Technology

Gunning for Google in Search


Google could be out-innovated if it ignores new technologies, including Powerset's semantic search and Microsoft's specialization

Why is Google (GOOG) still looking over its shoulder? The search giant has just locked up a deal with Yahoo! (YHOO) to run search-related text ads on its closest competitor's American and Canadian sites. That leaves Microsoft (MSFT) as Google's largest rival, with just 6% of all U.S. searches performed on its site. Google lays claim to 68% of U.S. searches and Yahoo has a 20% share, according to a June 10 report by research firm Hitwise. Yet, Google isn't relaxing. "We have to be on our toes all the time," says Udi Manber, head of Google's search quality team. "It is absolutely the case that a new technology can come along."

Manber knows an army of would-be competitors still live to fight another day. They have their eyes on the search advertising market, projected to reach $25.8 billion this year and $51 billion by 2012, according to research firm eMarketer. And they are dedicating significant money and manpower to developing innovations that could change the balance of power in search within the next 10 years. "Search can actually get better than it is today," says Brad Goldberg, Microsoft's general manager of Live Search. "We look at search as a long-term bet."

Search technology (even Google's) is still in its infancy. The form of online search familiar to most Web surfers still relies on relatively simple pattern matching: marrying query terms to words on Web pages and then ranking those results based on factors such as the frequency of a query's appearance and the number of links to a particular results page. In the future, the underlying technology behind search will almost certainly change. Search sites will implement upcoming technologies that understand human language and, in essence, read pages for answers to queries. They will adopt next-generation voice-, image-, and face-recognition software capable of better identifying multimedia content. They will incorporate systems yet to be conceived.

The result, many experts say, will be Web searches that return more answers and media, and fewer blue links. "Someone could come along and out-innovate Google," says Heather Dougherty, research director at Hitwise.

Semantic Search

One group that believes it has the key to outsearching Goliath is the so-called "semantic search" companies (BusinessWeek, 7/9/07). These companies, including Powerset and Cognition Technologies, use systems that understand the structure of the English language and the definitions of words to retrieve search results.

For example, in a semantic search engine, a query asking "which tennis players beat Andre Agassi?" would return a series of tennis champion names who "beat," "defeated," and "bested" Agassi. Powerset's first link in response to just such a query is a Pete Sampras page. Google, however, doesn't answer the query right. It links to a page on tennis-heroes.net listing the players Andre Agassi beat. "We are going to be looking back five or 10 years from now at the way we are doing search today and it is going to seem primitive," says Barney Pell, Powerset's co-founder and chief technology officer.

The technology behind semantic search engines has been in development for more than 30 years and is just now coming online in a form that consumers can use. In May, Powerset launched a version of its product that only searches through articles in online encyclopedia Wikipedia. Cognition launched a version of its search engine a year ago that searched through legal, health, and wiki documents.

Incorporating New Technologies

Both Google and Microsoft are incorporating some aspects of the technology in their search engines. But while the small, venture-funded firms have little traction in comparison, they believe they have the edge in the long term. "This is not easy," says Scott Jarus, CEO of Cognition Technologies. "It took us 10-plus years to develop the underlying technology."

Of course, should semantic technologies prove to be truly game-changing, there's nothing stopping Microsoft or Google from buying one of these small firms. Google generates around a billion in free cash flow per quarter. Microsoft does, too, and with its deal with Yahoo looking dead, the software giant has a spare $46 billion lying around. In fact, Jarus sees the future of semantic search as becoming incorporated into the big search engines, likely through acquisition. "I think what eventually happens is semantic search gets bolted into the general-purpose search," says Jarus.

What Microsoft or Google cannot simply buy is the open-source search effort led by Wikipedia co-founder Jimmy Wales. Known as Wikia Search (BusinessWeek.com, 1/7/08), the engine incorporates community input with open-source computer programs. The goal, says Wales, is to eventually have a product that rivals whatever Google can develop—using a combination of algorithmic results ranked, in part, with the help of users—but be easily incorporated by large sites across the Web, making Web search ubiquitous rather than largely confined to single sites. "We think it is going to take a while," says Wales. "But, if we are successful, search will become a utility service that everybody offers rather than a single dominant player."

Microsoft's Specialized Effort

Microsoft is taking another tactic to chip away at Google's supremacy: specialization. The company has recently focused its efforts on shopping search, launching a Live Search cashback product (BusinessWeek.com, 5/21/08) that gives users rebates on items they buy when searching on the site. Microsoft only charges participating advertisers when users make purchases, largely eliminating the barrier for Web marketers to try the service since they are, in essence, only paying for sales. "Our goal is to make Live Search the most rewarding commercial search destination on the Web," said Microsoft founder Bill Gates in a statement.

If Microsoft is successful, it will gain a share of the most lucrative queries on the Web because searches clearly tied to commerce often command the highest prices, says Microsoft's Goldberg. So, for example, a search for "iPhone accessories" may command more than a query for "who was the 23rd President?" "Commerce queries are about 30% to 40% of all queries, but they are a higher part of the revenue," says Goldberg.

Microsoft could use a shopping success to encourage users and advertisers to try other specialized search sites focusing on areas such as entertainment, navigation and local search, and research. "We believe that search will become more and more task-centric," says Goldberg. "The experience will look dramatically different than it does today and it will be customized around the tasks that users are trying to complete online."

Google's Search Relevance

Maybe. The success of Google's universal search product challenges the wisdom of Microsoft's specialized approach. Though topic-specific search engines have had some success in particular categories such as health, travel, local search, and shopping, none have come near the popularity of Google's one-site-fits-all offering. "We want to be very easy and very fast," says Google's Manber. Aside from the health space, Google's category-specific search features are all behind the scenes. Consumers see one central site.

Google won't cede its lead in search without a fight. The company sees search as central to everything it does and has "hundreds and hundreds" of engineers dedicated to improving its search relevance and product, says Manber. Google also has the strength of its brand name that can keep people going to its engine, even if there is a product out there capable of providing slightly more relevant results.

Still, brand name and money aren't enough to keep a new, innovative company from supplanting a leader. Just ask Yahoo.


Steve Ballmer, Power Forward
LIMITED-TIME OFFER SUBSCRIBE NOW

Sponsored Links

Buy a link now!

 
blog comments powered by Disqus