That 5 users is enough, is an often misused and misinterpreted conclusion from Jacob Nielsen’s Alertbox (dated 19. 3. 2000) "Why you only need to test 5 users?". In some circumstances this is the case, but in many not! Nielsen specifically talks about cost/benefit of software usability study with 5 participants in each of three stages of iterative software development process. And nothing else.
Here is an blog about how many users is enough, based on what kind of study and what kind of questions do we have! Worth reading!
Alan Dix, Are five users enough?, June 4, 2011