Venture capitalists have an old saying: "Even a turkey can fly in a hurricane."

In the five years that began with the initial public offering of Internet browser concern Netscape Communications in 1995 and ended with the dot-com stock crash in the spring of 2000, we witnessed an economic hurricane. The Web transformed—or at least threatened to transform—so many industries that anything that wasn’t nailed down looked like it might soar to the sky. Now that the hurricane’s winds have calmed, we can start to tell the turkeys from the eagles.

So, we have crawled through the Web’s wreckage in search of turkeys—by which we mean e-commerce predictions that missed their mark by an embarrassing margin. We have identified four of the most misleading and ruinous predictions of the past several years.

The exercise wasn’t just fun but also instructive. We still may be in the eye of the hurricane, with new winds of change about to buffet us. In any case, the mighty economic forces already unleashed by the Internet will wreak havoc on industries and companies for decades. So, it is useful to pause and consider the various versions of the future that haven’t come to pass, lest we repeat the same kinds of mistakes.

We will count down to the winner (the loser?), going from the smallest of the four turkeys to the biggest of them all:

NO. 4: MICROSOFT, THE RED-INK-STAINED WRETCH
The June 1996 cover of Wired magazine featured a doctored photo of Bill Gates lounging on a raft in a swimming pool, looking like a Hollywood mogul. The headline warned: "Microsoft morphs into a media company."

In a detailed cover story in the Columbia Journalism Review at the end of 1997, the headline asked: "Will Gates Crush Newspapers?"

The reason for the worry: Microsoft was investing in online articles and information and was wooing the same advertisers that had long patronized newspapers and magazines. For instance, Microsoft established Sidewalk to create online entertainment guides for dozens of the world’s major cities. Microsoft rapidly hired reporters and editors in cities such as Seattle, San Francisco, New York, and Boston and said many more cities were to come. Microsoft also started the Slate Web zine, purchased WebTV for $425 million, and invested $1 billion in cable company Comcast. Microsoft created MSNBC with NBC, invested in the production company Dreamworks, and amalgamated a bunch of Web sites into the Microsoft Network.

"Microsoft increasingly competes against almost everyone in this room for readers’ time and advertising dollars," Arthur Sulzberger Jr., publisher of the New York Times, told the Newspaper Association of America convention in April 1997. "If you don’t consider it a competitor today, in this increasingly digital world in which we live, all I can say is, ‘Wait!’" It didn’t help that Nathan Myhrvold, Microsoft’s chief technostrategist at the time, referred to the so-called Old Media as "road kill on the information highway" in a famous memo, adding that newspapers are in "probably the worst situation of any form of print media."

Even media titan Rupert Murdoch was aghast. "News and entertainment are changing the delivery system totally," Murdoch told a conference of chief executives in June 1997. "We have to stay on our toes to make sure Bill Gates doesn’t erect a tollgate in every house."

We are still waiting for Microsoft to become that dominant media company. To be sure, its Web sites are popular. But Slate and WebTV have yet to pan out, while MSNBC has attracted a rather modest audience. The investment in Comcast, while profitable, has done nothing to make Microsoft a media power. The Sidewalk site posted losses of tens of millions of dollars and never gained traction among users or advertisers. In July 1999, the unit was sold to competitor Ticketmaster Online-CitySearch, which essentially shut down the Microsoft ventures by the end of the year. At the corporate level, Microsoft gutted the famed "Red West" campus that housed its Interactive Media Division, which at one time employed 2,000 people. The head of that division, Pete Higgins, went on a sabbatical at the end of 1998. He then joined a start-up company.

It turns out that, despite all Microsoft’s money, brand, and leverage in the technology world, it just didn’t know all that much about producing articles and information for a mass audience. Newspapers, meanwhile, have gone on to have their best years in history in terms of advertising revenue.

NO. 3: THERE’S NOBODY @HOME
In the spring of 1995, legendary venture capitalist John Doerr teamed up with publishing scion William Randolph Hearst III to form a Silicon Valley start-up company called @Home. They imagined a world in which millions of American homes would use @Home’s cable modems to receive high-speed, broadband services via the Internet.

They knew that an earlier big bet on broadband had been a turkey. In the early 1990s, Time Warner and others used their cable-TV systems to test whether consumers would buy goods and order services such as pay-TV through "interactive television" systems. The answer was a resounding "no." The systems either didn’t work very well or cost too much. Still, Doerr and Hearst decided to try again. This time, instead of spending millions of dollars on tests, they and their partners resolved to spend hundreds of millions of dollars to provide nationwide, broadband access to the Internet.

Although it wasn’t completely clear just what services customers wanted at high speed, @Home assumed it could sign people up just based on the prospect that they would get Web pages delivered to them as much as 100 times faster than through regular modems. The company also figured it could push the service to consumers based on deals with more than 20 cable companies that promised to co-market the service to their combined base of 60 million subscribers. The deals were exclusive, but only for several years, so @Home needed to build its customer base quickly.

A confident Doerr predicted in 1995 that the company would sign up a million customers in the first year. In 1996, Doerr told Fortune that @Home would become one of the "five really great world-class companies built around the Internet." When he selected Tom Jermoluk, the president of Silicon Graphics, to become chief executive of @Home, Jermoluk said: "This is going to be like a rocket ship."

Certainly, @Home’s stock took off like a missile following the IPO, which was celebrated with a party at the fabled Hearst Castle. The doubling, then redoubling, of @Home’s shares put Jermoluk into the serious bucks. In 1998, when AT&T said it would buy TCI, one of @Home’s main backers, Jermoluk’s 3.2% stake in @Home turned to pure platinum. He was catapulted onto the Forbes list of Technology’s Wealthiest 100, higher up than John Chambers, the celebrated CEO of Cisco.

But where were the one million subscribers? In 1997, two years after the original prediction, @Home had signed up 12,000 homes. By June 1998, the number had reached just 150,000. It was taking time for the cable companies to upgrade their systems, which originally had been built to broadcast to homes, not to receive signals back from people. It also was taking time to win over customers. Although the service was priced at just $40 a month, installation could be complex, outages were common, and customer service was living up to the cable industry’s reputation—in other words, service was awful.

Undaunted, Jermoluk began saying that one million would be just the starting point. He told the Wall Street Journal in 1999 that he expected to have 10 million to 15 million subscribers by the time @Home’s exclusive cable deals start expiring in 2003.

By December 1999, @Home finally hit one million subscribers—more than three years late. And at what cost? The company posted a loss of $723 million in 1999, including charges related to the acquisition of portal Excite for $6.3 billion in stock, and the $788 million cash-and-stock acquisition of Blue Mountain Arts, an online greeting-card company. Investors, by now skeptical, sent the stock plunging to about $11 at press time from a high of $99 in the spring of 1999. At that recent price, the company has a market value of less than $4 billion—or much less then Excite, alone, at the time of the acquisition. By the spring of 2000, Jermoluk was out as CEO. By September, his successor, George Bell, had also resigned.

NO. 2: NO FREE LUNCH, OR PC
The "free" personal computer is another idea that has been kicking around for years, but the hype reached new heights in February 1999 when Bill Gross, founder of the Idealab incubator, announced the formation of FreePC. "Merchants will pay to reach you, so they essentially will subsidize the cost of the PC," Gross, a manic entrepreneur, said at the time. FreePC immediately became the highest-profile embodiment of the growing "everything will be free" movement—at least until the online music service Napster came along.

Orders flooded in, but FreePC had only 10,000 Compaq Presarios on hand, so it had to turn away hundreds of thousands of people. Actually, the lack of machines may have been a good thing, because the business model was so suspect that the more machines the company gave away, the more money it lost.

Here is the business model, as laid out by Gross at the Internet Summit conference in Laguna Niguel, Calif., in the summer of 1999: Customers would agree to use the PC for 60 hours a month for three years. (Gross didn’t say how this rule would be enforced.) FreePC would collect data on users’ demographics and on how they used the machines. Advertisers would pay an average of $1 a day to reach each user. With a PC costing the company $500 to $600 at wholesale prices, the point at which the machine was paid for would be reached sometime in the first two years. FreePC would then have the final year-plus of the contract to reap profits. "The monetizing effort requires a long relationship with the customer," Gross said.

There was just one problem: FreePC only lasted nine months. It needed too much capital to buy PCs and couldn’t get enough advertisers to pay for its data. FreePC was sold for an undisclosed sum in November 1999 to eMachines, a seller of low-cost—but not free—PCs that has by now racked up more than $86 million in losses as a public company.

NO. 1: THE BUY.COM SPREE
In 1997, Scott Blum founded Buy.com on the idea that "virtual retailing" would be an unstoppable business model. He argued that he would even surpass Amazon.com, which he felt had veered from a pure virtual model by building warehouses and hiring so many customer-service reps.

Blum’s idea was to sell everything at or below cost. He touted a plan to make a profit sometime down the road by selling advertising and ancillary items such as extended warranties. He planned to keep costs minimal by having the manufacturers of items sold on his site ship the items directly to customers. Every operation at Buy.com, including its minimal customer-service center, was to be outsourced. Perhaps the only thing Blum refused to skimp on was Super Bowl commercials.

In 1998, Buy.com broke Compaq’s longstanding record of $111 million in sales in its first year in business—and it did so with fewer than 100 employees. At the beginning of 1999, Blum told Fortune, "We plan to exceed Amazon’s revenue in December" of that year. He added that he intended to reach $10 billion in revenue by 2003. To foreshadow where the company was going, Blum bought 2,000 domain names, most beginning with the word "buy," such as Buyinsurance.com, Buycars.com, and Buycheeseburgers.com.

Blum found plenty of believers. Blum, who owned 65% of the company, sold a 20% stake to Internet incubator Softbank for $60 million. (He used some of the proceeds to build an ocean-front mansion in an exclusive Southern California enclave—a house so over-the-top that it drew bitter objections from the neighbors.) Fortune placed Blum on its list of the 40 Wealthiest Executives Under 40 and gushed about his prospects.

When Buy.com went public in March 2000, it was a huge hit—for about 24 hours. The stock price doubled on the first day. (Blum had, by this point, left the company because he had a skeleton in his closet: Financial crises at Pinnacle Micro, a maker of optical disk drives that Blum co-founded, led to an investigation by the Securities and Exchange Commission into whether that company was booking revenue before it should have. Blum settled with the SEC without admitting or denying guilt, then resigned from Pinnacle. Investors in the Buy.com IPO might have worried if he had kept a management role.) The company posted losses of $82.5 million in its first two quarters as a public company. Selling every item at a loss will do that to you.

In the six months following that first day’s close, Buy.com lost more than 90% of its value. Rival virtual retailer Value America went belly up, seeking bankruptcy-court protection, and investors were figuring it was just a matter of time before Buy.com did the same.

HELD ALOFT BY HOT AIR
The common thread running through these predictions has to do with how radical change really happens in the real world. All four of these predictions failed to materialize because these companies were throwing massive amounts of today’s cash at a far-off goal.

Yes, as Microsoft assumed, digital and analog media are probably converging. Yet because of the unpredictable timing of the convergence, Microsoft moved much too early. In starting Sidewalk, the firm clearly went way beyond its core competencies and paid a price for doing so—not just in dollars, but in how it antagonized the media.

Broadband eventually will arrive in the home. But a start-up company working against entrenched forces has little chance to be a pioneer in a market that might not be ready for it. Ironically, now that the world’s biggest cable company, AT&T, appears to be gaining almost complete control of @Home, broadband might happen the old-fashioned way. With a greater ability to withstand the initial losses, the old Ma Bell may pretty much impose the technology on the public by making it a standard part of its infrastructure.

FreePC probably was correct in assuming the cost of some desktop PCs eventually will drop close to zero. But providing free PCs probably will never be a standalone business supported by advertising. Most likely, the nearly free PC will be bundled as part of a complete telecom/computing/software package—perhaps from a big phone, cable, or software concern.

Virtual retailers such as Buy.com learned that long-term success in all forms of retailing always will be about more than just the lowest possible prices. Even though the Internet is a wonderful collection of technologies, it can’t enable a company to turn a profit on a $100 item that it sells for $85.

Perhaps that is the biggest lesson of the last five years: The Web isn’t magic after all.

Schwartz is the author of Digital Darwinism (www.digitaldarwinism.com). He can be reached at evan@webonomics.com.