Tree Testing: A Practical Guide to Streamlining Information Architecture

By conducting tree testing, you can determine how quickly and easily users are able to find information on a website. In this article we explain what tree testing is, what the pros and cons are and how to get started using our step-by-step plan. 

The advantages of tree testing

There are several benefits to performing tree testing: 

• Insight into the usability of the information architecture. Tree testing provides insight into how well the information architecture of a website or app is structured. This reveals which parts are easy to find and where improvements can be made. 
• Saving costs. Through tree testing, problems can be detected at an early stage. In practice, this often means that solving the problems is less time-consuming, thus saving valuable development time. 
• Improve conversion. Not being able to find the right information is one of the biggest annoyances for many users when using a website or app. By solving these problems, this can be overcome. Because users can find the desired information faster after implementing the solution, there is often an increase in the conversion rate. 
• Comparing different designs. Tree testing can be used to compare different designs of a website or app. This can be used to see which structure works best for users. 
• User feedback. Tree testing not only provides insight into the usability of the information architecture, but also into what users find important and how they categorize information. 

The disadvantages of tree testing and how to overcome them

One disadvantage of tree testing is that it is not always the most natural way users search for information. For example, they have to navigate through a tree structure, whereas in reality users often perform searches or find information through filters. 

To overcome this drawback, it can help to supplement tree testing research with other research methods, such as usability testing. This will give you a more complete picture of how users behave on the website. 

Another disadvantage of tree testing is that it can sometimes be difficult to set up the right tree structure. It is therefore important to take enough time and consider multiple options when creating the tree structure. For example, you can choose to test multiple variants.

The link between card sorting and tree testing

In card sorting, users organize the categories and subcategories of a website and put them in order of relevance. Tree testing then examines how easily users can find the right information based on the structure created. By combining these two methods, an optimal information architecture can be created that fits well with users' wishes and needs.

How do you conduct tree tests?

There are several ways to conduct tree testing. Below we cover the different variants.

Almost always online 

Tree testing is often performed without a moderator using a tree testing tool. This means that participants run the test from their own environment. This allows more participants to participate in the study and the results are often more representative. In addition, it is often possible to conduct the test online, which means that participants do not have to physically be present at a specific location. However, it is also possible to add moderated sessions for in-depth coverage. 

Without a moderator (unmoderated) 

In unmoderated tree testing, participants are shown a tree structure and asked to indicate where they would expect to find certain information. This involves tracking how long it takes the participant to find the information and whether it is the right information. The results are then collected and analyzed. 

With a moderator (moderated) 

Moderated tree testing can be conducted in several ways. For example, the moderator can observe the participant as they perform the test and ask questions about the participant's experiences. Also, the moderator can go through the tree structure with the participant and ask for feedback on the structure. 

How many testers are needed for tree testing?

The number of testers required to perform tree testing depends on several factors. These include the purpose of the study, the complexity of the test and the time and resources available. 

Similar to conducting card sorting, it is recommended to conduct the study with at least 50 testers to get a representative picture of how users use the website or app and search for information. 

When recruiting testers or respondents, it is important to make sure that there is sufficient spread in the selection. For example, in terms of age, gender and experience with the website or app. This way you ensure that you have a balanced and representative dataset.

Tips on how to conduct tree testing

Want to start tree testing yourself? Here are five tips to get the most out of your research.  

• Make the tree structure clear and logical. Make sure the tree structure is clearly organized and easy to understand for the testers. Avoid unnecessary subcategories and keep the structure as simple as possible. 
• Ask the right questions. Make sure the questions asked to the testers are relevant and clear. Make sure the questions do not lead to suggestive answers and that they provide the insights you need to improve the user experience. 
• Choose the right tree testing tool. There are several tree testing tools available. User Sense also offers tree testing.
• Recruit the right respondents. To get representative results, it is important that respondents fall within the target group and that there is sufficient spread in the selection. This could mean, for example, selecting testers based on their age, gender, demographics and/or other relevant recruitment criteria. 
• Analyze the results carefully. Carefully analyze the tree testing results and draw conclusions based on the findings. Be sure to compare the results with other studies, such as user profiles, statistics and user feedback. This can help you get a complete picture of the user experience and where improvement is needed.