Where would customers expect website maintenance: “Solutions”, “Services” or “Ongoing support”? A tree test can investigate that choice before the finished layout is built. You need the category structure and tasks that do not already supply one of those categories as the answer.

What tree testing isolates

Participants receive a hierarchy of categories and a task. They choose the place where they expect to find the relevant information. NN/g describes tree testing as evaluating findability in a structure without the finished website’s visual design.

That can investigate “Which entry point do people use to find maintenance?” Whether a menu is easy to open on a phone requires testing the actual interface. A sound structure can still be difficult to reach through an awkward menu.

Prepare a sufficiently complete structure

To interpret the choice, your prepared structure needs to reach the level where the task can be resolved. With only three top-level terms, selecting “Services” leaves it unclear whether the person expects maintenance, support or consulting underneath. The additional level makes that distinction visible.

Suppose you group agency services under “Design”, “Development” and “Ongoing support” and want to know where customers would look to repair a broken contact form. The task can explain that messages are no longer forwarded and the existing site needs to transmit enquiries again. It does not name the intended destination, “Form maintenance”, because you need participants to make that choice themselves.

Before testing, the team defines which destinations make sense. More than one route may be valid. A reasonable alternative should not count as a failure merely because the current plan places it elsewhere.

Examine the route as well as the destination

The route to a correct destination can still reveal where the structure causes difficulty. Recording the starting point, backward steps and final destination exposes detours that the endpoint alone would hide. In a moderated study, you can then ask what the person expected a label to mean without explaining it before the choice.

Distinguish three possible findings:

  • A label means something different to participants.
  • The requested service plausibly fits several categories.
  • The tested structure lacks the service or information.

That distinction determines what you can change. A misunderstood label calls for reviewing its wording; an offer that fits several categories needs an appropriate connection between them. A missing destination must first be added so a new test can meaningfully investigate the route to it.

Check the revised structure in the real menu

Another structural study can investigate whether revised categories make more sense. Then test the actual menu with content, search and mobile layouts. Record the version and participant group. Small qualitative rounds explain search paths; they do not establish a reliable completion percentage for every visitor.

If terminology and the service structure remain open, user interviews can explore the language people use. Usability testing can investigate selection within an agreed scope. For planning, prepare the current structure, disputed labels and most important search tasks.