Introduction
Hi, I’m Ayman Zahir, and I work as a Junior UX Researcher in the Digital Services team at the University of York. Recently, I’ve been involved in the new VLE deployment and wanted to share what I’ve learned with you. In particular, I want to introduce you to Tree Testing and how we use it to better understand our users.
What is Tree Testing?
Tree Testing is a powerful method for evaluating the navigation and hierarchical structure of your website or digital system. In a tree test study, participants are presented with a tree-like structure that represents the evaluated site. Participants are then asked to navigate and find locations in the tree where specific tasks can be completed.
It’s a method that tells you how easily and intuitively users can find information on your system by stripping it down to its organisation, categories, and labels, free of any user interface that might add bias. If users get lost, it tells you exactly where that is.
Tree testing can help answer the following questions:
- Do users understand labels as they’re intended?
- Is content split into categories that seem natural to users? Is it categorised logically for users to easily understand?
- Are users reliant on learnt understanding or retained?
- Can users find the information that they want easily and quickly? Are they looking for it somewhere else?
- What is stopping the user from finding the content they are seeking?
Why did we decide to use it?
Tree Testing definitely has a specific use case. In fact, until July of this year I hadn’t used this technique. But the VLE redesign and restructuring of the Navigation Bar provided the perfect opportunity to test this out. My work involved evaluating whether these new changes had improved findability, benchmarking against the previous to measure any improvements.
It’s especially useful very early in a project, when you’d like feedback on the proposed Information Architecture, but don’t have any prototypes or interface to test on. It’s really quick and easy to set up, allowing anybody in the team to give it a go. You can run a Tree Test online or in person. All you need is a printer, scissors, a desk and a few willing participants.
Pinpointing the Root Cause
In a usability test, factors such as animations and UI need to be considered and meticulously planned around when designing navigation tests, thus making it harder to pinpoint the root cause – whether navigation is difficult due to the Information Architecture or the User Interface.
For example, from the VLE Digital Experience survey, we discovered that students hadn’t been visiting an important section of the VLE as frequently as expected. But the survey has limits. A gap in our knowledge remained, why didn’t they visit certain sections of the system. At this stage we could only speculate. For example, it could be an Information Architecture issue, is vague labelling misleading? Or do we have a UI issue such as – users failing to notice the link to the sections due to its placement and presentation? Deploying tree testing alongside usability testing allows you to uncover the why (root cause) behind any self reported behaviours.
How we did it
Our Tree Test study was done using an industry standard UX toolkit, but plenty of free and paid for options are available. The tool allowed us to mimic the VLE site structure (see example below) alongside a set of tasks for the participant to follow.
I’ve identified a few tips when generating the task list:
- Try to form a balanced task list of no more than 10 tasks that encompass the most important things you’d like to test.
- There should be a good balance between simple and complex tasks to retain user engagement during the test.
- Ensure tasks are easy to understand while providing real user contexts
- Tasks should be generalised enough to allow exploration without inadvertently biassing their answer.
- Tasks with multiple answers allow you to understand the most popular paths taken.
To create the tree, it’s good to have a full understanding of the navigational structure which you intend to test. I found it useful to map out the different navigation elements first – the navigation bar labels, links, buttons, and other interactive elements along with the actual content users should be seeking. Remember, don’t just include the answers to the test but add dummy answers that accurately reflect how the content may be interacted with.
How it went
Typically, you want around 40 to 50 participants in your study. For our study we deployed 32 unmoderated sessions to uncover general trends and 10 moderated sessions which provided opportunity to probe into some select behaviours exhibited using post interview sessions. From the results we found that:
- 64% of participants failed to find the section academic support when tasked
- 60% of participants incorrectly thought notification and settings would be in the Tools (38%) or Marks (22%) area.
- When tasked to find a digital book to read, 57% navigated to Courses followed by Reading Lists to find it.
Next steps
Once you’ve uncovered opportunities, you’ll eventually turn some of these into potential improvements. Re-running Tree Tests will allow you to measure any change in performance metrics, validating if the changes envisaged are having the desired impact. Providing team confidence in the data and decisions made.