A SUS Score of 70 Is a C. Most Teams Report It as a Pass.


person holding pink sticky note

A System Usability Scale score of 70 is a C. Not a 70 percent, not a pass, not “pretty good.” On the curved grading scale most usability researchers now use, a 70 sits in the middle of the pack and carries the label “marginal.” Yet I have watched a 70 get reported upward as a win more than once, because it looks like a grade on a school test and everybody in the room grew up knowing what 70 out of 100 means.

The SUS score is one of the most useful numbers a product team can collect in discovery and one of the most misread. This is a short reference for reading it properly.

What the scale is, in two paragraphs

John Brooke built the SUS at Digital Equipment Corporation in 1986 as a quick way to measure how usable people felt an office system was, and it has since become the default. James Lewis, in a 2018 review in the International Journal of Human–Computer Interaction, called it the most widely used standardized questionnaire for perceived usability. It is ten statements, each answered on a five-point agree/disagree scale, usually right after someone has used the product to complete real tasks.

Scoring has one trap. The odd-numbered statements are positive (“I thought the system was easy to use”) and the even-numbered ones are negative (“I found the system unnecessarily complex”). For odd items you subtract 1 from the answer; for even items you subtract the answer from 5. Add the ten results and multiply by 2.5. The output runs from 0 to 100, which is precisely why people mistake it for a percentage.

The number only means something next to other numbers

Jeff Sauro’s analysis of scores from 500 products put the average SUS at 68. His own warning, from the same piece: SUS scores are not percentages. A 68 is 68 percent of the maximum score, but it falls right at the 50th percentile. Half of the products measured beat it.

Sauro and Lewis later turned that distribution into a curved grading scale. A condensed version:

SUS score Grade Percentile range Acceptability
84.1 to 100 A+ 96 to 100 Acceptable
80.8 to 84.0 A 90 to 95 Acceptable
74.1 to 77.1 B 70 to 79 Acceptable
65.0 to 71.0 C 41 to 59 Marginal
51.7 to 62.6 D 15 to 34 Marginal
25.1 to 51.6 F 2 to 14 Not acceptable

Two details in that table matter more than the rest. First, the range is compressed: climbing from the top of the C band to an A takes about ten points, so a product sitting at 72 is closer to average than its stakeholders think. Second, the adjective ratings Bangor, Kortum and Miller collected alongside SUS (summarized here by Sauro) attach the word “OK” to scores around 51. In the same write-up, Sauro notes that later research suggests “Fair” may be the better word, because “OK” implies an acceptable experience when it is at best marginal. If your research report says users rated the product “OK,” check which number that word is standing in for.

Three ways teams fool themselves with it

Treating small moves as signal. The sample sizes that work for finding usability problems are too small to measure a score precisely. According to Sauro’s data on small samples, with five participants the sample mean lands within roughly 17 points of the true score 95 percent of the time. Seventeen points is roughly the distance from a D to an A. If your SUS went from 69 to 74 between two rounds of five users, you have not learned that the redesign worked. You have learned that you need more participants or a confidence interval in your slide. I made the same case about quoting the five-user rule: five people is plenty for finding problems and nowhere near enough for measuring them.

Scoring it wrong. The alternating positive and negative items exist to stop people from ticking the same column all the way down. They also trip up the people doing the arithmetic. In Sauro’s review, 11 percent of researchers mis-scored the SUS because they forgot to reverse the even items, and in 17 percent of studies some participants had forgotten to flip their own answers on the negative statements. If a score looks surprisingly good or surprisingly bad, recheck the coding before you rewrite the roadmap. A spreadsheet with the formula locked is worth the five minutes.

Asking it to diagnose. SUS tells you how usable people perceived the product to be. It does not tell you why, and it was never meant to. Sauro is blunt about this: none of the individual items really tell you what to fix in the interface. The fix lives in the task observations and the comments, which is why SUS belongs at the end of a moderated or unmoderated test, not in a standalone email survey. A SUS survey sent to your whole customer base, with no tasks attached, measures a vague impression of the brand as much as the product. That is the same problem I described in surveys that answer the wrong question.

Why the scale on the slide matters more than the number

In more than twenty years running IT operations at a large telecom, the metric that caused the most confusion in quarterly reviews was rarely wrong. It was read on the wrong scale. A service at 99.5 percent availability sounded excellent to executives who were used to grading on percentages. On a year of 8,760 hours, it also meant roughly 44 hours of downtime, which nobody in that room would have signed off on if it had been stated that way. We started showing the hours next to the percentage, and the conversation changed in a single meeting.

A SUS score needs the same translation. When you present one, present three things together:

  1. The score with its percentile and grade. “71, which is roughly the 59th percentile, a C” is honest. “71 out of 100” invites a misread.
  2. The sample size and a range. “Based on 8 participants; the true score could reasonably be several points either way” protects you from celebrating noise.
  3. The two or three task failures that explain it. The number earns attention; the observations earn the fix.

There is one more reason to get this right. Sauro’s analysis of 4,664 respondents found that SUS explains between 30 and 50 percent of the variation in how likely users are to recommend a product, with promoters averaging a SUS close to 81. If your leadership team watches Net Promoter Score, perceived usability is one of the few levers you can measure in a test session before the NPS survey ever goes out. That makes it worth reporting accurately, especially when the accurate version is less flattering. A green status on a red project does more damage than the red would have, and a C reported as a pass is the same mistake in a research deck.

Ty Sutherland

Ty Sutherland is the editor of Product Management Resources. With a quarter-century of product expertise under his belt, Ty is a seasoned veteran in the world of product management. A dedicated student of lean principles, he is driven by the ambition to transform organizations into Exponential Organizations (ExO) with a massive transformative purpose. Ty's passion isn't just limited to theory; he's an avid experimenter, always eager to try out a myriad of products and services. While he has a soft spot for tools that enhance the lives of product managers, his curiosity knows no bounds. If you're ever looking for him online, there's a good chance he's scouring his favorite site, Product Hunt, for the next big thing. Join Ty as he navigates the ever-evolving product landscape, sharing insights, reviews, and invaluable lessons from his vast experience.

Recent Posts