Part of our series on the User Experience for Europe.

Deep vs flat hierarchy

A key decision we had to make early on is how we want to approach the general architecture of our site. In the past, websites have often been built around what’s called a “tree structure.” A pre-defined network of pages discreetly linking from one another, getting more and more detailed as you go.

It looks like this:

Tree structures are great for organising information, but when designing large platforms they have two major limitations:

Maintainability - introducing new content often requires restructuring entire branches or trees. With the amount of content at the EC, this can be an enormous job.

Overlap - the same, or very similar content can fit onto many different branches. Or users may want to find that content via different paths, not just the branch you specified.

You can avoid these issues by eliminating a rigid tree structure and creating a network of interlinked content instead. This entails creating consistent content types (one template for all new announcements, all policies, all documents, etc) and classifying them with descriptive labels. Check out our previous post on how we’re classifying content.

This network-structure is also much easier to maintain and faster to change. If a label isn’t working, you just change it once in the classification and it’s updated everywhere. If you want to move content, you simply update its classification and it moves automatically.

In addition, if one piece of content is relevant in more than one area, or users tend to look for it in different ways, you can simply tag that way without having to duplicate it in separate areas of a tree structure.

Flat architecture based on content classification better enables us to make navigation structures based on editorial choices, not hierarchy of information. Because in general, users don’t care about hierarchy. They care about content.

Add new comment