How to Score the System Usability Scale

Where SUS comes from

The System Usability Scale was created by John Brooke. In his 2013 retrospective, published in the Journal of Usability Studies, Brooke wrote that he developed it "Rather more than 25 years ago, as part of a usability engineering program" ¹. University of Washington course notes date the scale to 1986 ².

Brooke designed SUS to be quick and inexpensive to run, so that teams could compare the perceived usability of systems without a lengthy questionnaire. That simplicity helps explain why it is still in wide use decades later.

How the questionnaire is structured

SUS has ten statements. Participants rate each one on a five-point scale from strongly disagree (1) to strongly agree (5). The statements alternate in direction:

• Odd-numbered items (1, 3, 5, 7, 9) are worded positively. They describe things like wanting to use the system often, finding it easy to use, feeling that most people would learn it quickly and feeling confident using it. Agreement is good.
• Even-numbered items (2, 4, 6, 8, 10) are worded negatively. They describe things like finding the system unnecessarily complex, needing technical support, finding it inconsistent and needing to learn a lot first. Agreement is bad.

The alternation is deliberate: it discourages participants from agreeing with every statement without reading. It is also the source of most scoring errors, because the two kinds of item must be scored in opposite directions. Use the standard published wording of the ten statements and do not change their order.

Scoring, step by step

1. For each odd item, subtract 1 from the participant's response. A response of 4 becomes 3.
2. For each even item, subtract the response from 5. A response of 2 becomes 3.
3. Each item now contributes a value from 0 to 4. Add all ten values. The total runs from 0 to 40.
4. Multiply the total by 2.5. The result runs from 0 to 100.

The University of Washington course notes use a totals method instead ²: add up the five odd-numbered responses and subtract 5; add up the five even-numbered responses and subtract that sum from 25; then add the two results and multiply by 2.5. It is the same arithmetic as converting each item first. Subtracting 1 from five odd items removes 5 in total, and subtracting five even items from 5 each starts from 25, so both methods always give the same score.

A worked example

Here is one fictional participant's response set, scored step by step.

The contributions add up to 30. Multiplied by 2.5, that gives a SUS score of 75 for this participant. To get a study score, calculate each participant's score first, then take the mean of those scores.

A SUS score is not a percentage

A score of 75 does not mean the system is 75 percent usable, or that 75 percent of people found it easy. The 0 to 100 range is simply a result of the multiplier, which turns a 0 to 40 total into a scale that is easy to read. Reading it as a percentage misstates what the questionnaire measures.

Published benchmarks exist that compare scores across many studies. If you use one, check its source, how its data was collected and whether it fits your kind of product. Avoid inventing labels such as "good" or "excellent" without a cited basis.

Common scoring mistakes

• Scoring every item the same way. Forgetting to reverse the even items produces scores that cluster near the middle and mean nothing.
• Averaging raw responses. The mean of the raw 1 to 5 responses is not a SUS score.
• Changing the wording. Rewriting statements to fit your product, or making all of them positive, breaks comparability with other studies.
• Dropping items. If a participant skips an item, you cannot calculate a valid score for that participant without a stated method for handling missing values. Decide that method before the study.
• Over-reading small samples. A mean SUS score from five participants has a wide margin of error. Our sample size guide explains why quantitative measures need larger samples.

When to use SUS

SUS works best as a consistent measure over time or across comparable products. Administer it at the end of a session, after all tasks, so participants rate their overall experience. Use the same wording, the same scale and the same point in the session every time.

In a small qualitative study, SUS scores are context, not conclusions. They can show that one participant had a markedly worse experience than the others, which is worth investigating in their session recording. They cannot prove that one design is better than another. For that, you need a larger sample or a controlled comparison, as described in our A/B testing guide.

A quick scoring check

To check a spreadsheet formula, enter two test cases. If every response is 3, the SUS score should be 50. If every odd item is 5 and every even item is 1, the score should be 100. If your formula gives anything else, the even items are probably not being reversed.