Content Management Systems are only ten years old and yet there are already over two thousand types of them on the market. At the low end of the multibillion dollar CMS business there are weblogs, wikis, forums, and news portals--often called CMS-Lite. At the high-end are enterprise content management (ECM) systems, which integrate management of documents, records, customers, and expert knowledge, as well as provide ecommerce support. In either case, we need to structure information in these CMS to create the best interactive user experience. What are the differences between the Big CMS and the Small ones; and do they matter to optimizing the user experience?

First up, a big differentiator is price. Licenses from top CMS vendors like Documentum, FatWire, Interwoven, and Vignette are six figures--sometimes per CPU! Many CMS-Lite products border on free. You need only get an ISP who also runs one of dozens of open-source news portal systems like Drupal or social networking tools like TikiWiki. At the very lowest end, you can set up a low-cost WordPress weblog or even a free one on Google's Blogger.com. If your company already has an in-house IT shop, they are probably now readying blogs, wikis, or forums for your intranet.

The next, and most important, thing to consider is the information architecture of these tools. Big CMS are actually multiple, loosely-coupled, applications all running against a single content repository database (though they may integrate content from separate legacy databases, usually wrapped in XML data structures). Each application may have its own user interface optimized to its special function. These separate applications are often template-driven structures and today are more likely to be an XML application that collects user input via forms to generate valid XML data on creation. A large organization may have an in-house team or a professional services consultancy (often provided directly by the CMS vendor) to design all these templates and user interfaces.

So what are the big differences for the Small CMS?

Small content management systems still have three critical elements--user interface and interaction design, information architectures for their structure, and the core content itself. But the information architecture, the fundamental organizing principle of the CMS-Lite, has been highly restricted. Weblogs, for example, have been optimized and refined for the hundreds of thousands of users of these tools. All content is simply arranged in reverse chronological order and optionally categorized by topic and author. For wikis, organization is achieved by nothing more than simple hyperlinks between pages that are all at the root of the CMS.

Simplifying and standardizing the information architecture lies behind their great market success. What all of the CMS-Lite tools offer is an out-of-the-box well-proven user experience where you simply type in the content, do a modest amount of tagging, and voila--publish to the Web. They also build in RSS syndication that simplifies reusing your content via Web services.

Big IA and Small IA
Unfortunately, the highly successful simplified IA of the popular CMS-Lite tools does not scale well. In ECM there are many more content types than simple blog posts to be managed. And many more kinds of user interface.

This leads us to another structural difference, one often overlooked, between the IA of a document collection as a whole (the entire content of an organization) and the architecture of the individual documents, especially Web pages, in the collection.

The classic book Information Architecture for the World Wide Web (O'Reilly), by Louis Rosenfeld and Peter Morville, sets out principles of organizing, labeling (tagging) navigating, and searching (making documents findable) based on decades of experience in organizing document collections, especially the books in our libraries.

Whether you are organizing the pages in a Web site, the documents in a corporate information collection, or content into a portal like Yahoo and Google (which provide taxonomies for the entire Web) there are a number of basic principles/best practices for this kind of IA that I'll call Big IA.

Some of these principles are evolving, like folksonomies, which employ bottom-up tagging schemes versus top-down centrally designed taxonomies. This will all be reported in the third edition of the Morville-Rosenfeld opus, now in the works. also hope Morville and Rosenfeld will also include a lot more about the Small IA of the individual document, especially the Web page with its navigation elements. The trend in documentation is toward highly structured documents that encourage reuse, using XML and the new DITA standard. We also need a standard page structure description language in XML, one that will have room for all the content elements of a printed document, but also the navigation elements and other user interactions of a hyper-document.

Content managers call this the content model, the allowed elements and their attributes, specified in a document schema. Describing this structure layer, between the content inside and the design outside that provides the user experience, is the next great challenge for Information Architects and CM Professionals, both Big and Small.