Textbook of Usability

Evolved design practice refers to design conventions that have survived a process analogous to natural selection: designs that worked were imitated, designs that failed were abandoned, over many generations of use. The architectural proportions of Vitruvius, the typographic conventions of Gutenberg, the wayfinding traditions of medieval cities — all represent accumulated observational evidence about what serves human needs.

This evidence is different from controlled experimental evidence: it is observational rather than experimental, retrospective rather than prospective, and correlational rather than causal. But it is evidence nonetheless, and it has two distinctive strengths:

• Enormous sample size: centuries of buildings, inhabitants, and users
• Strong selection pressure: structures that fail to serve human needs tend to be abandoned

This does not mean evolved practice is infallible. The context in which a convention evolved may differ from the context in which it's applied (medieval cathedrals offer limited guidance for mobile UX). Evolved practices can also encode prejudices, inefficiencies, and historical accidents.

The strongest design argument is convergence: when experimental evidence, predictive models, heuristics, and evolved practice all point to the same conclusion, the decision is robust. For example, the case for larger touch targets is supported by Fitts's Law, controlled experiments, platform guidelines, and the practical convergence of iOS and Android minimums on ~44pt.

Evolved practice represents Level 2 in the evidence hierarchy for design (Chapter 20), above intuition but below controlled experiments.