metricData, experimentation & growth

Time on Task

Time on task measures the elapsed time users need for a defined task, providing an indicator of efficiency in a specified context.

Download Markdown (.md)

01What it means

Time on task is a usability and performance measure of elapsed duration between a defined task start and endpoint. It is commonly interpreted as one aspect of efficiency, but the meaning of faster or slower performance depends on the task, outcome, and user context.[1][2][3]

02Why it matters

Two designs can have similar completion rates while demanding very different amounts of user time. Task-time data can reveal friction, support benchmarking, and quantify efficiency improvements, particularly for frequent workflows where small delays accumulate across many repetitions.

03When to use it

Use time on task when start and end events can be defined consistently and efficiency matters to users or operations. It is useful in comparative usability tests and production telemetry, but should usually be analyzed only for comparable tasks and separated by success status.

04How it works

  1. Define the task, user group, start event, end event, assistance rules, and treatment of abandonment before measuring.
  2. Capture elapsed time consistently and verify that instrumentation or observation records the intended boundaries.
  3. Separate successful, failed, abandoned, and assisted attempts because their durations have different meanings.
  4. Inspect the distribution and report a robust summary such as the median when times are skewed, alongside sample size and variability.
  5. Interpret changes with task success, errors, satisfaction, and qualitative evidence; investigate outliers rather than silently deleting them.

05A practical example

Example

A warehouse team compares two picking interfaces for the same set of orders. The new design reduces median successful task time, but error rates rise for unfamiliar item codes. The team does not declare victory from speed alone; it revises code presentation and retests both efficiency and accuracy.

06Where it breaks

Faster is not always better: thoughtful reading, learning, safety checks, or creative work may appropriately take longer. Timing data are often skewed and sensitive to interruptions, device performance, expertise, and task scope. Excluding failures can make a difficult design look deceptively efficient.

Common mistakes

  • Comparing task times for different scopes, users, devices, or success criteria.
  • Averaging skewed durations without inspecting the distribution.
  • Optimizing speed while ignoring errors, comprehension, safety, or satisfaction.

07Sources

  1. A Practical Guide to the Common Industry Format: Usability MeasurementsNational Institute of Standards and Technology · Government measurement guideA Practical Guide to the Common Industry Format: Usability Measurements
  2. ISO 9241-11:2018 Usability: Definitions and conceptsInternational Organization for Standardization · International standardISO 9241-11:2018 Usability: Definitions and concepts
  3. Usability benchmarking a website or whole serviceGovernment Digital Service · Government service manualUsability benchmarking a website or whole service
View saved topics

Start typing to search titles, definitions, categories, and frameworks.