A number of companies have built their business models around this plan and -- almost without exception -- they have gone out of business or suspended operations. The latest to do so are @bigger.net, a California-based ISP, and a Cincinnati-based service called Tritium. "Having a company announce they're going to enter this market is like hearing someone say, with complete confidence, 'I'm going to step out of that fifth floor window and flap my arms, and I'm going to fly, because I've got really strong arms,' " said Charles Ardai, president of free e-mail provider Juno Online Services LP. "You know he's going to hit the sidewalk, but all you can do is watch." Juno has been offering free e-mail service since 1996, using a proprietary e-mail reader and supporting itself mainly by ad sales. But the company charges users for Internet access, and Ardai says he wouldn't dream of it any other way. "You don't have to remain connected to the Net to read e-mail," Ardai said. "It's a much less expensive proposition than full Internet access, two orders of magnitude less expensive. The average Net user spends 25 hours online a month. While our users might spend 25 hours reading their e-mail, they don't spend nearly that much time connected to our network." Companies such as Tritium, @bigger.net, or the newly-launched NetZero give users unlimited, lifetime access to the Internet in exchange for allowing an advertising window to remain on their desktops while they are online. Some of the services, including @bigger.net, charge a one-time access fee to sign up. Experts say that the two major barriers to a viable, free Web-surfing business are infrastructure costs and the difficulties of building a substantial audience base. Free Web services are ubiquitous online, and although few are making money, they're not expected to disappear any time soon. That's because the costs of running a Web business are peanuts compared to the investments an ISP must make in modems, dial-up telephone numbers and customer service representatives, not to mention buying the bandwidth customers will use. Industry analysts say the Internet advertising market is just not big enough to support those kinds of costs. "That business model is, in general, probably not viable," said Shin-pei Tsay, research associate with Giga Information Group. "Because of the way page views are going, the top four sites are getting 75 percent of the advertising money, and the rest is divided up among millions of other sites." Advertisers are expected to spend $1.87 on Web advertising this year, and that figure will grow to $7.7 billion by 2002, according to Jupiter Communications. The theory is that advertisers will pay more to be able to carefully target their ads, and an ISP can offer more comprehensive information about a consumer than any Web site can. But the value of such targeting may depend on a company's ability to sign up a large subscriber base. "Advertisers will pay more to target their ads, assuming they have enough critical mass to slice and dice along geographic or demographic lines," said analyst Chris Charron of Forrester Research Inc. "But if you only have 50,000 subscribers, it's not worth it. ... Advertisers are only going to pay if they get results." Others counter that the number of viewers might not matter so much if advertisers feel they're reaching the right viewers. "When you talk about targeting, you have to look to cable television, and the rise of cable as the definitive success story," said Drew Ianni, analyst for Jupiter Communications. "They don't have a huge reach, but they do offer very targeted viewers. The same can be said for the Internet." Nevertheless, subscribers are what advertisers are willing to pay for, and building a subscriber base is no easy task, if the experiences of @bigger.net or Tritium are any indication. Juno says it has 6 million members; in contrast, @bigger.net had only about 40,000 after nearly two years in operation. "We happened to start early ... and obviously, no matter how fast you grow, you can't get from zero to 6 million users instantly," said Juno's Ardai. "Web sites ... don't have the same kind of scaling concerns, but with an access business, you have to become good at providing access -- it's not easy." Still, some experts say that it may still be possible for a startup to make the free-access model work, if it has the expertise and the funding. "It may be that it just takes 24 to 36 months and takes $10 to $15 million in marketing to get the business working, where the infrastructure is solid, and you're acquiring more customers, and it looks like a good business," said Jupiter's Ianni. Other attempts at free Net access have included J3, a long-distance telecommunications provider in Dover, Delaware; CyberFreeway, based in Silicon Valley; and X-Stream Network, based in the UK.