How to do it

1. Build the tree.
   • Use a tool like Optimal Workshop or a prototyping tool to create a site hierarchy, including all the categories and subcategories you want to test.
   • Provide options for each subcategory to prompt realistic behavior from participants who often evaluate link labels by comparing them with alternatives.
2. Define the tasks.
   • Ask participants to find or do something within the site, focusing on key website goals and user tasks, as well as potential problem areas. Set at least one correct destination for every task.
   • Write tasks that reflect how people might naturally approach the site. Use plain language and avoid using tree labels in the tasks.
   • Limit the activity to 10 tasks to avoid low completion rates or participants getting too familiar with the tree.
3. Run the test.
   • Send participants the link to the test or watch them do it in person.
   • If possible, ask follow up questions to get more context that can otherwise be hard to spot in the quantitative data.
4. Analyze the results.
   • Optimal Workshop includes useful data for further interpretation, such as how many people completed each task successfully, how directly they found the destination, and how long it took them.