Mark Nolan has a nice article in the April/May 2008 issue of the ASIST Bulletin called “Exploring Exploratory Search.”
Citing an article by Gary Marchionini (“Exploratory search: From finding to understanding.), Mark points to three larger classes of behavior: Lookup, Learn, Investigate. Each has subclasses of behavior. These behaviors, however, aren’t linear. Makes sense: we can bounce back and forth between them when searching information.
This recalls Allan Foster‘s nonlinear model of information seeking he proposed in 2004 in an ASIST article. “The behavioral patterns are analogous to an artist’s palette, in which activities remain available throughout the course of information-seeking,” says Foster. He identified three large phases as well, which he calls Opening, Orientation, and Consolidation. Not quite the same, but similar.
Mr Nolan gives the Investigate mode of searching the most attention in his article–and rightfully so. It’s the hardest to understand and to design for. How can people find things they don’t know they need? How can a search system support unknown information needs?
Of course, Donna Mauer describes this mode of searching “don’t know what you need to know” in her Boxes and Arrows article “Four Modes of Seeking Information and How to Design for Them.”
The article ends with some high-level areas to consider in supporting exploratory search. I must admit I was hoping for more than a focus on improved search retrieval systems and better content. What about text analytics and automatic extraction techniques? What about semantic analysis of tagging and the like? Seems we’re still thinking in terms of an active information seeking model for the user when maybe passive models may be more fruitful in exploring information. In other words, people shouldn’t have to find information; the information should find the right people.