The percentage of participants who successfully finish a defined task during a usability study, or of all users who attempt it in a live product. It is one of the most direct indicators of whether a design actually works for its intended purpose — a low rate rarely lies.
Common contexts
- Reporting usability test results for a form redesign where the previous version had a 54% completion rate
- Setting a success threshold before a study — e.g. 80% completion required before shipping
- Tracking live product task completion via session recordings to validate post-launch assumptions
Use when
Use task completion rate as your primary success metric whenever you need a binary, undeniable measure of whether a design works — it's especially persuasive in stakeholder presentations because 'only 4 out of 10 users could complete this' lands harder than any qualitative observation.
Avoid when
Don't use task completion rate as your only metric when the task itself has ambiguous success criteria — if users can reach a 'complete' state through a frustrated workaround, the metric will read as a success while the experience remains broken. Pair it with time-on-task and error counts.
A 100% completion rate in a usability study is often a warning sign, not a win — it means your tasks may be too easy, your participants too skilled, or your success definition too generous.
Real-world examples
- GOV.UK's performance platform publishes live task completion rates for every government transaction; the voter registration service showing 92% completion became the benchmark all other services were measured against.
- Amazon's checkout task completion rate is one of the most-studied metrics in e-commerce; their 1-Click patent (now expired) increased checkout completion by an estimated 15% by collapsing a 7-step task to a single interaction.
- Nielsen Norman Group research shows that task completion rates below 78% in unmoderated usability testing are a reliable signal that redesign is required; above 90% indicates the flow is production-ready.