The measure of how long a user takes to complete a specific task. Shorter times generally indicate better usability, though context matters — a fast checkout is ideal, but a fast content discovery session might signal that users gave up rather than succeeded.
Common contexts
- Comparing average time-on-task between the old and redesigned onboarding flow across two cohorts
- Setting an efficiency benchmark for a critical enterprise workflow where slow task time costs employees billable hours
- Flagging suspiciously fast task completions in session data as potential rage-clicks or skipped steps
Use when
Use time on task when efficiency is a defined product goal — particularly in professional tools, high-frequency workflows, or checkout flows where every extra second has a measurable cost in conversion or user frustration. It's most meaningful when you have a baseline to compare against.
Avoid when
Don't optimize for time on task in exploratory or discovery contexts — a travel planning tool or a learning platform where longer engagement signals deeper satisfaction will look worse by this metric even as the experience improves. Match the metric to the intended user behavior.
Time on task is the metric most likely to mislead when you haven't controlled for participant expertise — an expert completing your task in 40 seconds and a novice completing it in 4 minutes may both indicate equally valid design problems.
Real-world examples
- GOV.UK reduced the average time to complete a driving licence renewal from 21 minutes (on the legacy DVLA site) to 4 minutes after their 2013 redesign — a time-on-task improvement they cite as the primary evidence of design success.
- Nielsen Norman Group benchmarks time on task across competitor sites for clients; a 30-second difference in insurance-quote completion time translates to a measurable difference in abandonment for the slower competitor.
- Stripe's checkout.js reduced average payment form completion time from 42 seconds to 22 seconds by auto-detecting card type, formatting numbers automatically, and pre-filling postal codes — each micro-optimisation measurable in time on task.