Choose the question before the testing method.

A navigation test asks whether people can find a destination. A usability session asks whether they can complete a task using the interface and its content. An accessibility evaluation examines conformance and barriers through a defined evaluation method. These activities inform one another but do not substitute for one another.

For a redesign, start with a consequential uncertainty: can a resident find the correct collection schedule, determine whether a permit applies or locate the agenda for a specific meeting? Select the page, document and form transitions that the task depends on. A test of the homepage alone misses much of the service.

Use GOV.UK’s moderated testing guidance (opens in a new tab) as a methodological reference. It describes observing tasks and asking participants to explain their thinking. The worksheet and example tasks here are Civicfield planning aids, not a prescribed municipal research standard.

Recruit for the task and support participation.

Include people who use or are likely to use the service, with relevant differences in language, device, familiarity and access needs. Department staff are useful for checking facts, but knowing the organization can make a difficult navigation route look obvious. Keep their internal review separate from resident observations.

Agree on accessible session arrangements and explain what will happen, what information is collected, who can see it and how participation can stop. Obtain informed consent before recording. Use agency-approved privacy and retention arrangements. GOV.UK’s consent guidance (opens in a new tab) is a helpful process reference, not a replacement for local review.

Use fictional information when the task does not need personal records. If testing involves a live submission, payment or official action, establish a safe test environment first. Let assistive-technology users use familiar setups where feasible and confirm that the prototype supports the intended test.

Write neutral tasks with a clear finish.

ScenarioObservable finishLeading instruction to avoid
You plan to attend the next parks board meeting. Find when and where it happens and what will be discussed.Correct meeting occurrence, location and agenda identified.Click Government, then Boards.
Your collection day falls near a holiday. Find what you should do this week.Current schedule change and applicable address conditions understood.Use the Holiday Collection page.
A street-name sign is unreadable. Find how to report it and what information is needed.Correct approved reporting route and required details identified.Press the Report a Problem button.

On a small screen, scroll the table sideways to see every column.

Read the scenario consistently. Ask participants to explain what they expect a link to do. If someone becomes stuck, note the point before offering assistance, and record what assistance changed. A rescued task should not become an unassisted success in the report.

Pilot the session with a colleague to catch broken links and unclear instructions. That rehearsal tests the research setup; it does not replace research with likely users.

A user researcher prepares paper task prompts beside a phone, notebook and a plain local street map on a real table.
Preparing neutral research tasks.

Use an observation worksheet that separates fact from interpretation.

FieldWhat to record
Session and taskAnonymous session code, task version, device and relevant setup.
Observed actionActual path, selected labels and hesitation point.
OutcomeIndependent completion, assisted completion, incorrect answer or not completed.
EvidenceShort consented note or timestamp; exclude unnecessary personal details.
InterpretationA hypothesis about why the barrier occurred, kept separate from observation.
Impact and next actionBlocked task or lesser friction, proposed fix and accountable owner.
RetestChanged version, task wording and result in the next round.

On a small screen, scroll the table sideways to see every column.

For example, “Opened Minutes when looking for tomorrow’s agenda” is an observation. “Does not understand meetings” is an unsupported judgment. A better hypothesis is that labels or date context did not distinguish upcoming materials from approved records. Test that hypothesis after revising the interface.

Two researchers review anonymized observation notes beside a revised website prototype on a laptop.
Recording observations and planning a retest.

Report the limits and prioritize the barrier.

Summarize task outcomes with the number of sessions and the conditions tested. A small qualitative study can reveal important failures; its completion percentage is not a population estimate. Do not claim that testing a fixed number of people proves the whole site usable.

Prioritize an incorrect service instruction or a blocked submission before a cosmetic preference. Look for repeated patterns, but investigate a severe barrier even if one participant encountered it. Preserve differences across devices and access needs instead of averaging away a failed journey.

User participation also does not establish WCAG conformance. W3C guidance on involving users in accessibility evaluation (opens in a new tab) recommends combining user involvement with standards evaluation. Document what your study did and did not cover.

Retest the changed journey and retain the decision.

Assign each accepted issue a change owner, target version and retest task. Keep the task’s goal comparable so a label change can be evaluated without quietly making the new task easier. Where returning participants have learned the route, include suitable new participants and note the difference.

Store a concise decision record: evidence, revision, remaining uncertainty and next check. Use findings to improve the service-page pattern and to define a representative acceptance slice in the redesign process.

The aim is a repeatable learning cycle with visible decisions. A polished report that never changes a page provides little help to a resident trying to complete the task.