Low growth, high "q"

"Arguing in My Spare Time" No. 2.17

by Arnold Kling

Oct. 9, 1999

May not be redistributed commercially without the author's permission.

In venture capital parlance, an "elevator speech" is a concise description of a business. The expression comes from the following thought-experiment: imagine you were an entrepreneur in search of funding, and you happened to get into an elevator with a venture capitalist. In the time that it takes the elevator to reach the floor where the venture capitalist is getting off, what would you say to convince him or her to invest?

The elevator speech is an indicator of focus and preparation. In order to give a good elevator speech, you must be able to boil down your business to its essentials. You need to have identified your customers, your value proposition to those customers, and your method of earning revenue from the value that you provide. If your strategies and plans are vague, you will not be able to give a good elevator speech.

When Yahoo.com announced its earnings for the third quarter, they allowed anyone on the Internet to listen to the conference call with their stock analysts. Earnings had surpassed expectations, so the analysts mostly praised "TK" (I assume this meant Tim Koogle) and asked non-controversial questions. It was like a bunch of 10-year-olds gathered around Michael Jordan asking, "Gosh, how do you score so many points?"

Had I been permitted to be a skunk at this particular garden party, I would have asked, "What’s your elevator speech?"

The original "Yahoo hot list" was a community-powered web index. It provided an expandable framework within which web users could recommend useful sites to one another within the context of a hierarchical directory.

The Yahoo hot list had some flaws, in my opinion. One challenge was that as the Web became more commercial, the entries in the directory were made increasingly by web site owners rather than by unbiased individuals. Another challenge was that as the scale of the Web grew, the lists of sites in the directory became large and unwieldy.

In theory, Yahoo could have attacked these problems directly. To address the need for editorial review and objectivity, Yahoo could have developed "collaborative filtering" methods. To address the scale problem, instead of attempting to catalogue the entire Web, they could have become a "directory of directories."

Yahoo could have stayed focused on its original model, which was to enable Web users to help one another find useful Web sites. Yahoo could have made money both from advertising and from creating a "premium version" for which consumers would pay a subscription fee.

In other words, Yahoo could have continued to try to be the best at one thing: providing a directory on the Internet. Instead, they chose to broaden their business model.

Yahoo competes in online classified advertising, which in an interactive medium has evolved naturally to incorporate auctions. However, eBay is the market leader.

Yahoo competes in offering web-based email, which is one of the first examples of net-centric computing. However, Hotmail is the market leader.

Yahoo competes in offering consumers the ability to put up web pages. They did this by purchasing Geocities, which at the time was the market leader.

Yahoo competes in offering community interaction, such as games, chat, and clubs. However, it is not clear that these categories ought to be aggregated. If I want to play bridge, should I go to a generic "games" site, or to a site that is dedicated to bridge? If I want to share vegetarian recipes, should I go to a generic "clubs" site or to a site dedicated to vegetarian cooking? My opinion is that the specific sites should win out over the aggregates. Regardless of that, to the extent that the aggregation model is correct, AOL is the market leader.

Yahoo sees its best opportunities for revenue growth coming from two areas. One is content aimed at markets in Europe, Asia, and Latin America. These markets already are contributing more growth to Yahoo’s page views than the nearly-saturated U.S. market. The second opportunity is to become more directly involved in electronic commerce, going beyond banner advertising to transaction processing, and garnering a share of merchant revenues.

In short, Yahoo’s business has become much broader in scope and much more complex in the past few years. The same phenomenon has occurred at Excite.com, Amazon.com, AOL, and many other well-known Internet companies. Why is this happening?

Recently, I attempted to forecast the growth in systems development costs for our homefair.com web site. Borrowing terminology from the "discipline" (I use the term loosely) of scenario planning, I came up with two drivers for our cost.

The first driver is growth in the user base for the Internet in the United States. If this were to continue to grow rapidly, then our existing applications would get increased use, and we would be under less pressure to develop new applications.

The second driver is Tobin’s "q" for Internet stocks. James Tobin earned the Nobel prize in economics. One of his contributions is the insight that firms embark on expansionary investment when the ratio of their market valuation to the replacement cost of their capital is greater than one. He called this ratio "q." With the buoyant market for Internet IPO’s, the value of Internet stocks is much higher than the "replacement cost" of launching a new Dot-com venture. There is a big incentive to launch new Web companies. As long as "q" for Internet stocks remains high, our web site will face more competition, which will put us under pressure to create and enhance our applications.

With these two drivers, using scenario-planning’s crude "either-or" rubric, we have four cases:

1. High growth, low q. With (U.S.) usage growing rapidly and a weak stock market holding down competition, we could get by with a relatively low budget for developing new systems.

2. High growth, high q. With usage growing rapidly and a strong stock market, we would want to increase our systems budget in order to add to our value.

3. Low growth, low q. With usage growing slowly and a weak stock market, we would want to focus our systems budget on applications that generate more revenue from the same size customer base.

4. Low growth, high q. With usage growing slowly and competition funded heavily, we will need a big increase in systems development. We will need to enhance our existing applications so that they generate more revenue, and we will have to build new applications to stay ahead of the competition.

We seem to be in the fourth scenario. Under that scenario, everyone is being driven to make their business more complex. They need to find more ways to get revenue from a given audience. And they need to fend off new competitors with new applications.

It is my suspicion that this complexity-increasing behavior is maladapted to the Internet. My intuition is that talent is highly diffuse, and that the nature of the Internet promotes decentralization and disaggregation. Ten companies, each trying to implement its own application, will be more effective than one company trying to implement ten applications. As I said in "Grover and the Silver Bullet," I believe that diseconomies of scope are large and important.

If this intuition is correct, then the high value of "q" is distorting the economy. The stock market is telling "TK" to try to be all things to all Internet users. It is telling new entrepreneurs also to think in grandiose terms. However, smaller firms with more modest goals might deploy talent more effectively.

Eventually, I believe that the Internet economy will be dominated by simple, focused businesses, rather than by sprawling conglomerates. When "q" settles down to a more reasonable level, I expect that the successful companies will be the ones that still can give elevator speeches.