How to measure user experience in your product
Measure user experience with a mix of behavioural metrics (task success, time on task, error rate, retention) and attitudinal metrics (System Usability Scale, satisfaction, Net Promoter Score). Google’s HEART framework (Happiness, Engagement, Adoption, Retention and Task success) helps pick the right few for your goals.
Why measure user experience
Teams make design changes all the time, but without measurement they cannot tell whether those changes made things better, worse or made no difference at all. Opinions about what users like are unreliable, and even experienced designers are often surprised by how real people behave. Measuring user experience turns these opinions into evidence, helping teams prioritise work, justify investment and learn what actually helps their users.
Measurement does not need to be complicated. A small set of well-chosen metrics, tracked consistently over time, usually tells a team more than a large dashboard of numbers nobody looks at. The key is to choose metrics that reflect real user goals and the outcomes the business cares about.
Two kinds of UX metric
| Type | Examples | Tells you |
|---|---|---|
| Behavioural | Task success, time on task, errors, retention | What users do |
| Attitudinal | SUS, satisfaction (CSAT), NPS | How users feel |
Behavioural metrics in detail
| Metric | What it measures | How to collect it |
|---|---|---|
| Task success rate | Share of users who complete a task | Usability tests, analytics funnels |
| Time on task | How long a task takes | Usability tests, product analytics |
| Error rate | How often users make mistakes | Usability tests, error logs |
| Adoption | Share of users who try a feature | Product analytics |
| Retention | Share of users who return | Product analytics, cohorts |
Behavioural metrics show what people actually do, which makes them reliable indicators of usability. They are most useful when tracked for specific tasks, such as completing a booking or finding a document, rather than for the product as a whole.
Attitudinal metrics in detail
Attitudinal metrics capture how people feel. Customer satisfaction scores ask users to rate a specific experience, often right after a task. The Customer Effort Score asks how easy it was to get something done. Net Promoter Score asks how likely people are to recommend the product. These measures are quick to collect, but they reflect perceptions rather than behaviour, so they work best alongside behavioural data.
Google’s HEART framework
- Happiness: Satisfaction and perceived ease, often via surveys.
- Engagement: How often and how deeply people use the product.
- Adoption: New users or new use of a feature.
- Retention: Users who keep coming back.
- Task success: Completion rate, time and errors on key tasks.
The System Usability Scale
SUS is a ten-question survey that produces a score from 0 to 100. Scores around 68 are commonly treated as average, which makes it a quick way to compare versions or competitors.
Combining numbers with qualitative insight
Metrics tell you what is happening; qualitative research explains why. If task success drops, watching a few usability sessions or reading support tickets usually reveals the cause. Pairing every key metric with a regular source of qualitative insight helps teams act on numbers with confidence. Our post on sharing research insights covers how to present both together.
Setting targets
Metrics become more useful when they have targets. Base targets on your own baseline rather than industry averages, and set them for specific journeys, such as raising checkout task success from 78% to 90% within two releases. Review targets each quarter. If a target is reached easily, raise it; if it seems out of reach, investigate what is blocking progress before changing it.
Setting up UX measurement
- Pick goals: Choose one or two HEART categories that matter most.
- Define signals: Which user actions show progress toward each goal.
- Choose metrics: Turn signals into numbers you can track.
- Baseline and track: Measure before changes and follow the trend.
For data-rich products, show these metrics clearly to stakeholders, as covered in UX for data-heavy apps. A UX audit is a good moment to set your first baseline.
Building a simple UX dashboard
A UX dashboard should be small enough to review in a few minutes. For each key journey, show one or two behavioural metrics, one attitudinal metric and the trend over time. Review it monthly with the product team, discuss what changed and why, and agree what to investigate next. When a design change is released, mark it on the chart so its effect is easy to see.
- Limit the dashboard to the most important journeys
- Show trends, not just current values
- Mark releases and design changes on charts
- Review monthly with design and product teams
- Link each metric to an owner
Common measurement mistakes
- Tracking too many metrics, so none get attention
- Measuring only satisfaction without behaviour
- Comparing results without a baseline
- Changing several things at once so effects cannot be separated
- Reporting numbers without explaining what they mean
Avoiding these mistakes keeps measurement useful and trusted. For new products, a UX audit is a good moment to set the first baseline.
Frequently asked questions
What is the most important UX metric?
Task success on your key tasks is usually the most direct measure of usability, combined with a satisfaction measure such as SUS.
What is a good SUS score?
Around 68 is commonly treated as average; scores above 80 are generally considered good.
Is NPS a UX metric?
It measures loyalty more than usability, so use it alongside task-based metrics rather than on its own.
How often should we measure UX?
Track behavioural metrics continuously, run satisfaction surveys regularly and benchmark usability with tests at least once or twice a year.
Which UX metrics should a small team start with?
Pick one key task and track its success rate and a simple satisfaction question. Add more metrics only when you have time to act on them.
Can we compare our UX metrics with competitors?
For some measures, such as SUS, yes. For most others, comparing with your own past results is more reliable.