I was intruiged by a recent post by Jeremiah at the Web-Strategist.com about how the search box in a browser is circumventing the address field. I read this after sitting with a client today and discussing their mamoth 3.2 million page website. Third client actually, the first one had 100,000 pages, and second one has about 1,000 pages. How do you organize all that content?!
So here is an idea i’ve been toying with for a while. A colleague at Jufa Intermedia and I drew it out on a whiteboard over lunch one day. Should we develop it?
The idea is developing a CMS that treats pages as stand-alone pieces of micro-content. Here is the gambit: the navigation structure is dynamically created, and the pages are easy to find using search tools. So you can both explore and/or search for specific elements. If you don’t know what you want but know the topic you explore, and if you do know exactly what you want you can narrow your search down rapidly. But wait, there’s more…
People don’t usually consume whole websites, they use parts of it at any given moment of need. Search boxes are the second most used navigation element within most sites, yet most site-specific search functions… well, frankly suck.
One of the biggest problems client’s have is
how they will structure their IA:
copy the organization’s operational structure
or
discover how the user consumes the information provided.
The answer is sort of obvious from an effective marketing stand-point – you design for the user. But it’s not that simple, because in larger organizations, the workflow of how content is published does not usually correspond to how the world consumes that information. So if you can’t create good content, then your “effective” user experience is for naught.
So instead of publishing to an area that you are responsible for maintaining, you simply publish for various categories (to seed the choice), you tag it and you tell it who your intended audience is from pre-loaded user-profiles. So you don’t publish to a part of the site, you publish to the TYPE of person you want to read the content.
The CMS will dynamically generate menus based on traffic patterns for each pre-determined user type.
It will guess at what a particular user is looking for based on what the user is looking at and how long they spend looking at it. That’s the first part. Second, is that each page must standalone. So you publish a page so it can be found using an onsite search engine. You can categorize and tag, as you may a blog entry. So instead of a site of pages linked together through a man-made and stagnant architecture, you effectively create a catalog of the content.
Imagine surfing a large site
(1,000+ pages) like you would looking for a book on Amazon.
You have recommendations based on previous experience, you have recommendations based on others experience, you have categories you can browse, and you have simple and advanced search functions for narrowing your results. If the content contains “web 2.0″ user-generated features such as comments, ratings, etc. then they can be integrated into the CMS as well, by helping rank search results, and offering new navigations methods.
So, any VC’s out there want to fund this project? 
Bruce
You might want to take a look at Endeca — it provides “guided navigation” (a form of faceted search) and can be integrated into a number of content management systems. (It actually has a lightweight CMS of its own).
There is also an open-source project called solr, which is based on lucene, that also provides some faceted search.
- Geoff