The 10 items alternate between positive and negative tone (odd-numbered items are positive, even-numbered negative) to reduce response bias. Scoring: odd items contribute (response − 1), even items contribute (5 − response); sum, multiply by 2.5. The result is 0–100 but is not a percentage.
Jeff Sauro’s 2011 analysis of hundreds of SUS studies established the industry average at 68 and introduced percentile ranks that let practitioners compare their score against the published distribution. Scores above 68 are above average; above 80 are considered “excellent”; below 51 are “poor”.
SUS has three properties that keep it dominant despite its age: it’s short (adds under five minutes to any usability test), it’s well-benchmarked (Sauro’s dataset lets you contextualise any score), and it’s technology-agnostic (works equally well for websites, apps, hardware, voice interfaces). No competing questionnaire combines all three.
At the end of any moderated or unmoderated usability test. Before and after a redesign to measure change. In benchmark studies comparing your product against competitors. Not useful during formative research — SUS tells you how usable the product felt, not what’s wrong with it. Pair with task success rates and qualitative themes.
A SaaS redesign ran SUS pre- and post-launch with 25 users each. Pre-launch score: 61 (poor-to-OK). Post-launch score: 76 (good). The team used Sauro’s percentile table to report: “we moved from the 30th percentile of SUS scores to the 75th” — a form stakeholders understood immediately.
Score it with the SUS calculator — it handles the inverted even items and produces the percentile automatically.
SUS sits alongside NPS (satisfaction, not usability), task success rate (effectiveness), and time-on-task (efficiency) as the measurable layer of usability. See usability for the broader attribute.
Reviewed 2 October 2026 · Editorial standards