Get the Picture: Visualizing the Future of Search

Page 2 of 3

Visual Search in the Enterprise
Most search companies have recognized that in order to make money in the search business, they need to move into the enterprise and help companies get on top of the mountains of available information; visually-based search tools are no different. "That's where I suspect these companies have the most traction is when they find the right niche. We have this kind of data that we need to plot a certain way, so let's make it come up graphically," Sullivan says.

David Spenhoff, VP of marketing at Inxight, an enterprise information retrieval company that provides results in a visual format, says, "We're seeing interest in enterprise search due to a recognition that unstructured source is without a doubt growing exponentially." In addition, Spenhoff points to government regulations such as Sarbanes-Oxley and a growing regulatory burden in Europe and the United States. "Being able to identify and categorize information and have it available for sharing, collaboration, and re-use, along with the drive for compliance, means a lot of companies are recognizing that this is a key to the business," Spenhoff says.

Matt Turck, president and chief operating officer at TripleHop, makers of the MatchPoint visual search tool, also sees it this way. "The problem we address is very much a pain point felt throughout a lot of different companies that are information intensive," Turck says. He says the typical TripleHop customer is a vertical such as media, finance, professional services such as law firms, or any Fortune 500 company looking at competitive intelligence.

These organizations need to be able to run searches against a variety of information repositories, including intranets, document management systems, email servers, external online subscription services such as Westlaw, and public search engines, yet they need to be able to get a handle on this vast supply of data in a way that Turck says is difficult to achieve with a text list of results. "Our approach to this is to add a level of intelligence and a form of visualization and clustering technology that is looking at results (summary, title, and other elements). Then we extract from this information, in real time, a taxonomy of results," says Turck.

MatchPoint generates a two-pane results window that has an Explorer-like view of the information on the left and a list of results on the right. The left view takes the information and places it into meaningful categories (folders) on the fly. When a user selects a result and the "article" appears in the right, the program highlights any words that match the search query.

Inxight takes a similar approach with their SmartDiscovery search system. Instead of a group of folders, they offer a cluster graphic, with each node representing a different piece of the taxonomy. Like MatchPoint, Inxight builds this graphic and taxonomy on-the-fly based on the individual query.

Tim Bray, CTO at Antarctica Systems, a software company that applies a graphical view of business intelligence information, started Antarctica based on the idea that, while we have a graphical user interface for the desktop, as soon as we turn to search we essentially return to a command line interface. Bray thinks that it only makes sense to add a graphical layer to search results as well, and he sees business intelligence as an area where visual search companies can make a mark. "The idea was to remove barriers, by providing point-and-click access to information," Bray says. Antarctica does this by building custom maps for each client, rather than working with an out-of-the-box product.

Page 2 of 3