All Internet Stocks Aren’t Equal(ly Overpriced)

Someday people (not many people, but a few–well, okay, at least my grandchildren) will look back at these Moneybox columns and wonder if Internet stocks really were the most important thing about the U.S. stock market in 1998. Of course, they’ll be looking at these columns on the Net (or some form of it), so the fact that the answer might very well be “Yes” may not seem that odd to them. Even so, it seems very odd to me.

On the other hand, except for the technology sector, which continues to deliver strong earnings, and the occasional anomaly like GE, most of the stock market is stuck in the middle, meandering ahead a few paces before being dragged back by earnings warnings from the like of Caterpillar or Mattel. The Internet sector, on the other hand, continues its amphetamine-fueled rise, smashing through every conventional valuation model and providing the best evidence yet that as long as you can get someone to buy a stock at $350 a share, you should buy that stock at $320 a share.

Now, how you know that someone will be willing to buy that stock (call it at $350 a share remains somewhat mysterious, and those who advocate jumping on the Net bandwagon without explaining how you know when to jump off should answer that question before they write anything else. (Useless rhetoric along the lines of “industry leaders” and “geometric revenue growth” won’t cut it, either.) But I’ve said more than enough here about Net insanity (that is, I’ve not made huge amounts of money). Instead, I want to suggest that in some cases, conventional valuation models really can lead you to underestimate the real value of a stock like America Online.

The obvious critique of America Online, which now trades at something like 170 times earnings, is that even though it’s profitable, it’s nowhere near profitable enough to justify its share price. (Call this the Barron’s take on the stock.) In theory, a company’s present value is the value of its future earnings flow, discounted appropriately. And even if AOL’s earnings growth is rapid over the next decade, its stock still looks overvalued.

The problem is that AOL is genuinely better off not concentrating on earning as much as it can from every customer right now, and instead focusing on increasing its customer base. If AOL just stopped spending money on customer acquisition, after all, its costs would drop significantly and its earnings would rise. But in the long run this strategy would doom it, if not to extinction then to stasis. AOL needs to keep growing in order to live. And because the future value of each customer is significantly greater than the cost of acquiring that customer, its marketing strategy is an efficient use of capital, which is the best mark of a good company.

At some point, of course, AOL needs to step up its efforts to milk more money out of each customer. But here again, the company enjoys increasing returns to scale. The more customers AOL has, the more it will be able to charge advertisers, and the more ways it will have to take money from customers (through e-commerce, etc.). AOL’s entire strategy, then, depends on the recognition that it can use capital better by spending it than by distributing it as dividends or keeping it on hand as cash.

Other Internet companies–most notably Amazon–make similar arguments about their up-front spending. But there are some crucial differences. AOL has a guaranteed revenue stream per customer, with its subscription fees. More importantly, e-mail addresses have become the kinds of things people don’t want to change, which means AOL has built-in customer loyalty. Amazon says it’s creating loyalty with its personalization features, but ultimately it’s unclear how it’ll be able to differentiate itself from its competitors.

Who knows whether this means you should buy AOL (though if it dips back down to $90 again I’m backing up the truck). But it does mean that all Internet stocks, and all P/E ratios, are not created equal.