Engines of Change

Search is dead. Long live search.

Don't call it a "search engine." It's a "knowledge" engine or "decision" engine. The technology continues to move
past hot-linked blue lines into text snippets and expandable video clips. And while executives at Google, Microsoft, WolframAlpha and others declined to speculate on what search will look like 10 years from now, they all agree it won't look anything like it does today. As search grows up, future innovations will expand people's expectations on the types of things that engines can offer. The changes are impressive. Google Squared, a Google Labs project in development, finds reams of statistical and historical data across the Internet and presents the information in a spreadsheet format. WolframAlpha lets people type into the query box "chocolate chip cookies" to discover that an average baked cookie weighing 30 grams contains 6.8 grams of fat. Searching on "Paul McCartney" in Microsoft Bing serves up a list of subcategories that suggest you might really want information on biographies, videos and concert information, rather than where to buy concert tickets.

Google group product manager Jack Menzel agrees search is changing. Today it can solve many queries, but it's far from being a problem solver, he says. In five years, search will travel with you, become much more ubiquitous, and provide up-to-the minute information on news similar to Twitter, only in a browser. Google Wave might be the first step toward this real-time news scenario.

People will default to the Internet to ask complex questions and find quick answers. Although an obvious observation, search "really will have a significant impact on society," Menzel says. Engines will understand context and geographic location better, and become conversational. "My fantasy is you will be able to ask Google to find all the beaches within five miles, and how much it costs to park nearby," he says. "We do a fair job of getting you to the beaches, but can't make the connection to the parking information. Today, it's totally impossible because we don't fully understand how it works."

Google held 81.5 percent of the global search market share in May - compared with Yahoo at 9.39 percent; MSN, 2.94 percent; Live Search, 2.49 percent; AOL, 1.80 percent; Ask, 0.94 percent; and other at 0.11 percent, according to Net Measurements of Mission Viejo, Calif. Microsoft's Bing reached 6 percent market share within hours after launch, according to Vince Vizzaccaro, executive vice president of marketing and strategic alliances at Net Applications.

Life-changing innovations in technology typically occur during economic hardships. In many ways, these Internet engines continue to recover from the 2000 dot-com bubble. As the economy struggles with the recession, Internet engines must innovate or die by perfecting advertising models and technology that bring content to people searching for answers, rather than the other way around.

Microsoft's new engine Bing changes search. Not by catapulting Microsoft ahead of Yahoo and Google, but by introducing a new way to find and display results: Think access to all the world's information, a mantra similar to what Google has been chanting for years. Mary-Pat Gonzalez, senior librarian at the Huntington Beach Public Library, likens Microsoft's search model to functions provided by librarians who conduct reference interviews to help patrons discover and make decisions on what they really want to know. Oftentimes, people aren't able to phrase a question to get the answer they want, so librarians offer suggestions on related topics. A simple concept tied to linguistics, she says, but it works.

Bing, which Microsoft pegs as a decision engine, got its name because people are turning toward the Internet to make decisions, according to Mike Nichols, general manager of search at Microsoft. "We dug into the data and understood where people succeed and fail when using search and realized there were numerous unmet needs," he says.

Those unmet needs turned into logical groupings of topics organized into subcategories, from text to videos. Algorithms were designed to determine probability, even when someone searches on a generic term. Type "New York City" into the Bing search box: Algorithms determine whether the person is looking for hotels, restaurants or shops.

Bill Mungovan, senior director of product marketing at Omniture, believes that to compete with Google and gain between 30 percent and 50 percent market share, Microsoft will have to do a lot more. Mungovan would like to see Microsoft introduce CPM pricing, rather than CPC, for search-brand campaigns. Many search marketers simply want impressions on high volume queries or to support large brand campaigns. "Bing has an opportunity to introduce an additional payment model that allows marketers to purchase impressions," Mungovan says. "That's something that has never been done, but Microsoft has an opportunity to broaden the way we buy search traffic."

The good news is search has matured enough to evolve. The bad news is it could cost advertisers more to run campaigns on engines. Forrester Research principal analyst Shar VanBoskirk believes search engines that serve up "decision-based" results will prompt ad-placement prices to rise. Managing increased demand for search engine inventory will come through a deal with Yahoo. Expect also to see a new approach to seo that will require higher investments to reach natural search results for related subcategories, she says.

Aside from Microsoft, WolframAlpha, the "computational engine" created by namesake Stephen Wolfram, provides what Google can't by serving up objective facts that have clear answers. It is not a search engine, but rather a mixture of natural language processing that understands linguistics, gigantic databases filled with raw information, and a computational engine built on Mathematica.

During a recent live Webcast, Wolfram, founder of Champaign, Ill.-based Wolfram Research, told onlookers his team continues to try and understand why the engine hasn't been able to answer some queries. The group has been tapping databases containing financial information, historical affairs, real estate trends, weather data, real-time flight tracking, obituaries and local data. He said users have requested legal and bibliographic information.

Ads have begun appearing on the right side of the query page under "Featured sponsors," but the service remains awkward, requiring manual input by WolframAlpha staff. Clumsy may say more about bugs in the process than the computational engine. Typing "computers" into the search box, for example, provides no ads, but "computer" returns Lenovo. People won't see intrusive banner ads that "ruin" the Web site.

Don't expect WolframAlpha to take on Google, says Danny Sullivan, editor-in-chief of Search Engine Land. There is plenty of statistical information the knowledge engine doesn't have because they simply haven't added it. That's one reason it won't pull people away from Google. He says it acts similarly to an almanac, but if you can't depend on it to have answers for everything, it won't become your first choice.

But look past WolframAlpha's shortcomings as a start-up to find future innovations. People looking for facts on topics ranging from cars to sports will soon find sports scores and statistical information from WolframAlpha in Champaign, Ill. "Is it imaginable to think we could cover as much information as exists in a reasonable length of time?" asks Theodore Gray, WolframAlpha's cofounder. Answering his own rhetorical question, he says, "Yes, we haven't done it yet, but it's not insane to think we could."

Graphs will become interactive, too, rather than static, Gray told OMMA. Inputting a mathematical equation creates a plot line. But drag the plot line to interact with the computation and the calculation changes. WolframAlpha plans to publish a XML applications programming interface (API) that people can query using code on their server to access data from WolframAlpha. The interface for programs, rather than humans, will provide the link for subscription-based services that could crunch numbers and return information similar to LexisNexis.

Gray says it's not unrealistic to think WolframAlpha would integrate with Google. A Mozilla Firefox add-on tool already allows people to receive results from both search engines simultaneously. "We're emphatically not talking about deals that may or may not happen," he says, pointing to Google cofounder Sergey Brin's summer internship at Wolfram years ago. "It's clear that people appreciate Alpha is something Google doesn't do, so there are some obvious things that one might imagine happening."

Next story loading loading..