The San Francisco-based startup, a semantic Web and real-world database company, aims to improve those search results Google doesn't do so well, such as finding information about topics that might have two meanings. Take, for example, Apple. The search engine might contemplate whether the query should return results related to the tech company or Apple Records or Apple as in the Big Apple, New York City. Or Boston as in Massachusetts or the band Boston.
Engines have had a difficult time perfecting semantic search. The same word can mean so many different things. Too many sites organize content with words. The idea to make the Web smarter could also give a higher IQ to Google's paid search platform, increasing the accuracy of targeted ads by matching the best keywords with the queries, similar to the way it serves up more relevant information on search queries. It's about leveraging efforts with rich snippets and search answers, both of which aim to give back smarter, more immediate results to specific queries.
Jack Menzel, director of product management at Google, tells us that with the Metaweb acquisition came Freebase, the company's free and open database of more than 12 million things, including movies, books, TV shows, celebrities, locations, companies and more. If you're looking for Barack Obama's birthday, Google delivers the answer -- 4 August, 1961 -- at the top of the search page results.
Metaweb helps the Web map content to an "entity" rather than words, because people use too many varying descriptions to find the same content or entity, which represent real-life things. What about mapping keyword queries to search results or phrases in display ads? That would ensure a much tighter match.
SEO by the Sea Founder Bill Slawski dug out a few patents at the United States Patent and Trademark Office assigned to Metaweb. He describes a patent on "Query Optimization." That's where he found the Metaweb ID number of Arnold Schwarzenegger. The patent filing describes how an ID number can be used to collect and store data about named entities, and information associated with them, and how queries are performed based on collected information, explains Slawski.