We have recently done a significant number of information architecture (IA) projects; IA development and validation. Consequently, it seems an appropriate time to highlight some of the trends that we have observed and share our thoughts on creating an effective IA; as such this article will cover some of the common mistakes we have observed when creating an IA and some strategies for overcoming them. One topic, central to all discussions around developing a website, which will not be covered is internal politics and dealing with internal stakeholder; this is a topic for another time.
Before highlighting the common mistakes it needs to be stated that the role of an IA is to assist users in locating content on a website.
Below is a list of the most common IA mistakes we see:
Lack of clear rationale for content organisation: At times, particularly when websites have grown significantly, it can be difficult to identify the structure used for organising content. When there is no clear logic for users to identify with, it can make seeking content like finding your way in a maze.
- IAs which reflect an organisation’s structure: Organisational structures tend to only be understood by the people within that organisation, not the end users of a website. As such, the end users are not equipped with the knowledge or understanding to use the organisational structure to find content. The users of a website tend not to care or be sufficiently motivated to gain this understanding.
- IAs filled with brands, jargon and internal language: Using language which users are unfamiliar with is likely to make it difficult for users to both understand the content and navigate the site.
- Lack of consistency: Consistency is important because as humans we use our experiences to build an understanding of how to interact with a website. That is, we assume that our learning’s from one section of a website can be applied to others. Consistency can be an issue in 2 ways:
- Across sections of a website; often a problem when sections of a website are developed independently.
- Across different levels of an IA; for example when the topic based navigation menus are used at the top level and then audience based options are used at a second or third level.
There are a number of ways to overcome these and other IA issues. First and foremost, the end user needs to be engaged in a number of different ways. Involving the end user allows you to gather key pieces of information which are essential to avoiding these common mistakes:
- Understand who the end user is: While this seems pretty obvious, there are plenty of organisations who make assumptions about who is (or will be) actually using a website. This is something you cannot afford to get wrong. One of the simplest ways to do this is to run a survey on a website and ask some basic questions about your users. Without understanding who your users are, it is not possible to build an IA for them.
- Understand why people are visiting a website and what they are attempting to do: This information is integral, knowing what users will be using the IA for, and what it needs to facilitate is essential when creating an IA. There are a variety of common ways to elicit this information including online surveys, focus groups and contextual enquiry.
- Understand how users see and think about the content: It is essential that content is grouped in a manner that is logical to the users of the website and that navigation labels are intuitive and easy to understand. Card sorting is an ideal approach for allowing users to communicate how they see the content.
Finally, in order to ensure that an IA is effective it must be checked. Formula 1 teams don’t simply create a new aerodynamic package for their cars and then race them, they undertake hours of wind tunnel testing to ensure that it will work in race conditions. IA validation is the same, can you afford to release an untested IA on an unsuspecting audience? A great way to validate an IA is via tree testing (or IA validation) where the IA is tested without the aid of any design. This is a process which facilitates the refinement of labels and content grouping. It also provides the confidence and evidence that an effective IA is being implemented.
While the suggestions made in this article may seem pretty obvious, all too often we see too many assumptions being made during the process of IA development. By undertaking due diligence and ensuring that the information identified above is in fact correct, an effective IA can be created every time.