SUS Score Calculator & Usability Test Metrics

About the SUS Score Calculator & Usability Test Metrics

The System Usability Scale (SUS) is the ten-statement questionnaire usability researchers give people after they have used a product: five positive statements (“I thought the system was easy to use”) alternate with five negative ones (“I found the system very cumbersome to use”), each answered from 1, strongly disagree, to 5, strongly agree. Paste your participants’ answers and this calculator scores each of them from 0 to 100, then gives the mean SUS score with its confidence interval, its grade and percentile on Sauro and Lewis’s curved grading scale, the closest adjective rating from Bangor, Kortum and Miller, and the mean answer to every statement.

The same sheet takes the rest of a usability test: a second design or round to compare with Welch’s t-test, the two UMUX-Lite items with their regression onto the SUS scale, and one line per participant and task for completion rates with adjusted-Wald intervals, time on task as a geometric mean, errors and the Single Ease Question, charted per task and per participant. Answers that look inverted — the same answer to all ten statements, or agreeing with the positive and the negative ones alike — are flagged, and you can leave those participants out.

Without a pass, the free preview shows the mean SUS score and its grade in full, with the first rows of each table and watermarked charts; every other figure, the copy and the CSV, PDF and PNG downloads need a Pro pass.

Common uses

• Scoring the SUS after a moderated or unmoderated usability test, with the interval and grade a stakeholder report needs.
• Comparing the SUS of a redesign with the current design, or of two rounds of testing, before deciding whether the change helped.
• Benchmarking task completion, time on task and ease (SEQ) per task across participants.
• Converting a short UMUX-Lite survey to the SUS scale for comparison with earlier SUS studies.
• Spotting questionnaires that were answered without reading the alternating statements.

How the SUS is scored

Each answer is a scale position from 1 to 5. Items 1, 3, 5, 7 and 9 contribute the answer minus 1; items 2, 4, 6, 8 and 10 contribute 5 minus the answer; the sum of the ten contributions (0 to 40) times 2.5 is the score from 0 to 100. Scores of single items are not meaningful on their own (MeasuringU’s guide to SUS scoring). The scale’s author, John Brooke, asks that every item is answered and that a respondent who cannot answer one marks the centre point (“SUS: a quick and dirty usability scale”), which is why a missing answer counts as 3 by default here. The all-positive version words every statement positively, so every item contributes the answer minus 1.

The score is a mean of people, so the calculator also gives its t-based confidence interval: mean ± t × standard deviation ÷ √n, with t from Student’s distribution for n − 1 degrees of freedom. With 12 participants and a standard deviation of 15, the 95% interval is about ±9.5 points.

Grades, percentiles and adjectives

Sauro and Lewis’s curved grading scale (Quantifying the User Experience, second edition) puts the mean of many SUS studies at 68, a C, and grades a mean score like this:

• A+ 84.1–100 (96th–100th percentile) · A 80.8–84.0 (90–95) · A− 78.9–80.7 (85–89)
• B+ 77.2–78.8 (80–84) · B 74.1–77.1 (70–79) · B− 72.6–74.0 (65–69)
• C+ 71.1–72.5 (60–64) · C 65.0–71.0 (41–59) · C− 62.7–64.9 (35–40)
• D 51.7–62.6 (15–34) · F 0–51.6 (0–14)

The scale gives a percentile range per grade (the table on MeasuringU, which splits F into 0–25, the 0–1.9th percentile, and 25.1–51.6, the 2nd–14th); the single percentile shown here is interpolated in a straight line inside each row of that table (65.0 → 41st, 71.1 → 60th, 25.1 → 2nd), so 68 comes out at the 50th.

Bangor, Kortum and Miller asked participants to describe the product with one adjective as well as the SUS, and published the mean SUS score of each adjective: Worst imaginable 12.5, Awful 20.3, Poor 35.7, OK 50.9, Good 71.4, Excellent 85.5, Best imaginable 90.9. The page names the adjective whose mean is closest to your score.

Comparing two designs or rounds

Paste the second set of answers under Compare and the calculator runs Welch’s t-test on the two sets of scores: the difference B − A, its confidence interval, the t value, the degrees of freedom (which Welch’s formula adjusts for unequal spreads and group sizes) and the two-sided p-value. A p-value below 0.05 at 95% confidence means a difference this large would be unlikely if both designs were equally usable; the interval says how large the difference could plausibly be.

Task metrics

• Completion rate: successes ÷ attempts, with the adjusted-Wald interval of Agresti and Coull: add z²/2 successes and z² attempts (1.92 and 3.84 at 95%) and build a normal interval around the adjusted rate, cut at 0% and 100%. It stays sensible for the small samples of usability tests. A skipped task is not an attempt; a partial success counts as not completed unless you tick Partial success counts.
• Time on task: the geometric mean with a t interval on the logarithms of the times, which Sauro and Lewis recommend for the skewed times of small tests; by default only successful attempts are timed.
• Errors: the mean number per attempt with its t interval.
• SEQ: the Single Ease Question, “Overall, how difficult or easy was the task to complete?”, answered from 1 to 7 right after each task (MeasuringU on the SEQ); the mean per task with its t interval.

The task-by-participant matrix shows every outcome at a glance, so a task that only one group failed stands out.

UMUX-Lite

UMUX-Lite asks two statements on a 7-point scale: “This system’s capabilities meet my requirements” and “This system is easy to use”. Its score is (item 1 + item 2 − 2) × 100 ÷ 12, from 0 to 100, and Lewis, Utesch and Maher’s regression 0.65 × UMUX-Lite + 22.9 puts it on the SUS scale (UMUX-LITE: when there’s no time for the SUS; the formula in use). The SUS-scale value runs from 22.9 to 87.9 and is graded on the same curved scale.

Limitations

• The percentile is interpolated inside the grading scale’s printed ranges: the scale itself gives ranges, not a percentile for every score.
• Welch’s test treats the two sets as different people. When the same participants rated both designs, a paired test is the right one: the T-Test Calculator has it.
• The inverted-answer flags are signs, not proof: check a participant’s comments or recording before leaving them out.
• One study at a time: segments, weighting and benchmarks by product type are not included.
• UMUX-Lite is read in its original seven-point form; answers given on a five-point scale are not converted.
• In the free preview the mean SUS score and its grade are shown in full, with the first rows of each table (up to 3) and watermarked charts; every other figure is hidden, and copying and downloads stay locked.

Privacy

On this device: nothing is sent

Everything happens in your browser. What you enter or open here is not uploaded or stored by MySmartCoPilot.