Part of the quality of a Web filtering product is the quality of its database (so-called pattern matching). Traditional filtering solutions have focused on the sheer size of the database, boasting numbers from 20 million to more than 35 million ratings. Although these numbers are impressive, what is lacking in those solutions is ratings quality (how accurate) and relevance (how often customers using the solution access the rated site).

In fact, tests have shown that the true test of a database is how often it can service the URL requests from its cache of information, and how quickly.

A superior Web filtering product should be able to service the bulk of the URL requests from its database cache. It is impossible for a database to service 100 percent of all URLs from its database, so a superior Web filtering product must include a second layer of filtering: real-time heuristic evaluation. This second layer reduces the number of noncached URLs down to an acceptable level. At the same time, it is able to provide ratings for many of the sites sent to it.