March 2, 2011 |
Q: What is BA Insight's approach to improving relevance? A New Approach to Relevance in the Enterprise
BA Insight has recently received a patent for a novel approach to improving relevance which I thought, when explained, would help some of you to think about relevance in a different light.
We call our approach AptivRank. The word apt means "quick to learn" and we named it such to indicate it is a self learning-algorithm; not a manual approach to relevance tuning.
In the Enterprise, users don't link or subscribe to content in the same way so Google's PageRank simply doesn't work. So you do you get users to identify relevant content without them feeling burdened?
While users don't link to valuable content, they do act on it. Enterprise search is about finding what you are looking for quickly and leveraging it so that you can complete a task. So action forms the basis of AptivRank. If a user acts on content it is clearly relevant to them. By capturing actions, relevant content can be identified and surfaced to users in subsequent searches.
The challenge for BA Insight was to create a new user interface so that users could act on content in rich ways in the context of search and this is where BA Insight's preview comes into play. In a standard search interface, users are forced to download documents to see if they are useful. After downloading a file, the search engine doesn't know if the user acted on the file. BA Insight preview provides the user with instant access to the relevant portions of the file which can then be acted on directly in the preview.
|
|
|