Oh What A Feature: Functional Usability of Web Content Management Systems

Page 2 of 3

May 04, 2005

May 2005 Issue


Browse and Search
If an author wants to modify some Web site information, the first thing she needs to do in a CMS is find the relevant content item. This can actually get tricky, especially in "placeless" environments where content does not map neatly (or evidently) to pages.

CMS vendors traditionally approach this problem by providing "in-context editing," where an author browses to a page and then clicks and icon to modify the content. However, it can be tricky to faithfully synchronize the content management and delivery environments and remember that authors may not always understand (and may never be prompted about) the implications of editing a piece of text that is re-used in multiple locations.

From a usability perspective, don't overlook how frequently your contributors will employ the repository search engine—not just to find individual content items, but to list entire ad hoc collections. They might want to find everything Fred edited last month, just before he left the company. Can your CMS search engine do that? Not all products are set up for this kind of fielded search or accommodate extensions to your custom fields.

Tagging and Classification
Some vendors are working on improved methods for browsing large repositories, such as by employing a taxonomy. Of course, usability greatly depends on how well content is classified in the first place. Without a doubt, one area that raises user hackles is applying metadata to content. To be fair, there's a real conundrum here: classifying content elevates it value enormously, but authors frequently hate to do the extra work, and autoclassification tools remain woefully inadequate for most use cases.

Simply requiring authors to apply multiple metadata fields when entering content into a CMS can brook a serious usability problem, especially for intranets. "Tagging documents and storing them properly in a proprietary repository can be so time consuming that employees will just bypass the tools and go for what works," argues Matthew Clapp, an independent consultant. "I've seen people collaborate via Yahoo! groups because of the time it takes to contribute a single document to the authorized space."

Not every CMS can be so easily bypassed, but large Web sites and intranets with heterogeneous content can mean lots of different taxonomy facets and, therefore, a variety of potentially complex choices to represent in a browser. Hierarchical categories are notoriously hard to navigate via selectboxes. Alternatively, many systems use checkboxes, which better handle tree-like structures, but such lists can run on for several screens depending on the breadth of your taxonomy.

Some CMS vendors are experimenting with systems designed for more contextual tagging—showing only the fields and values appropriate to an author's role, locale in the system, and/or content type. Of course, this approach puts pressure on the design and intelligence of the system to make sure authors aren't confronted with unexpected or inaccurate choices. At the end of the day, we still need humans to add essential context to content, but anything to make it easier will go a long way towards making your CMS more usable.

There is another lesson to learn here: Poor usability often becomes a stalking horse for other deficiencies, particularly those revolving around change management. An author may reject an interface, but the problem may really be that of a new content or organizational model, a corporate insistence on richer metadata, or a standardization of business processes that may or may not be warranted. If the change wasn't explained, wasn't sold, and if people at the front lines find it unnecessary, they may erroneously blame the application.

Page 2 of 3