a rickety old ladder. And we have plenty of shortfalls. We analyzed 1999 earnings estimates for tech companies in the S&P 500. Some 47% of those companies missed their initial estimates by 20% or more in either direction. In other words, if analysts had been projecting profits of $1 a share, half of the time the results would be either more than $1.20 or less than 80 cents.

It's horrible! It's not because we're all so stupid -- it's because the world is so unpredictable. But if we're building models where, in the first year, we can't get the numbers right half of the time, then what's the value of taking that model out ad infinitum?

What do you do in response to uncertainty?

You change your investment model, and you look for different clues about how companies are going to perform in the future. Over the past few years, the equity market has been conducting a large-scale experiment with Internet companies. It has morphed into an upscale venture-capital market. People and institutions are buying stocks in companies at a point in their development when only VCs used to make those investments.

Those of us who consider the Internet to be a great medium for connecting the world should be thankful that this experiment has been taking place because it speeds up the process dramatically. But it also changes how you value the players.

What should we do as investors? We should do what VCs do. We need to know what kind of "growth vector" a company is on. Not just quarterly financial performance necessarily, but all kinds of data points that get at such questions as "Is this company still on a trajectory that will make it a really big business in four or five years? Is it signing up customers at a rate that it needs to? Is it holding on to its customers? Is its revenue-per-customer where it needs to be?"

Now, what happens when data indicates that a company has moved to a slower growth vector? You get a violent correction. If you decide, for whatever reason, that you now expect a company to grow at, say, 40% a year rather than at 75% a year, that has a huge impact on projected cash flows. With that kind of change, you'd also probably wind up making a big increase in the discount rate. Put those two things together, and you get a massive impact on valuation.

Let's move from money to management. How do you size up the management of a specific company?

We follow roughly 1,200 companies. How many managers in those companies have real experience dealing with changes in their competitive environment that are driven by the Internet? Close to zero. That is a new phenomenon, and it requires companies to rethink strategies, to restructure portfolios, to redistribute power internally, and to break down barriers with customers.

But most of all it requires leaders who are willing to make mistakes, to learn from them, and to try again. To iterate. Here's the management philosophy that I like to hear: "I'm willing to screw up, then to figure out what and why I screwed up, and to change it. And then I'm willing to screw up again, and do it over and over, until I get it right. Oh, by the way, there are lots of people in this organization with the power to raise their hand when something's gone wrong, and we're going to reward those people, not punish them."