beyond the portal
Many internet marketing experts believe that search engines and other large portals dominate the Web. A small industry has even sprung up, promising to help sites improved their ranking in search engines. Search engines and directories themselves are beginning to offer premium listings or expedited handling for a fee. Guides and tutorials often argue that portal strategies are key to attracting visitors:
"Search engines and directories [are] the backbone of Internet marketing ... the way to attract hundred of new, targeted prospects each and every day." -- Ken Evoy, Make Your Site Sell
This perceived importance of portals is positively reflected in the soaring stock market valuation of portal sites, and negatively reflected in concern for the Disneyification of the Web and for the growing dominance of what Stuart Moulthrop has termed the military-infotainment complex:
"'Ours is an annotative age?' Well, take a note, radboy: ours is also an age of oligarchy. . . . As three or four exceedingly large enterprises come to dominate certain global markets, their enterprise service providers will be able to dictate standards and practices."--Stuart Moulthrop, Hypertext '98 keynote
Portal sites do attract tremendous traffic, and many people use them daily. But portals are much less important to the shape of the Web than most people believe.
Look at the Numbers
A look at Web logs reveals an interesting picture: while search engines and portal sites are important sources of visitors, most visits do not begin at a portal. For example, consider a sampling of visitors to Eastgate's site -- a small to mid-size sales site that serves an established, bricks-and-mortar software publisher.
|other sites and bookmarks
Yahoo sends more readers to Eastgate than any other site, but only 5% of all visits come from Yahoo. A far greater fraction of visitors come directly from small sites. In the 1999 sample, for example, 326 sites sent fewer than one visitor per day, but more than one visitor per week; these visits accounted for about 28% of the total visits for the period.
Implications for the Web Economy
Many observers have noted that the price of Web advertising has been falling rapidly. Accurate figures are difficult to obtain, in part because prices are falling so quickly. Weak demand leads to widespread discounting, and banner ads often sell for a small fraction of the list price.
It is quite possible that the falling price of advertising space stems not from weak demand, but from an abundance of supply. The Web is growing rapidly. New sites appear each day, and an interesting site on a hot topic (or simply a site that tells a good story, like Colors or Bruno) can suddenly attract thousands of readers. The number of Web pages on which banners might appear, and the number of Web pages viewed, are ballooning daily. Abundance drives down prices.
When banner ads first appeared, publishers charged about 4 cents per exposure. Since then, the number of Web pages viewed each day has increased dramatically; if prices had held firm, the Web would be generating wealth at a rate that dwarfs its already-phenomenal performance. A site averaging a mere 2000 visitors a day, with 5 impressions per visitor, would generate well over $100,000/year.
This would indeed be an economic revolution. Movie review sites would earn more than the movies they cover. It would be more lucrative to run a fan site about a television star or basketball player than to be an actor or an athlete. (This has not yet occured.)
The number of viewers who click on banners has declined rapidly as well. When the Web grows rapidly, click-through rates naturally decline because readers have so many new choices. Declining click-through rates, though, depressed ad rates, leading some to conclude that Web advertising is ineffective and that only high-volume portal sites can hope to realize even meager revenues from advertising.
Portals dominate Web advertising today because their sales costs are low. Large sites with large inventories of ads to sell can support sales departments. Small sites can sell fewer banner slots and must spend less on sales. Salespeople, like eggs, aren't easily divided; it's hard to hire 10% of a sales person, and overhead costs don't scale.
In time, as global Web traffic grows, an increasing number of topical sites will be able to employ sales staffs. Sites too small to support even one salesperson will band together through agents, networks and alliances. Portals will remain both important and highly visible, but their advantage will diminish as smaller sites become more efficient.
Implications for the Shape of the Web
Many have feared that the future of the Web will recapitulate the history of broadcasting: that it will soon consolidate into a handful of dominant sites eclipsing all the others. However, it is likely that the centrifugal force of links will balance the pull of centralization. Over the long haul, while high-traffic sites will prosper, the overwhelming majority of Web page visits will be made to and from smaller sites. (Indeed, one of the fastest-growing applications of the Web today is the highly-specialized site intended to specifically for a small group of family or friends, a supplement to occasional letters and phone calls. Trellix, originally an intranet-oriented company, has reinvented itself and its software to serve this neglected market.)
As smaller and more specialized sites regain confidence, we may also anticipate an richer design and production. This prediction may seem counter-intuitive, because small sites have small budgets. But small sites can afford to take risks -- and often could benefit from the attention and audience that successful design risks might attract. A larger site is more constrained; it needs to maintain its audience, its identity, and it is limited by legacy workflow applications -- its own and its customers'. Small sites can afford mistakes that, for big sites, lead to erosion of brand integrity and unproductive finger-pointing. Worry about blame leads to design by committee.
Some small sites will gain audience and become large, with the advantages and disadvantages of scale. Large sites that become dull and stale will wither, and as they decay may well spin off new, small sites. This cycle of regeneration, if it can be sustained, will keep the Web interesting (and volatile).