What Is Usability? Context, Evaluation Methods, and Improvement Decisions
Published:
Last Updated:
Category: Marketing Glossary, Content Marketing
Authors: Shusaku Yosa
Usability concerns how well particular people can perform a task in a particular situation. Specify the users, task, and environment before evaluating a screen. Then choose observable behavior and a success condition. This makes an assessment more useful than an unqualified statement that an interface is easy to use.
Usability depends on the user and task
An abbreviation that confuses a first-time user may be efficient for someone doing the task daily. One-handed phone use also differs from reviewing several documents at a desk. These contexts affect the information and interactions people need.
Nielsen’s usability explanation identifies learnability, efficiency, memorability, errors, and satisfaction as useful perspectives. Their importance varies by task: initial setup may emphasize understanding, repeated reporting may emphasize efficiency, and a consequential application may emphasize accuracy and recovery.
The official overview of ISO 9241-11 also treats usability as an outcome of use. Its public overview does not prescribe a particular participant count or testing method. Build the evaluation around the question your team needs to answer.
Clarify the relationship with UX, UI, and accessibility
UI is the set of interaction points, including buttons, displays, and inputs. Usability is a way of assessing work performed through those points. UX covers a wider experience, including expectations before use and experiences afterward. Comfortable operation cannot compensate for an offering that does not address the user’s need.
Accessibility addresses use by people including those with disabilities. W3C discusses its overlap with usability. Consider relevant assistive technology and actual conditions rather than relying on separate checklists alone. Successful task observations do not, by themselves, establish accessibility conformance.
Choose methods for the question
“Where do people hesitate?” and “What proportion can complete this task?” are different questions. Exploratory observation and a quantitative estimate or comparison require different recruitment and sample planning. The available evidence may also limit the question that can responsibly be answered.
Method | Useful question | Limit |
|---|---|---|
User testing | Where do people hesitate, fail, or recover during a task? | A small group cannot establish a population-wide occurrence rate |
Heuristic evaluation | Does the design have recognizable interaction problems? | An evaluator may miss problems specific to real users |
Usage records | At which stage does observable activity stop? | Records alone do not establish intent or cause |
Questionnaires or interviews | How do people describe satisfaction, expectations, and conditions? | Reported opinions may differ from behavior |
A heuristic review can draw on principles such as status visibility and error prevention. Follow up with user observation where appropriate. A low-traffic service does not benefit from adopting a formal comparison experiment that cannot answer its question reliably.
Plan participants, tasks, and success conditions
The following fictional quotation-service plan is exploratory. Explain the scenario, then let the participant choose how to proceed. Telling them which button to press would remove the opportunity to discover whether that button is understandable and findable.
Item | Fictional evaluation plan |
|---|---|
Participants | Business users requesting a quotation for the first time |
Environment | Their usual device and input method; note connection conditions |
Task | Compare two staffing quotations and save one |
Success | Correct assumptions, correct destination, and confirmation of the saved result |
Assistance | Record requested help separately from independent completion |
Observations | Hesitation, errors, recovery, unfinished reasons, and time |
Evidence handling | Collect necessary observations and explain recording and permissions in advance |
Suppose five fictional participants produce three independent completions, one assisted completion, and one unfinished attempt. Independent completion is 3/5 = 60%. This describes this observation; it is not a population estimate, proof of superiority, or a universal rule that five participants are enough. Different user groups and purposes require different plans.
State whether timing includes assisted attempts and whether it covers only completions. Excluding unfinished attempts can make difficult cases disappear. After the task, ask about observed hesitation and separate the visible fact from the team’s proposed explanation.
Separate impact, evidence, and implementation effort
Record the severity of a problem, observed occurrences, confidence in the evidence, and effort separately. Multiplying all these values in the same direction can make higher implementation cost appear to mean higher priority. Discuss the intended decision rather than treating an unexplained score as authoritative.
Fictional problem | Impact and evidence | Decision |
|---|---|---|
A changed condition is unnoticed | An incorrect quotation is saved; observed and reproduced | Test a correction early because accuracy is affected |
Supporting text is difficult to find | Some hesitation, but another route succeeds | Try a small placement change and observe again |
Color preferences differ | Opinions, without a demonstrated task barrier | Investigate operational impact before prioritizing |
A problem not observed in a small group is not necessarily absent. Consider important user groups and assistive-technology conditions that were not represented. Conversely, an individual preference is not sufficient evidence for a wholesale redesign.
Recheck the task and measure business outcomes separately
After a change, repeat the task with the same success conditions and look for both the original barrier and new errors. Record learning effects, participant differences, and environmental changes. If pricing or recruitment changed at the same time, do not assign the entire business result to the interface.
A clearer task outcome is valuable evidence. Revenue, retention, and support workload still require their own measurement. Keep the change, supporting evidence, remaining question, and next review together. Returning to the user’s actual work makes evaluation an ongoing decision process.
For the next step, use task-specific UI improvement examples; customer onboarding process design; app identifiers and the limits of observation data.
