Microsoft Flexes Search Semantics With Powerset Buy


While financial terms of the deal were not disclosed, it's rumored that Microsoft scooped up Powerset for more than $100 million to complement its search offerings in the wake if its massive bid for Yahoo going bust.

Last month Yahoo officially rejected Microsoft's $47.5 billion bid and then later brushed off $9 billion for just Yahoo's search business. The Yahoo offer was a bid by Microsoft to help it compete against search giant Google.

Powerset is the latest in a string of search-related acquisitions Microsoft has made recently. In January, Microsoft said it would buy Norway-based Fast Search and Transfer, a Web search services provider used inside corporations, for a price tag of about $1.2 billion.

According to a blog post by Satya Nadella, senior vice president of search, portal and advertising at Microsoft, Powerset will join Microsoft's core Search Relevance team and will remain intact in San Francisco.

Sponsored post

"Powerset brings with it natural language technology that nicely complements other natural language processing technologies we have in Microsoft Research," Nadella said.

Powerset basically takes the meaning of words and finds related concepts. Essentially, users do not have to type the exact word they search for, instead, the search results presented are based on semantics, hence the name "semantic search." Microsoft said a shared vision and passion for search were catalysts behind the Powerset buy.

"That shared vision is to take search to the next level by adding understanding of the intent and meaning behind the words in searches and Web pages," Nadella wrote.

Nadella noted that roughly one third of searches go unanswered on the first search and first click, and that users typically don't find their desired information until after multiple searches and clicks. Mainly, he wrote, that problem can be attributed to differences in the phrasing or context between a user's search and they way the same information is expressed on Web pages and the lack of clarity in the descriptions for each Web page in the search results.

Nadella said most search engines primarily match words in a search to words on a page, while the Microsoft/Powerset pairing is geared toward working to understand the intent behind each search and the concepts and meaning embedded in a Web page, therefore offering higher quality search results and more flexibility in how users can phrase their search terms.

"We will use knowledge extracted from Web pages to improve the result descriptions and provide new tools to help customers search better," Nadella wrote, adding that "working with our existing search team and other Microsoft teams that focus on natural language, Powerset will help us address all of those problems and opportunities."

Nadella added that Powerset brings to Live Search a staff of more than 50 including a team of engineers and academic experts in natural language processing and computational linguistics with a range of experience from other search engines and research organizations like PARC.

In a blog post discussing the deal, Powerset product manager Mark Johnson also said Microsoft and Powerset share a similar vision of how future searches should be conducted and noted that the Microsoft buy will alleviate some of the challenges faced by the startup, such as the massive expense, engineering efforts and computing resources required to build a large-scale semantic search engine.

"Powerset has always been a small company with big dreams, with the ultimate goal of changing the way humans interact with computers through language," Johnson wrote.