Look who's searching
New Economy 101 06/23/99 5:00 PM
By Claude P. d'Hermillon, Jr.
Must read rom Raging BULL
Because so much of what the masses do online involves some type of search, it's pretty clear that technology companies whose products connect surfers with content are in a market niche with plenty of growth potential. That's why the last few months have seen a gaggle of funding rounds, alliances, product launches and even a partial mutiny against a perceived leader.
With the release of its fourth product line last week - one that some analysts say puts that leader ahead of its competition by at least a generation - Inktomi (INKT) is the sector's 800-pound gorilla. Its core products are software applications engineered to boost the performance and intuition of massive computer networks. Though perhaps best known for its more generic search-engine technology, Inktomi's bread and butter comes from its Traffic Server, a highly scalable caching application used by Internet service providers and telecoms to ease bandwidth bottlenecks by storing, or "caching," frequently used documents that much closer to the user.
The company also has high hopes for its soon-to-launch comparison-shopping engine, which will enter a crowded market already occupied by the likes of ExciteAtHome's (ATHM) Jango.com, Amazon.com's (AMZN) Junglee and privately held my Simon.com Inc. Its latest offering, a search engine that categorizes Web content by topic - a task previously done by human hands or taxonomists, as Yahoo! (YHOO) calls them - is expected to boost the company's overall revenue by as much as 25% a year.
With a lineup like that, analysts generally like Inktomi's prospects and expect the company to move into the black sometime next year. So far, the Street has concurred, driving Inktomi's stock to an all-time high of 159 1/8 back in April. That's a more than 1,000% increase since the San Mateo, Calif., company went public last June. More recently, shares of Inktomi have traded down with the rest of the Internet stars, though last week saw an uptick to its current level around 117.
A crowded field
The search-engine industry has grown crowded almost overnight. Though counts and even definitions of what constitutes a search engine vary, there are more than 10 companies with some sort of a stake in the search-engine arena.
They include ventures that have mutated their search models into portals, such as Infoseek (SEEK), Lycos (LCOS) and ExciteAtHome, along with a slew of new ventures with names like Google.com and Yep.com.
Generally speaking, traditional search engines fall into one of two technology camps. Sites like Compaq's (CPQ) AltaVista and Infoseek locate relevant information by deploying spider-like technology that scours or "crawls" the Web for matches. The other style of search, known as indexing, takes a more subjective approach, deploying teams of editors to identify and index sites by category. That is the technique Yahoo pioneered way back when.
While much of today's search technology continues to rely on hit-or-miss keyword searches, which (Boolean delimiters notwithstanding) return hundreds if not thousands of off-topic results, armies of software engineers are scrambling to refine both accuracy and usability.
A different kind of browser
Among the biggest technological trends these days is a push to take restrictive search technology to the next level by focusing not on keyword matches but on browsing habits. The theory is that tapping people's opinions about a certain Web page or database will enable search engines to return more authoritative lists of sites that have proved popular with surfers looking for similar information. That, of course, assumes that the majority's opinion is worth mirroring.
At Google.com robots crawl the Web searching for clues about a particular site's popularity. The technique, known as "link analysis," rates a site based on the number and types of pages that contain links back to it. It may conjure chicken-or-the-egg questions, but several of Silicon Valley's most notorious money men think Google.com has merit. Earlier this month, venture capital powerhouses Kleiner Perkins and Sequoia Capital bought stakes in the venture worth an estimated $25 million. With business cards that might as well read, "the savviest tech investors around," these firms were the original sugar daddies to Amazon.com and Yahoo.
Strangely enough, though, no one associated with the venture will say much about the upstart's strategy. Judging from the slivers of information the outfit has circulated - Google.com is not a portal, won't feature content, and has no plans to compete with other search engines for portal business - some analysts are beginning to wonder just what lies behind the curtain. Aside from the fact that it has been buying mass quantities of computers, and that it plans to quadruple its staff by year's end, very little is known about how Google.com intends to make money.
At the same time, International Business Machines (IBM) has a search engine of its own under development, which it's beginning to shop around. Clever, as it's dubbed, is also said to utilize link analysis, though IBM says its technology will make the product far more effective than Google.com. No word yet on deployment, although the company says it's talking with portals about possible licensing deals.
Then there's Direct Hit, which launched its so-called “popularity engine” last August. Run by Direct Hit Technologies Inc., the unique search engine looks at the sites users choose from traditional keyword searches. By examining the amounts of time users spend on sites culled from standard queries, Direct Hit doles out popularity rankings based on stickiness, or time spent on a particular page. Back in February, the Wellesley, Mass., company generated a fair amount of buzz when it replaced Inktomi as the default search engine on HotBot, the site soon to be owned by Lycos. Inktomi continues to power some HotBot searches, but only for the most rudimentary queries.
There's also GoTo.com (GOTO), the newly minted public search engine that - gasp! - lets advertisers bid for priority placement in keyword searches. That concept caught so much heat last year when it was first rolled out that many thought it would quietly disappear or quickly remake itself. But when the Pasadena, Calif., company went public last Friday, the only gasps to be heard came from skeptics on the sideline who thought the market for cutting-edge Internet stocks (never mind a search engine) had dried up. Not so. In a mixed day for other initial public offerings, GoTo.com saw its market value double to $971 million by the end of its first day of Nasdaq trading.
The search-engine landscape looks promising, as hordes of programmers driven by crazed product managers starved for the next killer app, peck out tedious strands of code that might some day make life online a whole lot easier. Does that mean the geeks are on the verge of a truly liberating technology that will make today's cumbersome engine seem like it was programmed with a hole-punch card? Don't bet on it.
The more likely scenario, industry watchers say, is a gradual refinement of a few fundamental principals that are currently in vogue. Whether that work is done by an industry powerhouse like Inktomi or by some grad student working on a thesis-cum-business-plan is anyone's guess.
Nevertheless, it is widely believed that the next decent step in search-engine science will lead to a product that's an amalgam of artificial analysis and hands-on indexing.
- Talk Back: Chat with Author Claude
|
