metrics
System usability scale
What is the system usability scale (SUS)?
The system usability scale is a short questionnaire with ten statements rated from strongly disagree to strongly agree. It converts how usable something feels into a single score from 0 to 100. It’s quick, standardised, and surprisingly comparable across very different products.
Also known as: SUS, usability score
Prefer to watch?
Watch the recap 0:40
The demo
Rate these statements about a product you just used. Watch them fold into a single 0-100 score - and see where it falls against the average of about 68. (The real SUS has ten statements; this is a faithful five-item taste of it.)
SUS score: 50
Adjust the ratings and watch the score move.
What this demo shows (text version)
Five statements about a product, alternating between positive ("I'd use this often", "I felt confident using it", "It was easy to use") and negative ("It was unnecessarily complex", "It was awkward to use"), each rated from 1 (strongly disagree) to 5 (strongly agree). Each rating is scored the SUS way: for positive statements the score is the rating minus one, for negative statements it is five minus the rating, so agreeing with good things and disagreeing with bad things both raise the score.
The per-item scores are summed and scaled to a 0-100 figure shown live against the SUS average of about 68. The real System Usability Scale uses ten statements; this five-item version teaches the same mechanism. SUS is useful for trending and benchmarking usability, but it tells you how usable something feels, not why.
SUS turns ten quick agree/disagree statements into a single 0-100 usability score that’s standardised enough to compare across very different products. It won’t tell you what to fix, but it gives you a repeatable number to track over time and benchmark against the ~68 average. You can see and try this in action on this page.
You answered a few statements and they collapsed into one score on a 0-100 scale. That's the appeal of SUS: a messy felt experience reduced to a figure you can track over time and compare against a known average of about 68. The catch is remembering it's a temperature, not a diagnosis.
Two quirks make SUS trustworthy. The statements alternate positive and negative, so people can't just tick down one side without reading; and the scoring reflects that (agreeing with a positive item and disagreeing with a negative one both count as good). The 0-100 result is not a percentage - 68 is roughly average, and anything in the 80s is genuinely good.
What SUS gives you is a number to trend and benchmark; what it can't give you is a reason. A score that drops tells you something got worse, not what - for that you still need to watch people use the thing. Pair the metric with [usability testing](/entries/usability-testing/), and never let a tidy 72 stand in for actually understanding your users.