Internet content filters are implemented in a variety of different ways.  The following four methods have been widely adopted, some in conjunction with one another, by many of the major filter vendors:

Blacklists

Filter developers generate  "blacklists" - files categorically listing all sites that should not be accessible to anyone using the filtering software.  In theory these team will actually look at each page and decide whether or not it is offensive.  It is then placed into one or more categories, such as "Profanity," "Full Nudity," "Drug Use," "Violence," etc.  These categories will be different for each filtering product.  Sometimes the pages are put on the blacklist if their domain name contains certain keywords like "sex", "xxx", etc.  By using keywords to place sites on the blacklist, it eliminates time in looking at each page.  This method raises the issue of problems that arise from blocking perfectly safe and educational pages unintentionally. 

Keyword Blocking

Software developers come up with a list of objectionable words.  This prevents any page from loading if the page contains any of the words in the keyword list.  The problem with this method is that keywords give no context behind the words.  This method blocks many pages that should not be blocked at all like a site about breast cancer.  Although there are methods of writing algorithms to contextualize keywords, the current commercial products fall short of providing for an immense number of exceptional cases where acceptable content might be blocked by an imperfect filter.

Whitelists

These are similar to blacklists except they comprise a list of pages that can be seen.  Some products only have a whitelist giving the customer a very limited view of the internet, but it is very close to 100% effective in blocking all pornography and other offensive material. Again, the whitelist may or may not be published for the customer to view, but usually the customer can edit this list to add or delete certain sites.

Rating Systems