Cluster Analysis is the "search for homogeneous subsets". I am quoting Kachigan's definition. His book does a fine job of explaining the distance based kinds of cluster analysis (please see my review). In this post, I just want to send you to URLs that provide a fun way of understanding WHY you would want to engage in this process. In other words, what purpose would it serve.
Most folks that I meet that want to learn about Cluster Analysis, have yes/no or likert scale variables in mind. Claritas, the famous demographics company does something rather different, but it is still informative to play around with your zip code, and other zip codes that you know at one of their sites. http://www.claritas.com/MyBestSegments/Default.jsp
Also, one of my colleagues at Overbeck Analytics pointed out this excellent TED presentation. It is great fun, but is an excellent explanation of the value of this method. http://www.ted.com/index.php/talks/view/id/20