“How confident would you feel using this without anyone's help?”
From John Brooke · Usability researcher at Digital Equipment Corporation; creator of the System Usability Scale
Why it works
The real SUS is not one question — it's a battery of ten statements rated on a five-point agree/disagree scale, deliberately alternating in tone: the odd items are positively worded ('I thought the system was easy to use', 'I felt very confident using the system') and the even items negatively worded ('I found the system unnecessarily complex', 'I would need the support of a technical person to use this system'). That alternation is the trick — it forces respondents to read each item rather than straight-lining down one side, and it cancels acquiescence bias. The ten raw answers are recoded and summed into a single 0–100 score (note: not a percentage), which is why a 'quick and dirty' scale Brooke assembled by selecting ten discriminating items from a pool of fifty has become the most widely used usability measure there is — comparable across products, versions, and decades. The single confidence-without-help question on this card is a faithful representative of the battery (it blends the confidence and the technical-support items), useful as a lightweight probe — but it is a stand-in, not the score.
When to ask
After a user has genuinely worked with the product, when they can rate each statement against real experience. SUS measures perceived usability, so it needs lived use behind it — not a first glance.
Good follow-ups
- What would you need before you'd feel fully confident on your own?
- Where did you last reach for help, or wish you could?
- Which part felt the most complex or inconsistent?
Watch out for
Mistaking this one item for a SUS score. A real SUS result comes only from administering all ten statements and applying the recode-and-sum formula; a single question yields a feeling, not the validated 0–100 number, and the benchmark (roughly 68 as average) only applies to the full instrument at sample size. Two more traps with the real battery: dropping the negatively worded items to 'simplify' it (which quietly reintroduces the response bias the alternation was built to remove), and reading the score as a percentage — an 80 is not '80% usable', it's a position on a standardized distribution.
Where to ask
- In-product surveygreat
The native habitat — fire the battery after a user has worked with the product, so each statement is rated against real experience. SUS is a survey instrument by design.
- Long-form surveygreat
Travels cleanly in an emailed usability survey; the ten alternating items are self-contained and the 0–100 score only means something at sample size, which a survey delivers.
- User interviewworkable
Strong as a single representative probe in a moderated session — but one person's rating isn't a SUS score. The 0–100 number requires the full ten-item battery across a real sample.
Reworded for this context: “How confident would you feel using this without anyone's help — what would you need to get there?”
- Support chatpoor
A full ten-item battery is far too heavy to drop into a live support exchange, and a single lifted item there measures the open issue, not the system's overall usability.
- Sales callpoor
SUS measures perceived usability of a system the respondent has actually used. A prospect has nothing to rate, so the confidence question becomes hypothetical and the answer is noise.
Stage: Consider · A question popularized by John Brooke