System Usability Scale (SUS): One Number for Usability

"The app feels easier now." That's the kind of claim usability work drowns in — real, but unmeasurable, and impossible to compare against last quarter or against a competitor. The System Usability Scale, created by John Brooke in 1986, fixes that with ten standard questions that collapse into one number.

Because the questions never change, SUS scores are comparable across products, versions, and time — and against a huge body of published benchmarks. That comparability is the whole point.

Note

SUS trades detail for standardization. Ten fixed questions won't tell you which button confused people, but they give you a single score you can benchmark and track — something ad-hoc usability questions never can.

The Ten Questions

Respondents rate ten statements on a 1–5 agree/disagree scale. The statements alternate: odd-numbered ones are positive ("I'd use this frequently"), even-numbered ones negative ("I found it unnecessarily complex").

Tip

The alternating wording is deliberate — it keeps people from mindlessly agreeing down the list. It also means the scoring has to flip the negative items, which is exactly where hand-calculation goes wrong.

How the Score Is Built

The raw 1–5 answers become a 0–100 score through a specific recipe — not a simple average.

1For odd (positive) items, subtract 1 from the response.
2For even (negative) items, subtract the response from 5.
3Sum all ten adjusted scores (each now 0–4).
4Multiply the total by 2.5 to land on a 0–100 scale.

Watch out

A SUS score is not a percentage. A 68 does not mean "68% good" — it means average. The 0–100 range is a scale, not a proportion, which is the single most common misreading of SUS.

Reading the Number

A SUS score only means something against benchmarks. The widely-used interpretation from Bangor and colleagues maps the number to a grade and an acceptability judgment.

Score	Reading
≥ 80.3	Grade A — excellent
≈ 68	Average — the historical midpoint
≥ 70	Acceptable
50 – 70	Marginal
< 50	Not acceptable

A percentile makes it even clearer: a score of 68 sits at roughly the 50th percentile of all products tested, so "68" literally means "middle of the pack."

SUS in the SKARI Survey Lab

SKARI's Survey Lab scores SUS correctly and interprets it for you — handling the reverse-coded items so the number is right, then placing it on the benchmark scale.

The 0–100 SUS score with reverse-coded items handled automatically
Letter grade and percentile against standard benchmarks
Usability and learnability sub-scales
Item-level breakdown to see which statements dragged the score

Takeaway

You get the correct score, its grade and percentile, and an item-level view — so you know not just how usable it is, but which questions pulled the number down.

Frequently Asked Questions

Is 68 a good score?

It's exactly average. Aim above 70 for "acceptable" and above 80 for a Grade A experience.

How many people do I need?

SUS stabilizes with a modest sample — a couple dozen users gives a usable score, more tightens the estimate.

SUS or a custom usability survey?

Use SUS when you want a benchmarkable number over time; add task-specific questions when you need to know what to fix.

Key Takeaways

Questions

fixed, standard

Scale

0–100

not a percentage

Average

the midpoint

Grade A

≥ 80.3

excellent

The SUS turns a fuzzy sense of "easy to use" into a single, benchmarked number. Score it correctly, read it against the grades and percentile — not as a percentage — and you can track usability like revenue or retention.

Takeaway

Ten fixed questions, one comparable number — the value of SUS is that a 72 today means the same thing as a 72 next year, or a rival's 72.

CSAT and NPS

Satisfaction and loyalty metrics

Importance-Satisfaction (IPA)

Where to focus UX fixes

Kano Model

Which features drive satisfaction