Task Success Rate & Time-on-Task

Task success rate is the share of valid task attempts users complete successfully according to criteria defined in advance. Time-on-task is the time between a defined start event and end event. Together, these UX metrics show whether a task succeeds and how long it takes—provided the task, success criteria and timing boundaries are clear.

How Successful are Users? - The Task Success Rate

The so-called task resolution rate indicates how many people were able to complete a certain task completely and correctly. It is given as a percentage and is particularly suitable for measuring the effectiveness of a system.

Formula: Task success rate = (valid attempts meeting the complete-success criteria ÷ all valid attempts) × 100. An attempt counts as successful only if the predefined user goal is achieved without a critical error. Also define whether an abandonment is a valid failed attempt; technical failures unrelated to the product may be excluded under a rule set in advance.

AI

AI-generated image

Variants of the Survey

• Binary: Success yes/no - easy to survey, but not very differentiated.
• Report completion levels: Count complete success, partial completion and abandonment separately, alongside the binary success rate. This keeps the standard metric clear while making progress visible.
• Weighted scoring: If subtasks have different importance, a predefined score can be calculated: sum of weights for completed subtasks ÷ sum of all weights × 100. This is a supplementary weighted completion score, not the same as the binary task success rate; document its weights and criteria.
• Error-based: Additionally analyze why tasks fail - e.g. technical errors, misunderstandings, usability problems.

The success rate can describe and compare versions or time points when tasks, success criteria, and measurement conditions are comparable. An observed difference alone does not establish a cause; causal claims require an appropriate study design, such as a randomized experiment. With small samples, consider estimation uncertainty and chance variation. Trends over time can also reflect changes outside the product.

How Efficient is the Use? - Time-on-task

While the success rate describes the achievement of a goal, the Time-on-Task records the effort required to achieve it - usually in the form of the time needed.

Practical example: In an e-commerce test, a successful checkout takes a median of 3:20 minutes on the old version and 1:50 on the redesigned version. The difference is 90 seconds, or 45% less time relative to 200 seconds. This comparison is meaningful only if the task, sample and measurement rules are comparable and the success rate has not fallen.

But: Less time is not automatically better. An extremely short time-on-task can also indicate superficial processing or lack of interest. Therefore, this metric should always be interpreted in the context of the task and user goals.

How is it Measured?

• Formula: Time-on-task = timestamp of the end event − timestamp of the start event. Define both events in advance, for example from the first “Log in” click to confirmed sign-in.
• For moderated tests: by stopwatch or digital timer
• For digital systems: automatically via log files, session recording or analytics
• Important: Define the same start and end events for every participant. Set rules for pauses, moderator interventions, loading time and interruptions in advance.

Example: Eight successful attempts take 12 minutes in total. The arithmetic mean is 12 × 60 ÷ 8 = 90 seconds. With small samples, report individual values or at least the sample size and spread; for right-skewed times, the median and interquartile range are often more informative than the mean alone. Times from successful attempts alone can look deceptively good when many people abandon the task: report abandonments and time to abandonment separately rather than silently omitting them from the success or timing picture.

Interpreting outliers and abandonments

Check unusually short or long times against the recording first: was there a timing error, an external interruption or a genuine interface detour? Keep valid observations in the analysis; they may reveal barriers. Exclude values only under a transparent, predefined rule and document the impact. Avoid arbitrary time limits that turn slow but valid attempts into apparent successes.

Strong Together: Effectiveness & Efficiency

The two metrics only become fully meaningful when combined. A task can be completed quickly - but still fail frequently. Or it is usually completed successfully, but takes a remarkably long time.

Example combination: In an intranet test, 90% of users manage to find a specific piece of information - but it takes them over 4 minutes on average. The interpretation: The system works, but not efficiently. A better navigation or search function could help here.

Best Practices from the Field

• Formulate tasks realistically: The task should correspond to the real context of use.
• Operationalize start and end points: This is the only way to consistently record times and success rates.
• Do not simply delete outliers: A user needs 10 minutes because the interface is blocked in a certain constellation - this is not a “measurement error”, but a valuable insight.
• Triangulating with qualitative methods: Statements from interviews or thinking aloud help to interpret figures (“Why was the task difficult?”).
• Report time and success together: A faster time is not an improvement if more people fail or make critical errors.

How to measure both metrics consistently

1. Define the task, target users and the expected successful outcome.
2. Set the exact start and end event before collecting data.
3. Record completion as complete, partial or failed and capture the time for every participant.
4. Report the sample size, distribution and relevant outliers instead of only an average.
5. Compare versions only when task wording, context and participant criteria remain comparable.

Conclusion

Task Success Rate and Time-on-Task are proven tools for making user performance measurable. As quantitative metrics, they provide clear, comparable and meaningful data about the usability of a system - provided they are used context-sensitively and methodically clean. In practice, they are indispensable for benchmarking, redesign evaluations and continuous UX optimization.

Last modified: 1 November 2026