Creating a Usability Test Plan for a Redesign Project

A redesign can look finished on a slide and still fail in customers’ hands. A usability test plan gives the project a shared agreement about what will be tested, who will take part, which tasks matter, and how evidence will guide decisions. It turns broad goals such as “make checkout simpler” into observable behaviour.

The plan should connect the current product, the proposed prototype, and the decisions awaiting sign-off. It records constraints that could otherwise emerge late, including recruitment time, participant accessibility needs, device coverage, and privacy obligations. Product, design, engineering, and marketing can then work from the same assumptions.

For a moderate redesign, iterative evaluation is usually more useful than one large round. Early sessions expose unclear navigation or content; later sessions compare revised flows with the baseline. The plan should preserve those cycles rather than treating the first script as a one-time test.

A practical plan is specific enough to guide a moderator and flexible enough to accommodate a changing prototype. It defines success criteria before results appear, separates participant reactions from observed performance, and explains how the team will act when findings conflict. This clarity supports routine releases and contested design decisions.

Frame the Redesign and Evaluation Goals

Start with the redesign’s user and business goals. Identify the problem, affected audience, intended change, and release decision. A flow may aim to reduce abandoned registrations, improve self-service, or make essential content easier to locate. Choose two or three priority questions instead of testing every screen.

Turn each question into an evaluation criterion. Useful criteria include task completion, critical errors, time on task, confidence, observed recovery, and accessibility barriers. Record a baseline where possible and agree on meaningful improvement, rather than adopting a generic benchmark after seeing the data.

Keep the plan in a free UCD workspace so goals, prototype versions, notes, and decisions remain connected. Version the prototype and plan together; otherwise results may be attributed to a design that no longer exists.

Define Participants and Research Roles

Define participant profiles from behaviour and needs, not job titles alone. A first-time customer, frequent subscriber, and person managing an account for another may have different stakes. Document research user roles with experience, frequency, technology access, and accessibility requirements, then translate them into screener criteria.

Recruit a balanced mix across Sydney, Melbourne, Brisbane, Perth, and Adelaide, including regional participants where the service is used. Set quotas for customer status, digital confidence, age range, and assistive-technology use. Avoid recruiting only design-friendly participants who recognise the terminology under review.

Offer transparent compensation, such as Coles, Myer, or Woolworths gift cards, or a bank transfer where appropriate. Obtain informed consent under the Australian Privacy Principles, explaining recording, storage, withdrawal, and de-identification. Keep contact details separate from research files and follow a defined retention schedule.

Build Tasks, Scenarios, and Test Environments

Write tasks as goals with realistic context, not instructions that reveal the interface. “Find an eligible claim and submit supporting details” is stronger than “select Claims, then choose the first option.” Prepare success conditions, permitted assistance, and failure criteria so moderators score behaviour consistently.

Use the redesigned prototype for core tasks and, where helpful, the existing product for comparison. An Android application may serve as a familiar comparison artefact for a feed-based redesign, although participants should still test the team’s own prototype rather than a third-party interface.

Match the environment to real use. Test relevant browsers, screen sizes, and operating systems, with remote and on-site options. Australian participants may join from different cities during daylight-saving changes, so schedule across AEST/AEDT and ACST/AWST or provide recordings for people in other zones.

Practical Recommendations for a Focused Test Plan

Prepare a moderator guide, participant information sheet, consent form, task sheet, prototype links, and observation coding sheet. Pilot the flow with a colleague who did not design it; ambiguous prompts and broken states are easier to repair before participants arrive. The protocol should remain a living instrument as the prototype changes.

• Link every task to a research question and a decision, then remove screens that support neither.
• Record task outcome, assistance, errors, elapsed time, and confidence consistently for each participant.
• Test the full journey on one realistic device before expanding the device matrix.
• Label prototype builds, scripts, recordings, and findings so reviewers can trace the evidence.

Set the sample size according to the number of critical journeys and design maturity; formative studies can be compact, while comparative evaluations need stronger controls. Establish what happens if a participant cannot complete a task, the prototype fails, or an accessibility barrier appears. These contingencies protect data quality without pressuring moderators to improvise.

Run Sessions and Capture Reliable Evidence

Run a brief briefing, obtain consent, and explain that help can be requested without penalty. Use think-aloud prompts sparingly and neutral follow-ups such as “What are you expecting to happen?” Avoid defending the design or teaching participants how to complete a task, because assistance changes the performance being measured.

Assign one moderator, one observer, and a note-taker or live recorder. With permission, capture screen activity and audio, log deviations from the script, and tag assistance immediately. A short post-task interview can clarify intent without rewriting what participants did. Review the evidence promptly for missing data, moderator drift, technical faults, and patterns based on one participant.

Synthesize Results and Keep the Plan Actionable

Synthesize quantitative measures and qualitative evidence together. A high completion rate can conceal confusion or accessibility barriers, while one dramatic comment may reflect preference rather than a usability problem. Group findings by journey and severity, then connect each issue to its evidence and affected user role. Write recommendations that state the problem, likely impact, supporting evidence, and required decision.

Separate essential corrections from optional refinements, and identify which prototype changes need another session. Retest changed flows with relevant participants using comparable tasks and measures. Update the plan after each cycle with decisions, rejected changes, and remaining uncertainty, and store the report with project artefacts. Project owners can contact UCDmanager for help organising research roles, usability activities, or accessibility evaluations. A traceable record keeps later redesign decisions explainable to the project team and reviewers.