Gerry McGovern // New Thinking

The benefits of a single task-driven classification / navigation

December 20, 2015 -- Gerry McGovern

A single / unified navigation is generally the best option for a website or app.
The answer to complexity is simplicity and the number one principle of simplicity is the removal off all things that are not absolutely necessary. The natural impulse is to add and to complicate. Over time, most websites and apps grow, often in unnecessary ways.
Nothing is worse than the growth in navigation. Often what happens is that the initial navigation is seen to be not working properly. Instead of fixing or replacing this navigation, a new form of navigation is added. The intentions are good but the result is more confusion as the customer now has two competing navigation options to choose from. Over time, more navigation options are added, creating more confusion. 
A single, thoroughly tested navigation seems counterintuitive but we have found that it is the best option for many websites and apps. It seems that GOV.UK are coming to a similar conclusion. “This emphasis on a single unified list represents a shift in our thinking over time,” they wrote in November 2015. “Previously we’d assumed that we needed separate taxonomies for different groups of users and different types of content (mainstream/specialist/policy). We now think a single taxonomy which includes all content and meets the needs of all users will make it easier for users to get to the content they need, and for publishers to classify their content.”
In 20 years of doing navigation design I’ve gone from strong proponent of audience / group / persona based navigation to harsh critic. This approach is great in theory and intuitively it feels right, but I can’t remember a situation where it worked well in practice. 
When we observe people navigating using persona-based navigation we notice the following problems:

  • People don’t like have to put themselves into a particular group. They find it intrusive and a waste of time. They want to complete a task, not tell you what group they belong to.
  • People can belong to multiple groups. We had an agricultural website that had a navigation for “Farmers”, “Exporters”, “Researchers”. But what if you were a farmer who exported and did research?
  • Persona-based navigation requires more content. Instead of having one piece of well-written content, specially written content for each persona needs to be written. In reality, each piece of content tends to be not that different, thus creating overlap and duplication. In some websites, you’re sent down a persona journey only to end up on a generic page that is the same end point for every persona.
  • People don’t know if the persona is for or about the group in question. On a university website we had “Students” and “Professors” but many professors were clicking on the Students link because they wanted to do a task ‘for’ their students.

I always thought there were exceptions. For example, in a government tax website where you had Citizens and Businesses. But I’ve heard from the Canadian Tax Authority that they are finding lots of overlap and confusion with these two personas, and that research is pointing towards a single, task-driven navigation.
Developing a subject-based taxonomy for GOV.UK