Search engines are very good at finding keywords, phrases, images, people and crawling for metadata. What Google, Yahoo!, MSN, and the rest lack is the ability to contextualize their results.

NewsSift (www.newssift.com) is a new search engine from Financial Times that does just that. The new search engine is targeted toward business people and organizes search results by topic, organization, place, person, and theme. The interface looks familiar to those of us that are used to seeing PowerPoint slides.

It’s only in BETA now but I think this is a tool worth watching. To illustrate how it works, if you were searching for "Dakota" you would receive results related to two states, the child actress Dakota Fanning, organizations such as a universities, government agencies, and legislative branches, any organization with "Dakota" in its name.

So instead of getting 100 URL's with a short description of each in response to a search term query, NewsSift lists terms in each of the five categories previously mentioned –topic, organization, place, person, and theme. It provides a (contextual) path instead of a list.

NewsSift uses contextual and relational technology from Endeca that allows it to recognize the difference between Rolling Stone magazine and the Rolling Stones, for example.

Once users have selected their context from the five lists, they can drill-down further by adding related topics, more specific terms, or cross-referencing across multiple categories. For example, a user can search for "Charlotte," which will bring up the Charlotte Bobcats, the City of Charlotte, Charlotte Ross, and the Charlotte Region. If you then add the term "Stimulus," you will be presented with a list articles written about the Stimulus Package and Charlotte. In this regards, the value, if you want to call it that, is the site eliminates long search strings with quotations around some words and phrases but not around others, which is often the most effective way to use Google.

Google is slowly embracing the concept of semantic search -- putting terms in context. However, they have a lot of money invested in their existing architecture and the algorithms necessary for semantic search apparently don't scale to cover Google's billions of indexed pages.

Google did announce recently that queries are going to contain related terms (more than just different spellings or forms of the word searched, as is the case now), which signals that Google is moving toward what NewsSift has done.

NewsSift is hiring technologists while Google is reducing its workforce. NewsSift has a long way to go, but it is this type of innovation and willingness to take on a market leader that will turn this economy around. Will NewsSift take down Google? I doubt it, but it will be interesting to watch.

Update on decline of print media. In an eMarketer sponsored report, Consumer Magazines; Rethinking Paper and Pixels ($695.00) Carol Krol writes: “The numbers are sobering: 525 magazines folded in 2008. As of the publication of this report, so far in 2009, 87 other titles have closed. The bleeding will not be stanched any time soon.”

Steve