As long as we're now so many years into our fascination with the Web (which hit most people's radar screens all the way back in, oh, 1995), the editors at Context decided to pull together a list of the five dumbest ideas of the Internet Age. To be at least slightly politic, let's call them the Five Worst Predictions. We don't mean to be cruel about this—oh, maybe a little. We mainly wanted to make the point that, as the great Yogi Berra once said, an accurate prediction is very difficult to make, especially when it's about the future.

Predictions about technology have always been especially ticklish. Why, Bill Gates himself once assured us that 640 kilobytes of random-access memory was all that PC users would ever require—many PCs now have 200 times that much memory. The prediction business has become tougher since the Web burst on the scene. The half-life of prophecies has become so short that some of the boldest predictions go awry almost as soon as they are uttered.

It'd be nice to be able to just sit around and laugh at the gurus, and sometimes we do. But the unpredictability of the Web is a serious problem for anyone trying to do business on-line—which, these days, means just about everybody. How can you build a long-term business plan when the generally accepted "truths" about the Internet are overthrown so rapidly in favor of new ones? You can't, at least not with any certainty.

But how can you avoid trying? It seems to us that you have to try. But if our Worst Predictions are any guide, you should rely less on the pundits and more on your own experience. You have to test all the ideas, all the technologies, and all the business models for yourself so you're ready if they hit—while spending carefully enough that you don't look silly if they bust.

Let's get started. We'll begin with:

No. 5: Digital cash, whose proponents helped to convince the world that consumers wouldn't use credit cards to buy over the Internet because of fears about security.

With fear of hackers widespread in 1995 and 1996, companies rushed to develop complex, highly secure payment systems. Companies with names such as Cybercash, First Virtual, and DigiCash sprang up with solutions. Some required all shoppers to download special "wallet" software that would store specially encrypted digital money. Others required merchants to install special storefront applications that would know how to translate these digital debits into dollars and cents.

For a while, these digital cash systems were treated as a given, as a necessity if electronic commerce were to ever reach critical mass. In 1996, MasterCard Senior Vice President Edward Hogan said that electronic cash "will be a multibillion-dollar market" by the turn of the century. A January 1997 report from consulting firm Killen & Associates forecast eight billion e-cash transactions on the Internet in the year 2000.

It was not to be. Hordes of consumers began ordering books, CDs, airline tickets, computers, and the like over the Web, paying simply by typing in their credit-card numbers. Encryption programs bundled right into Web browser programs scrambled the card numbers as they traveled in packets over the Internet. In the process, the stocks of First Virtual and Cybercash lost more than 80% of their value. The marketers of the digital cash concept ended up burning through tens of millions of dollars before they phased out their products and began searching for Plan B, an entirely different business to enter.

No. 4: The doomsaying about America Online, which, again, was wildly far off. The predictions carried consequences because many marketers, at their peril, ignored it as a way of reaching customers and didn't foresee the potential for reaching customers through high-traffic "portal" sites like AOL.

In 1995, industry analysts sounded the death knell for America Online because, the thinking went, the explosion in use of the World Wide Web meant that people would no longer need to pay AOL for its information services or even its chat capabilities. "All of the proprietary on-line services are really on a march to extinction," Forrester Research analyst Mary Modahl said.

Sure enough, in the fall of 1996, after AOL surpassed five million members, the company seemed to hit a wall. While the number of users of the Web grew exponentially, AOL couldn't seem to keep up. Busy signals greeted users who dialed into the network. Anyone who managed to get on-line encountered painfully slow response times. Ms. Modahl foresaw consumers deserting.

But within a year, AOL's membership more than doubled, to 12 million customers, and it has continued to rise. Nowadays, AOL is often cited by analysts as the media empire to beat in the 21st century and the only company with the market clout and staying power to challenge Microsoft's dominance in the digital economy. "I was completely wrong," Ms. Modahl has said of her 1995 prediction.

No. 3: The "network PC," which made it seem that the Internet's success depended on the coming of a technology breakthrough that, in fact, wasn't necessary.

Back in 1995, Larry Ellison, the billionaire chairman of Oracle, declared that the Wintel (Windows/Intel) personal computer was "ridiculous"—too complicated to use and too expensive for the average user. "The price of PCs has been constant for 15 years because Microsoft insists on sticking more and more junk in them," Mr. Ellison said. Instead of $2,500 machines with huge hard drives and bloated software applications, his new, disk-less "network computers," or Net PCs, would sell for less than $500 and run simple "applets"—small applications written in the Java language from Sun Microsystems.

The new Network PCs didn't hit the market until 1997, and by then there was little demand for them. Even as PCs had become more powerful, prices had fallen to around $1,200, closing the price gap. Prices soon headed well below $1,000. Besides, people were so accustomed to the familiar design that switching to a Net PC was out of the question. Oracle soon disbanded its Net PC division and took a big write-down.

No. 2: The various predictions on the number of users of the Internet, which have consistently underestimated the speed with which the new medium would gather steam. Now, we acknowledge that predictions about violent change are treacherous. But estimates are supposed to be high. Everybody knows that technology forecasters get overly optimistic about the speed of change. Because, with the Internet, the estimates were so low, they allowed many companies to think they could afford to wait to form on-line strategies.

In 1995, for instance, Forrester Research projected that the Internet would have a worldwide user population of 34.9 million people by 1998. Notice the precision. Not 35 million, but 34.9 million. The real number ended up being in excess of 100 million users.

Researchers also grossly underestimated the size of the electronic commerce marketplace. In 1995, Jupiter Communications projected $3.1 billion in annual business-to-consumer revenue for e-commerce by 1998. Forrester predicted $2.3 billion. The real number turned out to be more than $13 billion.

No. 1: Drum roll, please. We made "push" technology No. 1 on our list of Worst Predictions partly because its proponents confused so many people into thinking that the Web would be a broadcast medium, much like television—but mainly because the hype was so deliciously flagrant.

The idea got rolling back in late 1996, when the start-up PointCast reported that more than one million users had downloaded its free "push" software. The media saw a crystal-clear vision of the future. Instead of having to troll the Web and retrieve, or pull, news and information, people could use software from PointCast and others that would push to their screens exactly what they wanted—accompanied, of course, by paid advertisements.

The media went into hysterics. Business Week heralded an age of "Webcasting" in a February 1997 cover story. "I see push creating almost a second Internet," said Halsey Minor, chief executive officer of C/Net, a leading Web news service. "We're huge believers," said Bruce Judson, then Time Inc.'s chief of new media. "Kiss your browser goodbye," the editors of Wired magazine wrote in a cover story, titled "PUSH!"

Based on nothing more than press releases and early product demos, the media anointed a former midlevel Sun Microsystems executive named Kim Polese of Marimba as the "queen of push" and the Web's new "It Girl." For a start-up with few real customers, Marimba received perhaps an unprecedented flurry of press coverage. She became part of a small group that advises Vice President Al Gore on technology policy. The delirium culminated when Time magazine named Ms. Polese one of the 25 most influential people in America for 1997.

We all know what happened next. PointCast users found the visual presentation annoying, and corporations realized that all the pushed content clogged their networks. Users disabled or deleted the software. Less than a year after the push hype reached its zenith, companies such as Marimba disassociated themselves from the very concept of push and began looking for new applications for their software.

Several predictions that didn't make our list deserve honorable mentions:

There was the talk of Apple's demise, which, while not all that significant a factor for companies setting Internet strategies, was just so far off....In 1997, a Business Week cover featured the Apple logo on a black background, speaking about Apple in a wistful past tense. But when co-founder Steve Jobs returned as chief executive, he orchestrated a remarkable return to profitability.

There was also the speculation that Yahoo and Amazon.com, among others, would flame out. Yahoo was supposed to face many of the same pressures as AOL. Amazon was supposed to have trouble sustaining its phenomenal growth. In fact, we at Context share some of the skepticism about Amazon. But, as of this writing, Amazon carried a $17 billion stock market value, Yahoo a $31 billion value. Not bad for a couple of failures.

There was all the talk about how the Java programming language would remove the need for operating systems such as Windows and could put Microsoft out of business. George Gilder, for instance, wrote in a 1997 column in the Wall Street Journal that "Java has the power to break Microsoft's lockin of applications profits and lockout of rival operating systems." Java's claim to fame was that programs written in that language would be able to run unaltered on every computer. But the last we checked, any given Java applet would run on about half the computers in the world. And Microsoft had added more than $200 billion in market capitalization since the Gilder column ran, making it the most valuable company in the world.
   

There are some lessons to be learned by looking at the common threads running through many of the predictions that turned out to be off the money. The ones that cause their makers to look back and wince seemed mainly to underestimate the twin forces of inertia and simplicity. The sheer fact that lots of people use a certain technology creates momentum that is extremely difficult to stop. In addition, the history of technology shows that, when faced with two choices, customers will almost always opt for the simpler solution.

AOL, for instance, had both forces going for it: the inertia of the largest on-line user population, plus the fact that AOL was the simplest way for novices to get on the Internet. Push, meanwhile, was fighting both forces. It faced complex technology and business issues; plus, it required breaking the surfing habits of tens of millions of Web users. Likewise, while Net PCs seemed to have simplicity on their side, they actually had some complex technology issues to overcome. The infrastructure necessary to serve software to millions of users and to store gigabytes of data on far-flung networks was quite troublesome. Net PCs were also fighting the inertia of years of PC purchases.

The real lesson, though, is that predictions are hard—something to keep in mind when experts say that they have now seen the future and it is...portals.

Perhaps we should pay less attention to the actual predictions and just focus on their entertainment value. Push technology? "It will be replaced by Poke technology, allowing users to play Three Stooges with people at remote locations," says cartoonist and amateur futurist Scott Adams. He was also among the first to foresee Intel's collapse into bankruptcy. "I give them two months, tops," Mr. Adams said not long ago. Now that's a good prediction.