Begin with the visitor’s goal
Describe what a person is trying to achieve in their own terms. A task prompt should give a reason to look for information without naming the category being tested. Otherwise, the study may measure the ability to repeat a word rather than find an appropriate route.
Identify the audience’s starting knowledge. An experienced staff member and a first-time user may approach the same task with different expectations, and both can be relevant to the product decision.
Separate the structure from the presentation
A navigation model can be examined with limited visual detail when the question concerns categories and labels. A more complete prototype may be necessary when layout, visibility, or interaction affects the route. Choose the representation that fits the uncertainty.
State which parts are simulated. A paper-like structure and a working interface can produce useful but different observations. A successful route through one should not silently be treated as a complete test of the other.
Observe the choices between endpoints
The final destination is only part of the evidence. A person may hesitate, explore several branches, or recover after an unexpected choice. These behaviors can reveal whether the organization supports the way they understand the task.
Ask for interpretation at appropriate moments without teaching the route. Record the distinction between what the person expected to find and what the chosen category actually contained. That gap can point to a naming or grouping problem.
Recommend a structural response
A useful result explains the route, the point of uncertainty, and the implications for the navigation model. The answer may be a clearer name, a different grouping, a cross-link, or a more explicit task description.
The report should identify the tasks and audience examined. Broad navigation changes can affect other routes, so a revision should be reviewed with related tasks before it is assumed to improve the whole experience.