Work from the decision outward. Collect what you already know, locate the unresolved question and test only when the next decision needs more evidence. You can finish an assessment without commissioning a new participant study.
1. Define the decision and what can change
Write the decision in one sentence, for example "keep or replace the top-level menu before the spring release". Then list what is fixed (platform limits, legal names, deadlines), what can change, and who decides. That sets the scope, while evidence may still be needed to distinguish an IA problem from another cause.
Questions to answer before going further:
- Which audiences matter most?
- What would make the decision obvious?
Name the decision-maker and the changes they can act on before you commission further research. That makes every result easier to turn into a concrete next step.
2. Gather what you already have
Start with the evidence your organization has already collected: top tasks, search terms, support questions, analytics, previous studies and the current structure. The table shows who usually owns each source and what it can contribute, so you can assemble a useful assessment without duplicating research.
Start with the pages and sections implicated by the decision. Expand to a fuller content inventory when coverage, duplication, ownership or missing navigation paths require it. 5
Reuse earlier tree-test tasks when they still describe the job well. Improve the ones that were leading, compound or vague. Tasks you write now become the baseline for the next round.
Participants, if you will need them
- Who: participants from each audience the decision affects. Results from one audience do not generalize to another.
- Where from: recruit participants who represent the target audience for the decision. Colleagues can help pilot the tasks; include their results in the study only when they genuinely fit that audience and familiarity does not bias the task.
- How many depends on the purpose. A qualitative pilot needs a few participants 1. For quantitative tree testing, about 30 participants seeing each task can start to show patterns within an audience, and about 50 seeing each task is a common target 11. If each participant sees only part of the task bank, recruit proportionately more participants so each task still gets that coverage. Comparing two structures typically needs about 50 or more observations per task, per structure for reasonably narrow confidence intervals 2. Each step up adds recruitment cost and time.
Review the structure itself
Trace the routes people would use for priority tasks and look for structural features that deserve investigation: ambiguous siblings, content without a browse route, unexpected depth and items with several plausible homes. The same pattern can be useful in one context and troublesome in another, so use it to form questions rather than assign automatic penalties. 12
To inspect one navigation against your priority tasks and record what you find, use the navigation audit worksheet.
Rules that mislead when applied everywhere
- Three clicks. What matters is whether each choice is predictable. A clear five-step path can work better than an ambiguous two-step one.
- Seven plus or minus two per level. This comes from working-memory research, not studies of menu scanning.
- No more than N levels. There is no universal correct depth.
- A structure health score. Tree depth and shape do not establish findability.
3. Read the evidence and identify the likely part
Bring complaints, search behavior, prior studies and structural inspection into one view. Sort the findings by coverage, grouping, labels, and hierarchy and depth. Where sources disagree, look at which tasks and audiences each one represents. That often reveals the next specific question to test.
Contradictions to look for
- Search logs show frequent queries for content already in navigation. The label may not match search vocabulary, the route may be hard to follow, or search may simply be preferred.
- Support reports findability failures while analytics shows high usage. The support audience may differ or arrive by a different route.
- An earlier study found no problem, but complaints continue. Check whether it covered the complaint themes with the same audience.
- The team agrees on placement, but search vocabulary differs from the label. The grouping may be sound while the label is not.
Compare query terms used for the same need and audience. Search logs often use specific-item names rather than category labels: they can reveal vocabulary gaps, but query frequency by itself does not measure browse findability.
What counts as a complaint. Classify a support request as a findability complaint only when it describes a failed attempt to locate content or complete a task.
Coverage gaps in existing evidence. Analytics, search logs and support records cover only observed visits, searches and submitted contacts. If a target audience is absent from those sources, you still lack findability evidence for that audience.
If the existing evidence resolves the decision for the affected audience, record the basis and move to Step 5. For critical or hard-to-reverse changes, confirm that the evidence is strong enough before ending the assessment.
4. Test only what remains open
The table connects each unresolved question to a suitable study, typical sample sizes and the evidence it can produce.
For a live hierarchy, start with the evidence you already have and add a baseline tree test when you need to evaluate its routes 1817. If the open question is how people would group the content, a card sort can explore it, and plausible groupings can become candidate structures. Alternatives are compared by tree-testing them on equivalent tasks, and a problem that appears only with the interface calls for testing the screen itself.
Tasks
Take tasks from the top-task list, search logs and support themes. Give each participant 8 to 10 tasks; more can create learning effects 15. Write tasks in user language and avoid tree-label wording; otherwise the task becomes a word-matching exercise 9.
Look for semantic cues as well as exact words. "Your green bin wasn't emptied" can cue the Environment section without sharing its label. When a frequent query matches the label, write the task around the underlying need ("you want to know what a mortgage would cost you each month"), not the query.
More on task design: writing tree test tasks and avoiding leading tasks.
Comparing a current and a proposed structure
Test the current structure as a baseline and the proposed structure with comparable participant groups, the same scoring and the same test format. Carry tasks forward where they still describe the job, and improve the ones that were leading or vague for both structures. See Reading results for how to read the comparison.
What drives effort: the number of audiences, whether you compare structures, task allocation per participant, the size of the tree you test (not the size of the site), and where participants come from.
When the remaining question concerns a different part of the experience, name the evidence you need and send the issue to the right method or owner.
Test a proposed navigation: see which routes participants choose and where they finish.
Start a tree test
Explore content grouping: learn which labels and organizing patterns participants use.
Start a card sort
5. Read the results and decide
Read tree-test results as routes, not just scores. Direct finishes reached the intended destination without backtracking; indirect finishes reached it after backtracking; elsewhere finishes ended at a different destination. Combine those outcomes with first clicks and final destinations to see where the navigation needs attention. 2
When a task's first clicks form a recognizable pattern, UXbeam names it on the task: most participants making the same wrong first choice, two competing first choices, a correct first choice followed by a lost path, or scattered first choices. UXbeam also flags a destination where many attempts ended without being accepted, so the team can review whether it should count.
Patterns to notice
Possible changes, once an inspection supports them: move an item, list it in a second place, rename a section, make two neighbouring labels distinct, fix the content, or hand the problem to the interface or search owner. List an item in a second place only when a substantial share of participants choose that location; doing it everywhere dilutes the structure 2.
How to interpret comparisons and retests
- Keep scoring, test format and participant audiences consistent between rounds, and report what each revised round achieved, tasks included. A 2025 peer-reviewed study found that different tree-test formats produced significantly different results 19.
- With modest samples, moderate differences in success, first clicks or final destinations can be uncertain. Assess the uncertainty of the difference between two results, rather than comparing either result against its own margin of error.
- If a lower score on a critical task would change the decision, plan the study to detect a meaningful difference. Around 50 observations per task and structure is a common starting point for comparisons 2, but a moderate task-level difference may require a larger sample.
- After each iteration, examine first clicks, paths and destinations to understand what improved and where further changes would help.
- If a destination originally marked wrong would genuinely satisfy the task, review the accepted-answer key on content grounds. Show both the original and revised scoring. Use the original scoring for a like-for-like comparison and present rescoring as a separate analysis.
- Set decision thresholds by task importance, audience and the consequences of failure. No single pass mark suits every structure.
When the evidence points elsewhere
- Content: required content is missing, inaccurate, or duplicated. It goes to the content owner.
- Interface and navigation design: the item is findable in the tree but not on the live site. Test the interface 3.
- Search: relevant content exists, but site search returns poor results. Route it to the search owner 6.
- Grouping: first clicks scatter across many tasks, or placement remains disputed. Run a card sort (card sorting, reading card sort results, organizing principles).
Prioritize findings by task importance and frequency. If you assign severity ratings, use at least three independent raters 13. For content, record whether the next action is to keep, update or remove it 5.