And some of the players are likely to strike out before long. Lycos, Excite, Infoseek, Yahoo! and others have been adding features at a furious rate in an attempt to turn themselves into Web portals, a term for multifeatured gateways to the Internet. The idea behind the recent deals between AT&T and the three search services is the same one driving those services to add features such as free e-mail and chat rooms: to attract users, and the advertising dollars that come with them. In the longer term, however, analysts say the portal market is headed inevitably for a shakeout. "The portals' share of ad revenue right now in relation to their traffic is quite high," explains analyst Chris Charron with Forrester Research. "They get about 15 percent of all traffic, but 59 percent of ad revenue online. That's going to change, they're going to get a smaller share. And even with total advertising revenues growing, that won't be enough to sustain nine broad-based portals. That's going to result in a lot of consolidation in the portal space." One thing appears certain: The battle is just beginning. "No one knows who the primary portals are going to be in the year 2000, but each of these brands are battling in 1998 to secure brand recognition and strategic positioning to ensure they're left standing after the shakeout," said analyst Ron Rappaport of Zona Research. "Marketing acumen and brand cachet will be the deciding factor for the victor in the portal wars." Building that cachet may depend on something that's difficult to measure: how well a site understands what its users want. "Yahoo has 30 million users, Excite has 20 million, whatever. But that's just part of the game," argues Gregory Wester, a research director with The Yankee Group. "What people aren't looking at is, who's got the knowledge of the customer? And in that space Yahoo! and Excite are much further ahead (of other search sites)." Personalisation could make the difference, because once users get hooked on using a service like My Excite or My Yahoo!, they tend to stick with it, Wester says. His research shows that out of all search engine users, only 13 percent stick to the same service in any given month. But out of those who use a personalised site, nearly all -- 95 percent -- use only that search service. Wester believes that advertisers will ultimately pay big bucks not just for volume of users, but for customers who come back day after day. Observers believe the portal area will soon fall into a hierarchy of a few big, well-trafficked portal sites -- Yahoo!, Excite, America Online and Netscape typically top the list -- along with smaller portals centred around a particular type of content. A third type of portal might revolve around life-stages. For example, pregnant women might flock to a site that would broker specialized information, advice and even speciality goods. The field could look something like the television market of today, with a handful of broadcast networks and a proliferation of niche cable channels, analysts say. In the long run, however, there's no guarantee that portal sites will continue to draw the huge amounts of traffic they do today. As the Internet's infrastructure develops, it could do away with the need for portals altogether. "Right now, the reason portals have value is because they collect a wide variety of content and services into one point," observes Michael Sullivan-Trainor, a program director with International Data Corp. "Now if you imagine a future where you have an automated agent or some other way to get where you want to go, then the need for portals disappears."