UMUX-Lite: The Two-Item Usability Questionnaire (Items, Scoring and Evidence)

What Is UMUX-Lite?

UMUX-Lite is a two-item questionnaire that measures perceived usability. Participants rate two statements on a 7-point agreement scale:

1. "This system's capabilities meet my requirements."
2. "This system is easy to use."

The two ratings combine into a 0–100 score that tracks closely with the much longer System Usability Scale (SUS). Use UMUX-Lite when you need a standard usability number but can't afford ten questions: in-product intercepts, post-interview wrap-ups, tracking studies, or any survey where length is the main cost.

It was introduced by James R. Lewis, Brian S. Utesch and Deborah E. Maher of IBM at CHI 2013, in a paper titled "UMUX-LITE: When There's No Time for the SUS".

Where UMUX-Lite Comes From

UMUX-Lite is a shortened version of the Usability Metric for User Experience (UMUX), a four-item questionnaire Kraig Finstad published in 2010. Finstad's goal was a short alternative to the ten-item SUS, with items mapped to the ISO 9241-11 definition of usability (effectiveness, efficiency and satisfaction).

The UMUX mixed positively and negatively worded items. Lewis and colleagues found that this mix produced an artificial two-factor structure (positive items clustering apart from negative ones) rather than a single usability factor. Keeping only the two positive items gave a cleaner, one-dimensional measure, and UMUX-Lite was born.

The two items map neatly onto the Technology Acceptance Model: one asks about usefulness (do its capabilities meet my needs?) and one about ease of use. The original paper also found that scores correlate with likelihood-to-recommend, as the evidence below shows.

How Well Does It Work? The Evidence

UMUX-Lite is one of the better-studied short usability measures. Key findings:

• Reliability is high for a two-item scale. The original CHI 2013 paper reported reliability estimates of .82 and .83 across its two survey samples.
• It correlates strongly with SUS. The same paper reported correlations with SUS of .81 and .81, and correlations with likelihood-to-recommend of .74 and .73.
• Raw scores run a little lower than SUS. MeasuringU's summary of the original study notes that the mean difference between SUS and UMUX-Lite was about 3.5 points on the 0–100 scale, which is why the authors proposed a regression adjustment.
• Across many studies, the gap is close to zero. A later summary by Lah, Lewis and Šumak (2020) found that across 16 published comparisons, the mean difference between SUS and UMUX-Lite was −0.4 points, less than half a point on a 100-point scale.

In the authors' own words from the 2013 abstract, "UMUX-LITE score means were slightly lower than those for the SUS, but could be adjusted using linear regression to match SUS scores."

The regression adjustment is no longer recommended

That regression adjustment has since been retired. Lah, Lewis and Šumak (2020) found the regression formula seriously underestimates SUS at higher levels of perceived usability. MeasuringU, which publishes much of the ongoing work on these scales, states: "the regression equation was found to seriously underestimate the SUS at higher levels of perceived usability … so we do not recommend its use."

Current practice is to use the plain linear interpolation score below. Treat older articles that tell you to apply a regression with caution.

How to Score UMUX-Lite

With the standard 7-point scale (1 = Strongly disagree, 7 = Strongly agree):

1. Take each participant's two ratings.
2. Subtract 1 from each (so each runs 0–6).
3. Add them together (0–12).
4. Multiply by 100/12.

UMUX-Lite = ((Item 1 + Item 2 − 2) × 100) / 12

Average the per-participant scores to get the study score. Because UMUX-Lite tracks SUS closely on average, many teams read it against their existing SUS benchmarks. Do that loosely: report it as "UMUX-Lite score", not as a SUS score, and keep in mind that individual studies can differ from SUS by several points.

A 5-point version also exists

Lewis has also published validation work on a 5-point version (2019), which some teams use to match other items in their survey. The scoring logic is the same with a different divisor: ((Item 1 + Item 2 − 2) × 100) / 8. Keep the scale consistent across waves of a tracking study; switching between 5 and 7 points makes comparisons harder.

When to Use UMUX-Lite (and When Not To)

Use UMUX-Lite when:

• You want a standard usability number but survey length is tight.
• You're adding a usability check to a study whose main purpose is something else, such as a discovery interview or a churn survey.
• You're tracking a product over time and need a light, repeatable metric.
• You want separate signals for usefulness and ease, since each item is interpretable on its own.

Use the full SUS instead when:

• You need direct comparability with a long history of SUS data in your organization.
• You're running a formal summative evaluation where a reviewer expects SUS.
• You have room for ten items and want the extra precision of a longer scale.

Use a task-level measure instead when you want to know about one specific task rather than the product overall. The Single Ease Question (SEQ) is the standard post-task item; UMUX-Lite is a post-study, whole-product measure.

How to Run UMUX-Lite Well

Keep the exact wording

Keep the wording identical across studies so scores stay comparable. Use the two statements as written, with "this system" replaced by your product name if that reads more naturally ("Acme's capabilities meet my requirements"). Keep the order: capabilities first, then ease.

Ask it after real use

UMUX-Lite measures perceived usability, so the participant needs recent, real experience. Ask it at the end of a usability session, after onboarding is complete, or of users who have been active recently. Asking people who signed up and left after five minutes measures something closer to first impressions.

Label both endpoints

Anchor 1 as "Strongly disagree" and 7 as "Strongly agree". Leave the middle unlabeled or label only the midpoint.

Report both items, not just the composite

The composite is useful for benchmarking, but the two items often tell different stories. A product can score high on ease and low on capabilities (simple but limited) or the opposite (powerful but hard). Report both item means alongside the score.

Pair it with a "why"

A number tells you how people feel, not why. Follow each item, or at least the lower one, with an open question: "What would need to change for you to agree more strongly?" That turns a benchmark into a list of fixes.

Mind the sample size

With small samples, a UMUX-Lite average can move a lot from one study to the next. Report a confidence interval alongside the mean so readers can see how precise the number is. Our guide to confidence intervals for small-sample UX research shows how.

How Koji Helps

Koji is an AI research platform that runs interviews by text or voice and produces a report as responses arrive. It is well suited to UMUX-Lite because it collects the rating and the reasons behind it in the same conversation.

Add both items as scale questions. Koji's structured questions include a scale type with a range you set (1 to 7 for standard UMUX-Lite) and labels for each endpoint. Add the two statements as two scale questions with "Strongly disagree" and "Strongly agree" as the anchors. See the structured questions guide.

Get the "why" automatically. For each scale question you can set how deeply the AI interviewer follows up (up to three follow-ups) and turn on a follow-up anchored to the participant's own score, which asks what would change it. Participants who give the capabilities item a 4 are asked what's missing, in the same interview, by text or by voice.

See the numbers immediately. The report shows each scale question as a distribution with its mean and median, and links participant quotes to the answers so you can read why a score was low. The Responses tab shows every participant's answers in a grid with a summary row, and exports to CSV with one click, so you can apply the UMUX-Lite formula per participant in a spreadsheet.

Only usable interviews count. Each interview gets a quality score from 1 to 5, and only interviews scoring 3 or more enter the report. Low-effort responses (people clicking through ratings without engaging) don't drag your benchmark around.

Use it alongside discovery. Because the two items take seconds, you can add them to the end of any Koji study, such as a discovery interview, a churn interview or a feature feedback study, and build a usability trend without running a separate survey.

Compared with a standalone survey tool, the advantage is that the benchmark and its explanation come from the same participant in the same session, so you never have to run a second round of research to find out why a score moved.

Related Resources

• Structured Questions Guide: set up scale questions with custom ranges and labels
• System Usability Scale (SUS) Guide: the ten-item scale UMUX-Lite was built to replace
• Single Ease Question (SEQ) Guide: the standard task-level companion metric
• Usability Metrics Guide: success rate, time on task and error rate
• Scale Questions Guide: choosing ranges and anchors for rating questions
• Confidence Intervals for Small-Sample UX Research: how precise is your score?

Sources

• Lewis, J. R., Utesch, B. S., & Maher, D. E. (2013). UMUX-LITE: When There's No Time for the SUS. Proceedings of CHI 2013, 2099–2102.
• Finstad, K. (2010). The Usability Metric for User Experience. Interacting with Computers.
• Lah, U., Lewis, J. R., & Šumak, B. (2020). Perceived Usability and the Modified Technology Acceptance Model. International Journal of Human-Computer Interaction.
• MeasuringU, "Measuring Usability: From the SUS to the UMUX-Lite" (measuringu.com/umux-lite).