How many times have you searched for something and come up short? Searching for content over and over, with hundreds or thousands of results, is all too commonplace. It is inefficient, ineffective and frustrating. Popular search engines are proud of the multiple-millions of “results” they achieve in a few milliseconds time, but is this really what we want? Just because a search engine “can” produce thousands of results, doesn’t mean it “should”. Most people are searching for a single result, not thousands. That said, the purpose of these search engines is not to give you the single piece of content that you want, but to give you all of the potential content that you might be interested in, especially the results that also drives advertising revenue. This scenario is not useful in a business context.
Business users need to find a very small and relevant set of content based on the information they have. Going back to the library example, when I walk into a library to find a book about Java Programming, I want to be directed to the shelves that contain only books on my subject, not books that contain the words Java, which are out of context to the subject at hand. This is why there is a card catalog and a Dewey Decimal System.
A business user is no different. If we are looking for resumes for Java programmers, then we want to be given the small subset of content relevant to Java programmers, not content that includes the word “Java’, which could return documents that include the Indonesian island, different types of coffee or a company that has Java in its name.
Classification systems that use Natural Language Processing and text analysis can provide context to content and therefore, organize it properly. Furthermore, by combining classification with enterprise search applications, businesses can provide a robust and effective conceptual search that delivers a highly relevant set of results.
The Bottom Line: Content must be easily accessible to those who need it.