Building intranets on the fly: welcome to a whole new world of content automation

Communication World, March, 1998 by Bobby Minter

Now imagine this process is repeated over and over in dozens of different ways for different information - from policy directives to news from world-wide affiliates to create a dynamic, fully automated internal web site, where employees receive information based on who they are, what their interests are and when in the day they might be retrieving it.

Sounds too good to be true, but a new generation of Internet applications for automating content management is making this level of sophistication and page composition a reality, enabling content providers to dramatically reduce the cost of distributing focused messages to targeted users on their web sites.

While these industrial-strength tools for web production are in use primarily in external sites today, they will soon be part of the standard tool kit for intranet development as well. The reason is simple: They are critical to lowering the extraordinary cost of maintaining a content-rich site; a site capable of keeping up with the ebb and flow of corporate change, from information to company demographics, while providing ever-increasing levels of customized information. In fact over the next several years, producing web sites without these tools will be much like trying to produce an annual report using a word processor and spreadsheet instead of professional design tools such as QuarkXpress or PageMaker.

Streamlining Production and Reducing Costs

Research suggests that the high-end content providers on the World Wide Web can spend an average of U.S. $1.2 million in developing a site, nearly a third of which goes to managing content. While many corporate intranets are developed for a fraction of such a budget, the cost of constantly updating and maintaining a site, in terms of money and resources, can severely limit development. What the new applications do is shift the development, from flat and static web pages to dynamically composed pages. The result reduces costs and opens the way for a greater and greater degree of automation in producing the site and customizing its information.

Most corporate sites today follow a fairly traditional model for content production: The company newsletter or magazine, a model most communicators find familiar. Content is created by writers and editors, approved and forwarded to designers for layout, approved again and sent to print. Similarly, most web sites are static enterprises. Content is written, edited and approved, sent for layout and HTML coding, then shipped to a webmaster for placement on a site.

As a result, each viewable page within a site corresponds to one HTML file on a web server. When the site is a small one, say 100 files or so, with few internal links, keeping it updated and under control is a relatively easy task. But few corporate sites are that small or simple. And the task of managing these sites with thousands of files of multiple links and intricate interrelationships is daunting, if not almost impossible.

The issue is further complicated because most site management and development tools today do not address the essential problem with file-based sites. As the company adds files, the site's complexity grows exponentially. Doubling the content base, for example, doubles the navigational complexity and page links, which creates a four-fold increase in the interrelationships among files.

To overcome this critical issue, content management applications change the paradigm for producing web sites from an interwoven series of static pages to a rich complex of information stored in databases and made available when a user requests it.

It is a digital twist on the proverbial question of whether a tree falling in the forest makes a sound. Does a page exist if a viewer doesn't see it?

For example, a page may be defined by four content components - a navigation bar global to the application, a navigation bar specific to the current content, a body section containing the current content, and a teaser section with links to other related content. Yet these elements are stored in distinct databases and not composed to create a "page" until a user actually requests the "page." This modular approach can be applied to any type of web application, whether pure news publishing, customer support or a product catalog.

This means that the job of composing a page with text, images and links is handled by the server. And that makes possible a level of information distribution unavailable under the traditional file-based models.

For instance, consider the savings in time and resources required to gather news from worldwide affiliates and distribute it via the corporate intranet. In the traditional static page model, a correspondent in each division might compose an item, send it to a centralized editor, who would compose each item into an HTML file, then place the new document on the server, updating the links as necessary. Using a content management application, the repetitive tasks of coding HTML pages and updating links would be easily handled by the server. A correspondent would simply prepare an item, complete a form to provide the server with identifying information, and submit the content for updating by the server, with little or no human intervention.