Heuristic Evaluation: Nielsen's 10 Heuristics with Examples

Heuristic evaluation is an expert-based method for usability testing. Between three and five UX experts assess an interface using Jakob Nielsen's 10 heuristics, document violations, and prioritize them by severity. The method does not require actual users.

This article explains each of the 10 heuristics with a real-world practical finding, outlines the five-step process, and clarifies the severity scale.

Key facts about heuristic evaluation at a glance:

The Method: An expert-based usability assessment. Fast, cost-effective, and applicable even to prototypes.

The Foundation: Jakob Nielsen's 10 general principles for interaction design, which serve as broad rules of thumb. The first version was published in 1990 (Molich & Nielsen), and the refined 1994 version still contains the most widely applied usability principles in the industry today.

The Evaluators: A single evaluator typically finds only about 35% of usability problems, while five evaluators find around 75%. Three is the practical minimum for a tight budget, and five is already optimal in terms of cost-benefit ratio.

The Limitation: The method reveals problems from an expert's perspective. Only a usability test shows what users actually do in real-world use.

Nielsen's 10 Heuristics (with practical examples)

Nielsen originally developed the heuristics in 1990 together with Rolf Molich (Molich & Nielsen) and refined them four years later, in 1994, based on a factor analysis of 249 usability problems. Today, they are maintained and published by the Nielsen Norman Group. Since then, they have been the standard framework for heuristic evaluations worldwide.

Each heuristic is followed by a brief explanation and an anonymized finding from our audit practice.

1. Visibility of System Status

The system should always inform users about what is happening. This is achieved through appropriate feedback within a reasonable time. If this feedback is missing, uncertainty and double-clicks arise.

Example: Instead of a frozen button, a progress bar with a percentage during file upload shows that something is happening.

2. Match Between System and Real World

The interface speaks the users' language, not the system's. Concepts, words, and sequences are based on the user's real world, not internal database structures.

Example: A booking tool names the field "Travel Date" instead of "date_field_01".

3. User Control and Freedom

Users sometimes make mistakes and need a clearly visible "emergency exit" without long detours: Undo, Cancel, Back.

Example: A multi-step form has a clearly visible "Back" button on every page.

4. Consistency and Standards

Similar actions and terms should always look and be named the same. Users want to transfer what they've learned from one place to the next, and inconsistency forces them to rethink every time.

Example: The primary action button is colored and named consistently across all pages.

5. Error Prevention

Good design prevents errors through locks, confirmation dialogues, clear constraints, and inline validation, before they occur.

Example:The password field validates the requirements live during input, not just upon submission.

6. Recognition Rather Than Recall

Options, actions, and objects should be visible so users can recognize them rather than having to recall them from memory. No mental effort for routine tasks.

Example: Recent searches appear as suggestions as soon as you click into the search field.

7. Flexibility and Efficiency

Experienced users want shortcuts: keyboard shortcuts, saved settings, macros. Beginners should not be overwhelmed, and experts should not be slowed down.

Example: Experienced users can trigger actions via keyboard shortcuts (e.g., Ctrl+S).

8. Aesthetics and Minimalist Design

Any additional information competes with relevant information. Dialog boxes should only display what is immediately needed.

Example: Infrequently used options are hidden behind 'Advanced Settings' instead of being permanently visible.

9. Help Users Recognize, Diagnose, and Recover from Errors

Error messages should be expressed in plain language, precisely state the problem, and suggest a constructive solution.

Example: A failed payment states the reason ('Card expired') and the next step.

10. Help and Documentation

Although the system should ideally be usable without documentation, help, when needed, must be easily findable, task-oriented, and specific.

Example: A '?' icon next to a complex field opens a short, context-sensitive explanation.

‍

Conducting a Heuristic Evaluation: The 5-Step Process

The heuristic evaluation follows a clear process. These five steps lead to a prioritized action plan.

‍

Step 1 – Scope and Briefing

‍Define the evaluation scope: which pages, which flows, which user scenarios? Provide personas and context, brief evaluators. A clearly defined scope prevents evaluators from straying off course.

‍

Step 2 – Individual Assessment

‍Each evaluator independently reviews the interface using the 10 heuristics. Findings are documented with a screenshot, the affected heuristic, and an initial severity level (more on this shortly). Important: no consultation during this phase. Independent assessments prevent groupthink.

‍

Step 3 – Documentation of Findings

‍Use a consistent format: heuristic, description, screenshot, severity level (0 – 4), affected area.

‍

Step 4 – Team Consolidation

‍Consolidate all findings, merge duplicates, and determine final severity levels by consensus. Recommendation: Workshop format, 60 – 90 minutes. In this round, it often becomes clear that different evaluators have identified the same problem from different perspectives. This strengthens the quality of the findings.

‍

Step 5 – Recommendations and Prioritization

‍Sort findings according to an Impact/Effort matrix and, based on that, create a prioritized action plan, ensuring its implementation. Anyone who skips this step will have conducted the evaluation for nothing.

Severity Levels: Prioritizing Findings Correctly

Not every finding is equally urgent. For each violation found, evaluators assign a severity rating from 0 to 4 to prioritize the most important issues to be addressed. The following table shows the scale and the typical recommended action for each level:

‍

An important note on the process: Severity ratings should be assigned individually and only finalized as a group after the consolidation round. This helps avoid anchoring effects – if the first evaluator loudly declares "Severity 4," others tend to be influenced by it.

For more than 30 findings, we recommend an additional Impact/Effort column. This highlights quick wins – issues with high impact and low implementation effort rise to the top.

Strengths and Limitations of the Method

Every method has its place. Heuristic evaluation also has clear strengths and equally clear limitations that one should be aware of.

‍

Strengths:

‍

• Can be used quickly, even on a prototype or wireframe.
• Cost-effective compared to user tests.
• Repeatable: before and after changes, at any project stage.
• Provides a structured action plan instead of vague opinions.
• Identifies problems for which users have long developed a workaround, and which therefore remain invisible in tests.

‍

Limitations:

‍

• Identifies expert problems, not necessarily real user problems. What evaluators consider a violation is not necessarily what users struggle with in practice.
• The outcome heavily depends on the quality and experience of the evaluators.
• No quantitative statements about task completion rates, error rates, or processing times.

‍

What users actually do in real-world use is only revealed by a different approach: direct observation during a moderated usability test. Both methods complement each other. In many projects, we recommend heuristic evaluation as a first step to identify and fix the most obvious problems. Afterward, a usability test with real users validates whether critical tasks actually work. This creates an optimal cost-benefit ratio.

‍

The Role of Heuristic Evaluation in a UX Audit

Heuristic evaluation is not a complete UX audit, but rather a single component of it. In a typical audit, it is combined with data triangulation (analytics, heatmaps) and, if necessary, with user tests.

What does this mean in practice? The evaluation identifies the violations, but only a complete audit connects these violations with business KPIs:

Where do we lose users?

Which conversion hurdles correlate with the identified problems?

‍

Anyone conducting a heuristic evaluation internally needs at least three UX specialists, clear briefings, and time for consolidation. For complex products or when BFSG compliance needs to be checked, the internal team often reaches its limits. Beyond this threshold, external support in the form of a professional UX audit with heuristic evaluation as an integral part.

‍