If you think about looking for a bottle of wine on one of your favorite wine sites, how would you begin? You might start by going to the primary navigation and clicking on "wine". Or better yet, you might locate the search box and type in a more exact phrase like "red wine" (probably because they did not have "red wine" in the navigation in the first place). The site will then provide you with a long, list of red wines based on one of any number of ranking systems. Or, if they subscribe to some school of usability, it will be in alphabetical order.

Now you begin to think, "oh great, but I have a limited budget and that bottle of Colgin Cariad Red far exceeds my price range." Not only can the number of bottles in the list far exceed your budget and imagination, but the number of questions you can ask yourself to find your next possible selection of wine to meet your budget is overwhelming.

This is where Guided Navigation begins to excel.

In the navigation scheme (made up of navigators), we can show the user every bottle of red wine as well as a set of related, primary attributes (called "dimensions" by Endeca). These related dimensions (in the form of clickable navigation) might include years, countries, ratings, wineries, price, etc.—all geared towards helping the user refine his/her choice. On top of the related dimensions, the Guided Navigation can dynamically discard invalid attributes.

For instance, if red wine is not made in a particular country, it will not show up in choice to the user. This "pruning" is necessary and will eliminate any "no results" found message. The best part? These navigators and dimensions are derived from the data input by the content managers on your site. There is no need for re-design.

Now, Guided Navigation can organize your next wine choices on several different levels. For example, a wine navigator may group categories into dimensions such as Vineyard, Wine Type, Vintage, etc. The categories within each dimension can be organized in a hierarchy so users have the option of zooming in from an aggregate view (French wines) to a lower view (Bordeaux wines) to the finest grain of detail (Pinot Noir wines). Again, these categories present themselves based on information stored in your database—structured or unstructured.

Most content object models have hundreds of attributes. Guided Navigation shows users only the dimensions that are most relevant for their current results. Data sets might have some dimensions that are available at all times because they are shared by every item. Other dimensions are only shared by a subset of the items and won’t appear until a user sees just that group.