Tag Archives: Google

Since the rumor about Microsoft’s $100M acquisition broke last week (now no longer a rumor, but a fact), I’ve had several people ask me on my take of where Microsoft is headed with this new tool in its search toolbox. First let me frame my angle here. Over the last four years I’ve worked in the small but interesting enterprise software space of knowledge management. One of the key elements of successful knowledge management is the ability to ‘find’ knowledge. As I recently wrote in a KM World article (Your Customers can Search, but do they Find?), finding implies a level of intelligence beyond simple keyword search. This is where natural language processing (NLP) technologies come into play. As with the field of knowledge management, NLP has been in and out of favor over the last decade. With the Powerset acquisition, Microsoft is clearly betting that NLP is not only back in favor, but Powerset’s brand of NLP is the best available in the market. That being said, here’s my take:

Most of the quick analysis of Microsoft’s move was focused too narrowly on the entire Yahoo acquisition drama, and Microsoft’s attempts to challenge Google, but Microsoft’s own Don Dodge stepped in with an excellent post of where the real potential of the Powerset acquisition lies:

1. Powerset technology is more about indexing the content and understanding its meaning, than the query itself. This has enormous implications.

2. There are many lucrative markets for this technology…not just consumer web search.

The second point is worth noting first. Of course Microsoft will use Powerset to enhance its struggling consumer search properties, but incorporating Powerset into its many enterprise applications has as much potential as a consumer solution. The enterprise approach can manifest itself in two of Microsoft’s most critical corporate infrastructure properties, Sharepoint and Exchange. Once a clunky file sharing server, Sharepoint has evolved into a venerable knowledge management platform that can also handle many modes of collaboration. If Microsoft incorporates Powerset’s NLP into Sharepoint, that platform will emerge as a serious threat to many of the pure enterprise knowledge management, content management software vendors that currently dot these markets. Many of these vendors routinely sell their technology as a replacement for the inadequate built-in search for Sharepoint. A Powerset integration could change the equation. Additionally, incorporating Powerset NLP into Exchange, Microsoft’s anchor platform in the enterprise, could add a layer of intelligent search that has yet to be addressed by other vendors. While it may prove impossible to regain significant marketshare in the consumer world, Microsoft has a significant opportunity to consolidate control in the enterprise knowledge management arena. Powerset’s technologies could play an integral part of this consolidation.

Now, back to Dodge’s first point. Indexing and enabling textual search on content is a relatively old, and easy task. Grafting meaning to that index is where the game changes. Intelligent search, or search with an implied meaning, will return ‘answers’ as opposed to ‘results’. There is a significant point to be understood here. In the traditional mode of searching, search engines are designed to bring back the entire subset of content where there are keyword matches. This can result in hundreds, thousands, and often times hundreds of thousands of ‘hits’ that are returned. Google’s strength is identifying these hits and then representing the most popular results at the top of the result set. Intelligent search is different. In fact a good measure of intelligent search is how few results are returned. Since intelligent search first attempts to find meaning on indexed content, the resulting hits returned by an engine like Powerset should only include that content which is in context to the entered query. This can have a huge impact in enterprise deployments of traditional search, but as Dodge says in his post, the utility of receiving ‘answers’ can be extended beyond content search to include advertisement targeting, and other enterprise focused solutions.

While it’s too early to know exactly what Microsoft will do with Powerset, I think it is as important to watch the enterprise software angle on this acquisition as the consumer angle.