Nielsen's 5 Components of Usability
Jakob Nielsen's five components of usability: learnability, efficiency, memorability, errors, and satisfaction, and how to use them to evaluate a design.
What are the five components?
Nielsen Norman Group presents them in its usability introduction as five questions you can ask about any design:
- Learnability: how fast do people learn to accomplish basic tasks with the design?
- Efficiency: once people have learned the design, how quickly do they perform tasks?
- Memorability: when someone returns to the design after a period of not using it, how easily do they reestablish proficiency?
- Errors: how many errors do people make, how severe are they, and how easily do they recover from them?
- Satisfaction: how pleasant is the design to use?
The list is not a technical standard. It is a framework that helps you decide what to measure. NN/g adds two cautions: usability is not a property of the design alone (you have to say for whom, for which task, and under what conditions), and being easy to use is not enough, because a product also needs utility, meaning it offers the features people need.
Why does this matter for a junior designer?
Because when someone says “this is not usable”, they almost always mean one of these five components, and each one calls for a different answer. A flow nobody understands the first time has a learnability problem. One that feels slow to expert users has an efficiency problem. One that nobody picks up again after a week has a memorability problem.
Separating them helps you write concrete goals for a test. “We want it to be usable” cannot be measured. “We want a new user to finish sign-up without help” can.
How are the components measured?
In its metrics article, Nielsen Norman Group proposes four basic measures that map well onto several components:
| Measure | What it captures | Related component (this wiki’s reading) |
|---|---|---|
| Task success rate | Whether the user reaches the goal | Learnability, errors |
| Time on task | How long it takes | Efficiency |
| Error rate | How many errors occur | Errors |
| Subjective satisfaction | How the user feels | Satisfaction |
Note that those four metrics do not include a specific measure for learnability or memorability. A reasonable way to measure them (this wiki’s suggestion, not NN/g’s) is to compare a new user’s first attempts at the same task, and to repeat the task after a period away from the product. Treat it as practical guidance, not as a standard.
On sample size, NN/g distinguishes: to find problems in a qualitative test a few users are enough (five per round, across several rounds), but to measure metrics it recommends more, around 20 per design.
How do you apply this when designing?
- Choose the components that matter for your product. An emergency system needs learnability and errors. A tool used every day needs efficiency. A service used once a year needs memorability.
- Turn each component into a test goal. For example: “three of five users complete payment in under two minutes”.
- Design for errors as well as success. The errors component includes recovery. A clear message that says what happened and how to fix it is worth more than a technical error code.
- Design for experts without losing beginners. Efficiency should not sacrifice learnability. Keyboard shortcuts are a good example of that balance.
- Ask about satisfaction after the task, not before. That way the answer reflects what the person just experienced, not what they expect.
For errors, the prevention vs. recovery article goes deeper into the difference between avoiding an error and helping someone get out of it. In practice, the two complement each other.
When should you NOT use this framework?
- When there is no clear user goal. Without a task, there is no way to measure efficiency or errors.
- As a substitute for testing with users. The components guide a test; they do not replace one. That is what usability testing is for.
- As a mechanical checklist. A design can meet all five questions and still be confusing for a specific audience.
- To compare products from different domains. NN/g notes that usability depends on for whom, for which task, and under what conditions. A bank and a game do not weigh the components the same way.
Example
Illustrative example: a team evaluates a medical appointment booking system.
- Learnability: new patients take four minutes to book their first appointment. The goal is under two.
- Efficiency: frequent patients book in under a minute, which is fine.
- Memorability: after six months away, patients forget where the cancel option is. The team decides to add a more visible entry point.
- Errors: 20% of bookings are duplicated. A confirmation step is added before payment.
- Satisfaction: the score drops on the clarity question. The team reviews the wording of its messages.
Each component has a symptom, a measure, and a fix. That is what makes the framework useful.
To evaluate the interface without users, heuristic evaluations can help you find problems before testing. They are complementary tools, not substitutes.
Common mistakes
- Confusing the five components with Nielsen’s ten heuristics. They are different tools.
- Reducing efficiency to speed, without considering what the user has already learned.
- Treating errors only as prevention, without thinking about recovery or severity.
- Asking about satisfaction before the person has used the design.
- Assuming one component matters more than the others in every product.
- Taking usability as enough without checking utility: an easy product that lacks the features people need is of no use.
Sources
- Nielsen Norman Group, Usability 101: Introduction to Usability
- Nielsen Norman Group, Usability Metrics
- Nielsen Norman Group, Why You Only Need to Test with 5 Users