Structure

by Geoff Hart

Previously published as: Hart, G.J. 2001. Structures should pave the way to data. Computerworld Canada, Nov. 16:12.

This article is part 3 in a four-part series on the keystones of a successful online presence: content, structure, and relevance. View the introduction or the previous article in this series.

What I’ve called structure comes under many names, the most familiar of which are probably “hierarchy” or “information architecture”. Whichever word you use, structure encapsulates the relationships between the components of a site that people use to find their way to the information they seek. Structure is simple to define but can be devilishly tricky to create. A succesful site structure must create what psychologists refer to as a schema: a mental model that visitors can use to understand where you’ve hidden the content I discussed in the previous column.

To create a successful structure, you must understand your audience well enough to know what kinds of schemata they can benefit from. Structures fall into four broadly familiar categories, each of which is based on a schema so familiar that readers can use it without much thought to find what they’re looking for: ordered (e.g., alphabetical or numerical), functional, hierarchical, and web.

Ordered structures use our knowledge of some inherent order to help us know where to look. Alphabetical structures are the most familiar, and use the well-known order of the English alphabet to find information that can be defined by its position in the alphabet; dictionaries and encyclopedias are the obvious examples of such structures. Numerical structures rely on the familiar sequence of numbers; products grouped by price and historical data arranged in timelines are obvious examples. Many other orders exist, including physical (e.g., a geographical map, a blueprint of a house), logical (e.g., you can’t print your document until you’ve turned on the printer), and organizational (e.g., the well-defined ranks in a military organisation).

Functional structures rely on the human ability to group things based on similarities in their function (purpose). One familiar example might be the Web site of a company that sells several products, and devotes a separate part of the site to each product. Companies that sell computer-related materials on the Web, for example, typically divide their site into at least three different functional groupings: hardware, software, and consumables. An equally logical functional structure might be to gather all the products for printers (the printers, paper and ribbons, and printing software) in one area and gather storage products (disk drives, diskettes, and disk utiliity software) in another. Which approach makes the most sense depends on how your audience will approach the site.

Hierarchical structures depend on our ability to recognize how broad groups can be subdivided into narrower categories, each of which can in turn be subdivided into other categories, with the items grouped under these categories becoming more similar the further down the hierarchy you go. Organization charts represent a good example of a hierarchical structure, though as I noted above, they can also be ordered if the hierarchy is clearly defined. Hierarchies differ from ordered structures in that their order is arbitrary, based on degrees of similarity rather than a universally acknowledged sequence; for example, technical communicators fall under product development in some companies, under sales and marketing in others, and stand as their own department in others.

Web structures are the source of the name for the World Wide Web itself: highly interconnected, with a bewildering variety of links between related topics. In web structures, as in the Web itself, paths potentially exist between any two related topics. The problem with such structures is that their unparalleled flexibility comes at the cost of unpredictability: nobody knows all possible paths, nor even the best path to a specific piece of information.

In practice, most sites combine all four structures, with the most appropriate structure chosen for each component of the site. To link these structures, you can rely on familiar, time-tested schemata: a table of contents (such as a site map) to provide a high-level view of what structures exist on your site, and an index that provides a low-level view of individual topics for those who aren’t interested in the big picture.

View the final article in this series.