Mara logo

Tree Testing

Testing whether people can find things in a proposed structure, using a bare text hierarchy with no design, navigation or search.

Process & Methods

Tree testing strips everything except the hierarchy. Participants see nested text labels and are asked to find something — “where would you go to change your billing address?” — clicking down until they commit. Because there is no visual design, no search and no cross-linking, the result isolates the structure itself. A structure that fails a tree test will also fail in the product; the design was just hiding it.

Three numbers come out: success rate (did they end up in the right place), directness (did they get there without backtracking), and first click (where they went first, which is the best predictor of eventual success). Low directness with high success means the labels are ambiguous and people are recovering; low first-click accuracy means the top level is wrong, which is the expensive kind of problem.

In practice

A SaaS product tested its settings hierarchy with 40 users on eight tasks. “Change who receives invoices” had a 34% success rate — most people went to Account rather than Billing, because Account held the email address. Moving invoice recipients under Account raised it to 89% without changing a single label.

Where teams get it wrong

  • Wording tasks with the same words as the labels, which tests reading rather than structure.
  • Testing a structure with fewer than about 30 participants, where percentages are noise.
  • Ignoring first-click data, which predicts success better than the final answer.
  • Concluding that a failing structure needs better design. The design was never in the test.
  • Testing only the happy paths the team already believes work.

Learn more

You may ask

Frequently Asked Questions

What is tree testing used for?

Validating whether a navigation structure lets people find things, independently of visual design. It is the cheapest way to catch an information architecture problem before it is built.

How many participants do you need for a tree test?

Around 30–50 for reliable success rates. It is a quantitative method, so small samples give you anecdotes rather than percentages you can act on.

Related terms

All terms

Defined by Mara Last reviewed .

Let's talk about your product.

Happy to look at what you're building and say where design would move the needle.

Contact Us