Google (GOOG) dominates Internet search, but a growing number of companies are trying to come up with something better. On May 15, British mathematician Stephen Wolfram plans to launch an online service intended to provide more useful answers to search queries than the standard list of Web pages. IBM (IBM) just took the wraps off a computer program designed to field questions well enough that it can compete on Jeopardy! with the game show's best human contestants. And Microsoft (MSFT) is planning to relaunch its own search service this spring, though the details are top secret.
Why challenge a company that has crushed every contender to date? Certainly, rivals want a slice of Google's $20 billion in search-related revenue. But they also see that search has loads of room for improvement. Too often, search engines return a list of Web sites people must comb through for information, and that's if they get the right sites at all. The Google challengers, as well as Google itself, aim to divine the essence of what people are searching for and to provide answers that come closer to what they actually want than the standard list of blue Web site links.
These services have been hyped as "Google killers," but they're really no such thing. Google remains essential to most Internet users, and the new services are often more complementary than competitive. For its part, Google has been relentlessly tweaking its own search engine, trying to prevent others from getting a toehold. "It's a huge challenge for anyone to break the Google habit," says Danny Sullivan, editor-in-chief of the Web site Search Engine Land. "But there are a lot of things Google doesn't do that are necessary, so there is a need for alternatives."
The most ambitious is WolframAlpha, which takes a new approach to collecting information and presenting it. Its staff of 250 culls government and other public databases and crunches the data so they can be presented quickly as useful facts and figures. The idea is to "give everyone expert-level knowledge of everything," says Wolfram.
For instance, a search on "New York Tokyo" gives you the populations of both cities, a map of their locations, and the estimated flying time between them. Search on "$100,000 at 5% ARM," and you'll see tables on mortgage payments and principal balances over time for several interest-rate scenarios. "When you're doing [traditional] search, you're being told: 'Here are some places where you might look,' " says Wolfram. "We're trying to compute answers for questions that people have." It is hardly an all-purpose search tool, though. For many everyday queries, say, "Chicago restaurants," WolframAlpha produces few helpful results or nothing at all.
Other startups are staking out new territories of data that Google hasn't yet conquered. Twitter, which lets people post short public messages about what they're doing or thinking, has just added a way to search all posts. Twitter has quickly become the go-to place to find out what's happening in real time—from airplane crashes to the latest Apple (AAPL) rumors. Because it takes hours or days for Google to index most Web pages, the search giant's results generally don't offer the same immediacy.
Perhaps the most promising new search enhancement is bringing people and their knowledge and contacts more overtly into the search results. The startup Aardvark, for instance, whose staff includes five former Google employees, lets people send questions by instant message or e-mail to friends whose social networking profiles show they're knowledgeable about particular subjects. Ask Aardvark what's the best off-road bicycle to buy, and a friend who's a cyclist might answer with a model and a good local store. Another startup, Mahalo.com, uses a staff of people instead of computer algorithms to organize search results for the most popular search terms. That helps eliminate unhelpful sites and save time.
Then there's Microsoft, which has tried unsuccessfully for five years to slow Google. When the software giant relaunches its search engine later this spring, it's likely to focus on giving people all the tools and sites they need to accomplish a given task, such as booking a hotel room in the city they're flying to. "[Users are] looking for insight and knowledge rather than just links that navigate you to a Web site," says Yusuf Mehdi, senior vice-president of Microsoft's Online Audience Business Group.
Google isn't standing still. Already, its searches usually return not just text links to Web sites but also photos, maps, and other information. On Mar. 24, Google unveiled a technology that helps it understand the meaning of words and their associations with similar words. For instance, if you search "Star Trek," you get videos, news on the new movie, and a list of related searches at the bottom of the page, such as "spock star trek." "You should expect to see great new things in the next few years," says Udi Manber, Google's vice-president of engineering for core search. "We should just solve your problem."
But the main reason the new services face steep odds is that they're not yet businesses—or not lucrative ones. Google's huge impact was not so much in search technology but in perfecting a way to make a lot of money by matching relevant ads to search results. Wolfram says his company can profit in several ways, including by posting advertising alongside search results and by licensing WolframAlpha technology to companies to help them crunch internal data. He says he may yet turn to an established Internet company, such as Google or Yahoo! (YHOO), to develop the advertising business. "There are certainly discussions along those lines," Wolfram says. "We're open to all sorts of partnerships."
Business Exchange: Read, save, and add content on BW's new Web 2.0 topic network
Search Engine Face-Off
How does the new WolframAlpha compare with Google (GOOG)? Technology Review put the two search engines through several tests and gave results side by side. One search was for "Microsoft Apple" and another for "Sydney New York." The comparison underscores that Google and WolframAlpha try to deliver very different information for users.
To see the Technology Review comparison, go to http://bx.businessweek.com/search-engines/reference
Hof is BusinessWeek's Silicon Valley bureau chief.