Response to Gurley on Internet valuation and proxies

I just read an interesting article on Internet valuations by Bill Gurley, who is very perceptive. It's worth reading. However, I suggest here that there is a large caveat to keep in mind, which relates my love for 'option valuation' to Bill's points.

It drives me crazy that business people are so blind to the role of the unknowable and unpredictable in their view of the future, and also that managers so devalue the value of their future contribution as decision makers in any business that has flexibility to respond to that unknowable. The tools to measure and manage future uncertainty are well known, but ignored and discounted by managers and analysts alike. Perhaps looking rationally at an unknown future requires confronting squarely one's degree of control (or lack thereof).

Gurley's rant on valuation is Business School 101; his points that standard measures of businesses are proxies for the correct value is really important.

It has one serious error, but this one is subtle and made over and over again by the best analysts. The 'proper' way to value a business by discounting the entire stream of future cash flows back to the present omits two key things. For early-stage businesses, especially, these omissions from the standard model are especially important. (In mature businesses where disruptive change can be ruled out, these omissions are not so important)

The first is that a priori valuation of uncertain future cash flows need not equal a posteriori valuation. If you have exactly one cash payment that depends on a coin flip ($10 if heads, $20 if tails), the value of the business a priori is $15, but the value a posteriori has two possibilities $10 or $20. In either case, its a posteriori value will not equal the a priori value. This is why a company can be correctly valued at $5MM today, and still go bankrupt next year. The proper valuation should be a probability distribution a priori - that distribution collapses to a single number a posteriori. The variance of the distribution may differ among companies that have the same correct value.

The second error in the standard cash flows-based valuation analysis arises because future actions of management, customers, suppliers, and competitors affect the achievable cash flows. This is not probabilistic - it's better modeled by game theory. So the proper valuation always is with respect to an assumed future context. In game theoretic terms, a profit-maximizing competitor may not always optimize by minimizing your profit. In fact, in some markets, there is a benefit for all to cooperate (coopetition) and in others, there is a tragedy of the commons where each participant is incented to move in a direction that will ultimately destroy the goods for all. (this, by the way, is why libertarians' simplistic philosophies may be doomed to failure). As a result, active, flexible management of the company adds a great deal of value to any strategy.

A proxy measure that helps with both of these errors is "option value", which captures the value of the options (potential future decisions) that a business holds.

A very large part of the a priori value in Internet businesses is "option value" - creating the opportunity to be in the right place at the right time and boost investment, or to be able to shut down or defer plans if unanticipated costs or market delays crop up. Option value captures in a single number the combination of the a priori uncertainty and the value of the company's power to make choices that affect its destiny as it navigates the unknown.

Many Internet businesses have huge option values, derived from their ability to respond to the vortex of change that surrounds the Internet's technology and its ecology of businesses.

© 1998 David P. Reed.