The science of classifying and labeling content is essential to creating a delightful digital experience. It's the intuitive organization of information that allows users to seamlessly navigate through a site without ever feeling uncertain about where to go next or trying to shake a nagging feeling that they're missing something relevant that's been placed in a different section of the site.

The practice of labeling and organizing information, Information Architecture (IA), is certainly a science. However, any time people's reactions and emotions are brought into the mix there's always room for error. To mitigate the risk of launching an information architecture that users don't understand, we use tree testing to measure how users respond to the proposed structure.

What is tree testing? Tree testing is an effective method of evaluating a website’s IA by asking potential users to complete “tasks” using the site’s IA. Tasks should be based on the services or content the site offers its users. For example, a financial services company could have a task asking users to locate information on loan interest rates while a hospital might want to know if users can find the address of their cardiology clinic. Tree testing software allows us to track and analyze what percentage of participants successfully completed each task. It also gives information on task completion times, paths that participants take through the tree, and rates of direct success vs. indirect success.

 

Top 5 Reasons for Doing a Tree Test

 

1. Inform future research and design.

One of the best things about tree testing is that it always provides actionable results. Users’ overall ability or failure to complete tasks provides you with useful information about whether the site’s IA makes sense to real users. For example, a low task completion rate tells you that your site may need a complete overhaul of its IA and/or labeling conventions.

On the other hand, even high task completion rates can signal the need for more research and redesign if other metrics confirm that your site isn’t performing well. Verifying that users don’t have trouble navigating your IA means that problems may exist elsewhere. For example, a strong IA on a site with low conversion rates means that there are likely usability issues elsewhere in the flow. Tree testing can help you get to the root of users’ problems and uncover navigation issues that may go unnoticed with other types of research.

2. Get solid research for an affordable price.

Tree testing is very affordable when you have a limited budget. On average, participants spend 10 minutes completing the tree testing tasks, so compensation and recruiting costs are minimal compared to other types of research.

3. Easily compare alternate versions of an IA.

Tree testing allows you to test more than one IA to see which one performs best with users. This is worthwhile to do anytime you have more than one idea about how the IA should work. Whether you want to test several IAs prior to launching a site or you want to compare a new IA to the existing one, it’s a worthwhile investment to prevent user frustrations from happening in the first place.

4. Ask the experts.

Are you a designer or information architect working on a site full of esoteric or complicated topics? No one expects you to fully understand the material, but that doesn’t mean you can’t ask the people who do. Put your IA in the hands of those who will actually be using it, whether they be biochemists, stock brokers, or any other group with specialized knowledge. When you’re dealing with a site that is tailored to a specific group of people, you want to make sure they are able to find what they need. Cooper uses detailed recruiting screeners in tree tests to ensure you are only getting results from people who are representative of your target audience.

5. Get clear results, fast.

Tree tests are great when time is of the essence. Compared to more qualitative research techniques, tree testing requires very little time for recruiting, programming, and execution. Graphical representations of the data let you see what problem areas exist at a glance. This means less time spent on interpreting results and more time implementing changes on your site.

When you partner with Cooper to conduct a tree test, you get the benefit of having a trained team of professionals who provide expertise in the areas of recruitment, test design, analysis of results, and recommendations. To find out more about tree testing and how it can best serve your needs, send us a message.