I’ve been highly critical of many of Microsoft’s moves lately, mostly because they have felt like the awkward ramblings of a company one step behind the curve, but the other day Microsoft made, what in my view, is a smart play. They bought Powerset, a San Francisco-based semantic search company and with it, they got themselves a nice piece of technology that could vault them ahead of the pack.

Semantic search involves looking beyond keywords at the meaning of the search itself. When you search for “jaguar” are you looking for information on the car, the operating system or the animal? The context of your search, information about your location and machine, other searches you’ve made; can all give contextual clues as to what your true intent is. At the very least, it can give you tools to begin to break down the search to be more meaningful for you.

It would seem on its face, that this purchase is a pure play to one-up Google by buying this technology while it’s still fresh, but John Blossom speculates on his ContentBlogger blog that Microsoft is really going after Ask more than Google with this purchase. Blossom writes:


But when I asked him directly, Blossom surprised me by concluding that it wasn’t the consumer side of search Microsoft was going after, but the enterprise. He sees Powerset as a compliment to Microsoft’s purchase of Fast earlier this year. “Powerset offers Microsoft a new leg up in value-add search applications using semantic analysis to extract content related to a topic. It's most useful application is likely to be in enterprise search, where there are going to be many documents with a structure that would feed well into semantic analysis tools.”

Blossom thinks the open web would be a greater challenge for a semantic search tool because it involves interpreting all of the nuances of the different types of online content through a common filtering system, but he also sees potential for this technology in mobile applications using voice-activated response to natural language queries. “Whichever the application,” Blossom says, “being able to answer questions triggered by natural language queries is becoming increasingly important in content services."

But Dan Keldsen, who is director market intelligence at AIIM, and recently finished a study on the effectiveness of enterprise search (described here in my EContent article AIIM Study Finds Enterprise Search Still Lacking) isn’t so sure that natural language query is really all that. “From a “pure” search experience, people can barely be bothered to enter more than 1-3 search terms – what would make us think that users are going to type in fully formed questions as a query?” He points that in fact, most search engines have trained us *not* to enter fully formed questions.

In any event, Keldsen is simply not impressed with this move. “All the acquisitions or investments (such as into Facebook) that Microsoft has made in the consumer-facing, public web search market - as interesting as the techniques that Powerset uses are (Semantic Web, where art thou?) - this is much ado about nothing. It’s one small, subtle capability that is going to take quite some time (if ever) to make a significant impact on this aspect of the MS search business.”

I’m not so sure I agree completely with Keldsen here. I think the semantic web is the next frontier. People didn’t take Web 2.0 very seriously for a long time. Social networking was for teens, but today it’s on everyone lips. I have the feeling the semantic web or Web 3.0 as it’s been called could change the way we interact with web sites, but I do agree that it’s true potential is likely well into the future.

For now, I think Microsoft was wise to be forward-thinking. I’ve written before that Microsoft has always been a company to stand back and watch, then react to trends. In this case, they are getting ahead of the curve, and even if Keldsen is right and it’s well off into the future, I think Microsoft should be applauded for taking an educated risk. After all, what have they got to lose?