Expert evaluation (heuristic evaluation): find usability problems before recruiting users

Expert evaluation (also called heuristic evaluation) is a review conducted by one or more people who know interaction design and usability. Evaluators work through a set of heuristics or guidelines, note where the design meets or misses them, and rate the severity of issues they find. The method applies at nearly any stage—sketches, prototypes, or finished products—and is cheaper than formal usability testing because it skips recruitment, lab time, and scheduling.

• Evaluators with practical knowledge of good design and current interface conventions.
• A defined set of heuristics or checklists to guide reviews.
• A way to record findings and severity so they can be triaged and tracked.

Nielsen's ten heuristics (practical summary)

Jakob Nielsen's set of ten heuristics is widely used because the rules are general and work across different interfaces. They frame the usual usability goals designers aim for:

• Visibility of system status: users should always know what is happening (for example, current basket contents, progress indicators, or spinners).
• Match between system and the real world: use language and conventions that fit users' mental models and natural mappings (like slider direction matching quantity increase).
• User control and freedom: provide clear undo, back, and correction paths, especially in multi-step flows.
• Consistency and standards: keep labels, actions, and conventions stable across sections so users don't have to relearn.
• Error prevention: design to prevent mistakes (for example, accept varied input formats and normalize them in code).

• Flexibility and efficiency of use: support both novices and power users (shortcuts for experts, discoverable features for novices).

Expert evaluation uncovers a significant proportion of usability problems before any user session. It's faster and cheaper for early-stage checks. Usability testing with real users finds issues that depend on users' actual behavior and context—things experts can miss. Treat the two as complementary: use expert reviews to iterate quickly, then validate with users.

The same principles can be formalized into benchmarking: define detailed criteria, score designs, and compare performance across sites or over time. Benchmarking requires more structure and consistent scoring but produces comparable metrics useful for tracking improvements or comparing alternatives.

Run expert evaluations early to catch many obvious and moderately complex usability issues quickly. Use them as a preparatory step before investing in user recruitment and testing, and formalize the process when you need comparable measures across designs.