Value Metric

Time on Task

Time on Task is how long it takes users to complete a specific task — usually, faster is better.

Type
Usability
Funnel
Activation

What is Time on Task?

Time on Task measures the elapsed time from when a user starts a task to when they complete it. For most functional tasks, shorter times signal a more efficient, learnable experience, which is why it is a staple of usability testing and a sensitive detector of friction.

Because task times are typically skewed (a few slow users pull the mean up), the median or geometric mean is usually a better summary than the arithmetic average.

How to calculate it

Average Time on Task = Total Time Across Users ÷ Number of Users (prefer median for skew)

Total Time
Sum of completion times for the task
Number of Users
Users who completed the task

Worked example

If five users take 40, 45, 50, 60, and 120 seconds, the median time on task is 50 seconds — a better summary than the 63-second mean.

What good looks like

  • Relative to a baselineFaster is usually better

    There is no universal target; compare against a prior design, a competitor, or an expert benchmark. Only compare times among users who succeeded.

Why it matters

Time on Task is a precise, sensitive measure of efficiency that often reveals friction before success rates drop — users may still finish, but slowly and painfully. It is ideal for comparing designs and tracking improvement. The nuance: only compare successful attempts, and remember that for exploratory or content experiences, longer is not necessarily worse.

How to improve Time on Task

Streamline the task flow

Remove steps, clarify labels, and improve defaults so users reach the goal with less effort, then re-measure.

Speed the path to the first correct action

Guide attention so users take the right first step quickly — often the biggest chunk of total task time.

Frequently asked questions

Should you use the mean or median for time on task?

Prefer the median or geometric mean. Task-time data is usually right-skewed — a few very slow users inflate the arithmetic mean — so the median gives a more representative picture of the typical experience.

Is faster always better for time on task?

For most functional tasks, yes — faster reflects efficiency and learnability. But for exploratory, educational, or content experiences, longer engagement can be desirable. Always interpret time on task against the goal of the task, and only compare successful attempts.