01 / UNDERSTAND
What is time on task?
Time on task measures elapsed time for a defined attempt under stated start and stop rules. It can help assess efficiency, but only in relation to the goal and outcome. A faster failed attempt is not an efficiency improvement.
Task times often have a long upper tail. Keep individual durations and use a summary suited to the analysis, such as a median or a justified geometric mean. Separate successful attempts from failures and do not silently omit the latter.
Why use Time on Task?
A flow can work but take avoidable effort because people repeatedly search, reread or wait for feedback. Timing helps locate that cost when paired with observation. It also protects against a misleading comparison in which a design appears faster simply because participants give up sooner or receive more help.
When to use it
USE IT WHEN
The task is understood and its duration has a meaningful consequence.
Comparable designs or versions need an efficiency comparison.
Observation suggests repeated work or uncertainty within a successful flow.
When to choose another method
Use Task Success Rate when the primary concern is completing the goal.
Use Single Ease Question (SEQ) for perceived task difficulty.
Use User Interviews to understand whether speed is actually valuable in this situation.
What you’ll need
Define when timing begins and ends, what counts as an interruption and whether system waiting is included. Keep protocol, task wording and help consistent. Think aloud can alter duration, so document it rather than comparing it casually with silent use. Decide how timeouts, failures and pauses will be reported before collecting data.
What defines a comparable task duration
Keep the elements that help answer your question and support a decision.
The goal and attempt whose duration is measured.
The observable event that begins timing.
Completion, abandonment or a stated timeout.
Help, think aloud, equipment and test conditions.
Individual durations, spread and declared summary.
Completion, mistakes, waiting and confidence.
02 / PUT IT INTO PRACTICE
How to measure time on task
Keep evidence and assumptions distinct. Revisit your work when new information changes your understanding.
01
Choose the goal
Measure a complete useful task rather than a convenient screen fragment.
Include confirming the replacement visit, not only opening the picker.
02
Set timing boundaries
Use observable events and apply them consistently.
Start after the scenario is understood; stop at verified completion.
03
Declare pause rules
Explain whether interruptions and system waits remain in elapsed time.
Keep a slow service response visible if it occurs in normal use.
04
Hold protocol stable
Record assistance, narration and device conditions.
A think aloud session should not be treated as identical to silent use.
05
Collect outcomes too
Store duration together with success and failure evidence.
Record a quick abandonment as a failed attempt, not a fast completion.
06
Summarise the distribution
Report counts and spread with an appropriate centre, retaining the individual data.
Five successful times can be ordered to inspect the median.
07
Interpret the cost
Locate where time was spent and whether removing it would preserve understanding.
Distinguish necessary review from repeated searching.
Example: timing a confirmed appointment change
Fictional scenario: five successful attempts take 40, 50, 60, 80 and 150 seconds under identical rules. The median is 60 seconds and the arithmetic mean is 76 seconds. A sixth attempt ends unsuccessfully at 25 seconds and is reported separately.
The illustrated strip places each successful duration on an axis and marks the median, with the failed attempt in its own row. Observation notes identify repeated checking during the 150 second attempt. All times are invented.
The design question is why checking was necessary. If clearer status reduces repeated searching while preserving a deliberate confirmation, that could improve efficiency. Removing the confirmation merely to lower the time would need a different judgement about risk.
03 / BELOW THE SURFACE
Below the surface: useful friction can look inefficient
The shortest path can remove the moment when a person notices a consequential mistake.
BOUNDARY
Are start and stop events equivalent?
Keep timing rules reproducible.
WORK
What was happening during the slow segment?
Separate waiting, searching and deliberate review.
TRADEOFF
Could speed increase errors or misunderstanding?
Check outcome quality alongside duration.
An appointment change might require reviewing the new date and the old booking’s status. A rushed flow could finish faster but increase incorrect changes. Time cannot distinguish unnecessary friction from useful attention without examining what happens during it.
Before optimising a slow segment, label the work it performs. Is it waiting, searching, understanding, choosing or verifying? Remove repeated work before removing considered choice. Then pair timing with accuracy, comprehension and recovery observations.
This editorial decomposition does not introduce a new timing formula. It helps explain what a shorter duration would mean. In a high consequence task, a small increase in deliberate review may be acceptable if it reduces errors; the team should make that tradeoff explicit instead of hiding it behind average speed.
Time on Task companion worksheet
The original two page companion records timing rules, protocol, durations and outcomes. Page two examines where time is spent and which tradeoff a change creates. It is a measurement planning worksheet, not an automated stopwatch or statistical estimator.
Practical tips
Keep unsuccessful attempts visible in a separate outcome group.
Show spread, not only an average.
Review the protocol before comparing durations.
Common mistakes
Pooling completed and abandoned tasks into one speed score.
Starting the timer at different points across sessions.
Optimising duration without checking errors and understanding.
What you’ll get
A reproducible duration measure with outcome context. Use it to remove a specific avoidable cost, then review whether the change preserves successful and understood completion.