(FORTUNE Magazine) – Internet CEOs crave many things: world domination, instant service in bistros, fawning media attention. But what they crave above all else is eyeballs. That's less ghoulish than it sounds. In Webspeak, you see, eyeballs mean customers. Since the typical dot-com lacks the one metric that Wall Street has traditionally used to evaluate companies (you remember--earnings) analysts and investors have contrived other ways to size up Net stocks. One now stands out: market capitalization per pair of eyeballs. It's a useful first step in explaining why a company garners a certain kind of valuation. For instance, a pair of eyeballs at Web portal Lycos, with a $7.4 billion market cap, has a value of just $244; at Schwab, which has a $30 billion market cap, a pair is worth $4,562 (ironically, this also happens to be around the price a pair of real human corneas reportedly commands on the black market).

If the Internet market were rational, the market cap per eyeball would represent the total profit that you could reasonably expect a company to get from its average customer, adjusted for risk and the length of time before those profits are realized. Internet analysts are the first to admit that today's is not a rational market. So correlating the lifetime value of eyeballs to a fast-growing dot-com's stock price is not perfect science. As Lise Buyer of Credit Suisse First Boston puts it, whenever you try to relate the price of a Net stock to the underlying economic fundamentals, "the math does not work." Eyeballs aren't earnings.

Nevertheless, trying to assess how much a Net company's customers are worth does help shed light on why the market values some (America Online, eBay, and Yahoo, say) more dearly than others (MindSpring, barnesandnoble.com, or SportsLine.com). It also raises troubling questions about low-margin e-tailers, such as eToys, that spend more money to acquire new customers than those customers spend on their Websites. And focusing on the value of eyeballs makes it harder to avoid the real business questions facing dot-coms; like, Is there a chance they will ever turn a profit?

In what is beginning to emerge as a tougher market for Net stocks, investors are demanding results--not profits, necessarily, but at least a glimmer of profitability. No longer is the "relative valuation" rationale enough. It's simplistic to say that since SportsLine and Quokka are both sports-content sites and since Quokka's market-cap-to-revenue ratio is twice as high, SportsLine is undervalued. Instead, investors are focusing on eyeballs. After all, customers are the driver of every Website's revenues: buying stuff at e-commerce sites, raising the ad rates of popular content sites or portals, determining the prices that heavily trafficked sites charge affiliates that want Web real estate. So investors want to know how many customers visit a Website, how much revenue each generates, and even the gross profitability of each. With the help of Credit Suisse First Boston, we've done that analysis for some popular sites on the Web, and we've dug deeper to calculate how much money companies spend on marketing as they seek to acquire and retain customers. Says James Marks, another CS First Boston e-commerce analyst: "The sad truth is that many, if not most, CEOs we talk to cannot pin down the value of a customer. Therefore they do not know how much they should be paying for a pair of eyeballs."

The table on this page ranks a representative selection of Net stocks by the value that the market places on their companies' customers. It's not as if investors should buy or sell based solely on the numbers in the table, but understanding eyeballs can be the first step to making smarter Web investments. We've also broken down some companies' numbers in the sections that follow, in an effort to show how analysts use eyeballs as the basis for a more rigorous understanding of the value of Net stocks.

CONTENT VS. COMMERCE

Consider the relative merits of Yahoo and Amazon. Yahoo's market cap is $86 billion, vs. Amazon's $26 billion. Yahoo, which has 42 million customers but doesn't sell much of anything, derives only $18.99 a year in revenues per customer. Amazon, on the other hand, averages $160.01 a year from each of its 17 million customers. While those revenues include the money Amazon charges affiliate e-tailers, they mostly reflect actual cash paid for products like books and CDs. Yet the market believes that each Yahoo customer is worth $2,038, vs. $1,400 for each Amazon customer. How can someone who just clicks through a search engine be worth 46% more than someone who buys Graham Greene novels and Smashing Pumpkins CDs?

Brad Garlinghouse, a venture capitalist with CMGI's @Ventures, answers, "Ads are worth more than commerce because of price pressure [in commerce]. You can't make money if you ship books at a loss." If you look closely at the numbers, you see he has a point. Amazon's gross profit per customer is only $20.79--in other words, the cost of the CDs, books, and consumer electronics Amazon sells to reach that average of $160.01 per customer is typically $139.22. But then you subtract another $42.47 per customer that Amazon spends on marketing and the company is losing $21.68 a head. That's merely a partial measure of loss because we haven't deducted an additional $19.83 for warehousing, shipping, customer service, and other operating expenses.

Yahoo, on the other hand, doesn't ship physical products, so its 86% gross margins give it a gross profit of $16.42 per customer. It spent only $6.11 per customer on marketing, so its partial profit is $10.31 a head; subtract all other operating expenses, and Yahoo still earns $7.53 per eyeball pair.

Looking through the eyeballs lens, there's only one conclusion to draw: Yahoo now has the better business model. Its new customers add virtually no incremental costs, and it has accumulated so many eyeballs that advertisers pay Yahoo more than they pay other portals. Whereas every customer that Amazon adds simply increases the company's loss--they're not spending enough to make up for the huge marketing costs Amazon incurs in luring them in. (Disclosure: This reporter owns shares of Amazon.)

The promise of Amazon has always been that as it grows in scale, its purchasing power will increase, its operations will become more efficient, and its gross margin per customer should improve. That's been the idea for years, yet each year the losses just keep on going up. But Amazon has a new and interesting way to leverage its eyeballs. As Buyer points out, "The most interesting business models belong to companies that are able to sell their customer base." In a spate of deals in late January, Amazon sold prominent positions--called tabs--on its site, one to drugstore.com for $105 million and another to home-furnishings site living.com for $145 million. It also sold a link to online car-dealer Greenlight.com for $83 million. Since Amazon won't be physically fulfilling orders for shampoo, couches, or cars, it will realize an estimated 80% gross margin from these deals (compared with its usual margin of 20%). Merrill Lynch analyst Henry Blodget thinks such agreements could push Amazon into the black by next year. And, he adds, "they are definitely going to do more."

THE VALUE OF FREE

Evaluating eyeballs helps explain why giving stuff away--content, Web access, even hardware--can propel a company to a high valuation. At www.quokka.com, for example, Quokka Sports offers sailing, auto-racing, and adventure-sports enthusiasts in-depth coverage of their favorite events. The America's Cup trials don't bring in a Super Bowl-like throng. But advertisers pay Quokka about $70 per thousand ads shown, because the site gives them a way to zero in on the eyeballs of wealthy sports fanatics. So Quokka's potential profit per pair of eyeballs is pretty high, which helps explain why it has a $500 million market cap, despite having just 680,000 visitors per month.

NetZero, meanwhile, has earned a high valuation despite that its offer of free Internet access plays to the Kmart crowd. Those who use its software have a running ad banner on their screen as they surf the Web, and since NetZero monitors their every move on the Web, it has the potential to deliver targeted ads during the 35 minutes per day that each of them, on average, is online. For this reason, the market pegs the lifetime value of a NetZero customer at $1,140--even though he or she gets everything for free. That's pretty close to the value the market places on the average customer of MindSpring, another Internet-access provider--whose customers pay $20 a month for service. Combine NetZero's targeting potential with its low customer-service and marketing costs, and you understand how the two companies might be judged equal.

If you doubt that the free Internet business model is catching on, consider Netpulse, a private company that gives health clubs free Internet appliances to put on their gym machines. As people work out, they can surf the Web or watch TV via a touch screen that runs a steady stream of ads across the bottom, targeted to the yuppies who populate these gyms. Netpulse originally tried to sell its appliances for $2,500 a pop, but gyms weren't too pumped up about them. Now that CEO Tom Proulx--a co-founder of Intuit--is giving away the machines, the adoption rate is skyrocketing. Proulx thinks that his young, captive, desirable audience is worth more to advertisers than it costs him to provide the Net appliances to the health clubs. "People in health clubs are totally bored, actively seeking distraction. Their minds are up for grabs," he says. As of Netpulse's last round of private financing in September, each of the half million people who use its machines was valued at roughly $200. They may be worth several times that amount if Netpulse goes public.

THE NEW MIDDLEMEN

E-commerce sites salivate over customers like Andrew Goldstone. In the past six months he's dropped over $1,500 at eBay on prints by artists Josef Albers, Robert Rauschenberg, and James Rosenquist, as well as on rare books and a bootleg Oasis CD. eBay's cut of these transactions was probably around $100, with a 71% gross profit margin. That's not shabby.

But eBay's market cap assigns its average customer a lifetime value of $1,952. To get a sense of how outlandish that may be, consider this: Goldstone (and eBay's ten million other customers) would have to increase spending to $3,000 a year and keep that up for ten years straight to justify eBay's current stock price. (And this does not even take into account the fact that a dollar tomorrow is worth less than a dollar today.) Not likely--especially for Goldstone, whose cramped Manhattan apartment has only so much wall space.

eBay may be able to get out of this fix by adding customers. The market now says that the lifetime value of each eBay customer is a high $1,952--but last June, when the company had just 5.6 million customers and the same market cap, that lifetime value was $3,486. If the company continues to add them more quickly than its market cap grows, the pressure on its eyeballs will decrease even more.

Consumer exchanges like eBay must amass tens of millions of customers to succeed in a big way. B2B exchanges such as Chemdex or VerticalNet, on the other hand, need only link up hundreds of thousands of corporate buyers and sellers to score big.

Putting it bluntly, "B2B customers spend more money," says Chemdex's head marketer Martha Greer. She backs up this claim with evidence that the average order for laboratory chemicals and supplies on Chemdex's Website is around $500 and that a typical registered user--a research scientist at a pharmaceuticals or biotech company--orders over $20,000 of supplies a year. Chemdex's take is less than 10% of each transaction. Even so, it is a lot easier to see how Chemdex could generate $2,000 per researcher than to see how eBay could milk $3,000 from the likes of Goldstone.

Furthermore, B2B exchanges are possibly the stickiest Websites. Buying through Chemdex cuts a customer's procurement costs by as much as 90%. If that isn't incentive enough, Chemdex customizes each buyer's experience, so that a chemist logging on from SmithKline Beecham sees something very different from what a Bristol-Myers Squibb scientist sees. Such customization takes the time and participation of the customer, so the cost of switching to another exchange is considerable. Furthermore, says Greer, since Chemdex's relationship is with the enterprise, not individual researchers, "there is no discretion on the part of the end-user to go somewhere else. Their organization has chosen to buy this way."

While B2B customers are worth more than individual consumers, the value the market is conferring upon them ($172,100 for a Chemdex customer and a gargantuan $3,064,873 for a VerticalNet one) is highly distorted. Nobody's average customer is going to deliver over $3 million dollars in lifetime profits. Investors are betting that these companies are going to grow a lot by adding many more customers (the VerticalNet figures, for example, are based on just 2,903 storefronts that customers have established on its Website). If that happens, these companies will see their value-per-customer ratio decline to something closer to reality. But probably not all the way down. The gap between a rational appraisal of eyeballs and the irrational pricing of this market is still huge. Assessing the value of eyeballs gives us a less fuzzy image of Net valuations, but it's still not a crystal-clear view.