In this stage of the web design process we've considered the client's content and looked at different information architecture approaches to ensure usability by providing sufficient "scents of information" to guide the user. To aid us in the process of organizing content and move us towards the web site mapping step let's look at different types of web site structures.
While we may be inclined to use a "web page" as the most basic unit of a web site's information architecture, Jesse James Garrett recommends referring to these units as nodes:
By dealing with nodes rather than with pages, documents, or components, we can apply a common language and a common set of structural concepts to a diverse range of problems. The abstraction of nodes also allows us to explicitly set the level of detail we will be concerned with. Most Web site architecture projects are only concerned with the arrangement of pages on the site; by identifying the page as our base-level node, we make it explicit that we won't be dealing with anything smaller. If the page itself is too small for the project at hand, we can have each node correspond to an entire section of the site.
(Jesse James Garrett, Elements of User Experience, "The Structure Plane")
A web site's nodes can be arranged in a number of different ways, but we can generalize the structures and find examples on the web. A thorough understanding of these structures is extremely useful for planning the structure of a site, and how navigation will work between pages. Lynch and Horton identify three information structures in their book Web Style Guide 2.0: sequences, hierarchies, and webs. Garrett describes sequential and hierarchical structures as well, but he breaks "web" down into matrix and organic.
The sequential flow of language is the most basic type of information architecture there is; books, articles, audio, and video are all designed to be experienced in a sequential fashion. The simplest way to organize information is to place it in a sequence. Sequential ordering may be chronological, a logical series of topics progressing from the general to the specific, or alphabetical, as in indexes, encyclopedias, and glossaries.
Sequential structures on the Web are used most often for smaller-scale structures such as individual articles or sections; large-scale sequential structures tend to be limited to applications in which the order of content presentation is essential to meeting user needs, such as in instructional material which utilize linear navigations.
More complex Web sites may still be organized as a logical sequence, but each page in the main sequence may have links to one or more pages of digressions, parenthetical information, or information on other Web sites:
Hierarchical structures are sometimes called a tree because nodes have parent/child relationships with other related nodes. Traditionally web sites have been organized around a single node: the home page. Child nodes represent a narrowing of information within the broader category as represented by the parent node. Garrett adds, "Not every node has children, but every node has a parent, leading all the way up to the parent node of the entire structure (or the "root" of the "tree," if you prefer)."
This type of structure is the most common, because it's familiar, fits in with our understanding of computers, and is easy to implement and maintain.
No, not the movie trilogy with Keanu Reeves and Lawrence Fishburn (I still cringe over Revolutions), though the "layers" of reality provide an interesting possible analogy for this information structure. The rather uncommon matrix structure in information architecture allows the user to move from node to node along two or more "dimensions." Garrett explains:
Matrix structures are often useful for enabling users with different needs to navigate through the same content, because each user need can be associated with one "axis" of the matrix. For example, if some of your users really want to browse products by color, but others need to browse by size, a matrix can accommodate both groups. A matrix of more than three dimensions can cause problems, however, if you expect users to rely on it as their primary navigational tool. The human brain simply isn't very well equipped to visualize movement in four or more dimensions.
Elements of User Experience, "The Structure Plane"
Organic structures don't attempt to follow any consistent pattern, and so it seems an unusual web site structure, but in my opinion these are becoming more and more common. In an organic structure the nodes are connected on a case-by-case basis based on relationships that may or may not be well-served by the use of "sections" or "categories".
With the advent of "Web 2.0" we've seen an explosion of these organic structures through the storage of metadata, namely tags and keywords. If you were to look at the WWW as a whole, the structure would be organic, with web pages linking to web pages based on relative information, content, sequence, source of origin, etc. Blogs, for instance, utilize some aspects of an organic structure when the author interlinks between different posts within the post, and when the author utilizes a tag cloud as an accessory navigation device.
Garrett adds,
Organic structures are good for exploring a set of topics whose relationship is unclear or evolving. But organic structures don't provide users with a strong sense of where they are in the architecture. If you want to encourage a feeling of free-form exploration, such as on some entertainment or educational sites, an organic structure can be a good choice; however, it can present a challenge if your users need to reliably find their way back to the same piece of content again.
Based on your content outline begin to consider what information structures will be used in your Web site. You will represent the relationships between pages in these structures as you create a web site map.