Vertical Search Overview
IT.com is the flagship application of our technology that leverages advanced statistical techniques to produce search results tailored to a user's intention. Let us explain.
What is Search?
One can view search from the perspective of how structured or unstructured a query must be, and how structured or unstructured the document set to be searched must be:
An example of a structured document set would be a database, where each document is broken up into many different pre-defined "fields" (e.g. 'title', 'author', etc.) An example of a highly structured query would be a query on a database, or a query involving a lot of Boolean logic.
Using this graph, one can see the two poles of search: highly structured queries on highly structured document sets, or unstructured queries on unstructured document sets. An example of the first is a database lookup. An example of the second is traditional Internet search.
The advantage of the database lookup is that it has perfect recall and perfect precision, meaning that all of the results, and only the results, that you are looking for are returned. But database precision comes at the price of a great deal of pre-processing, lack of flexibility, and a great deal of education on the part of the user.
The advantage of web search is the lack of any of that — documents don't need to be processed into a rigid structure, and users can query them using their natural language, without any training. The problem with traditional search, however, is that you lose the perfect recall and precision of databases: ambiguity in search terms and general limitations of search technology mean that often documents you're looking for are not retrieved, and documents you're not looking for are.
Can one combine the precision of databases with the comprehensiveness and ease of use of traditional search? Yes, with vertical search, understood in a precise way.
What is Vertical Search?
Vertical search adds a third dimension to traditional search: context.
Operating within a given context allows vertical search engines like IT.com to combine the intuitive nature of natural, unstructured search with our knowledge of the searcher's intention, delivering precise results that directly address the problems they're trying to solve.
This technique actually adds information to search results that is useful to the searcher. In the case of IT.com, we use a repository of authoritative enterprise IT reviews and information to answer the question behind our searchers' intent: what solutions are most talked about and respected by peers with deep knowledge of enterprise IT?