Krug's web usability primer has stayed relevant for over two decades because its core argument — that good interfaces require no instruction — is timeless. The chapters on navigation conventions and usability testing with five users are still the most practical introduction to the discipline.
Krug wrote a short book on web usability whose central rule is that a page should be self-evident, so that a user never has to work out what something is or where to click. He argues that users do not read pages, they scan them; that they do not make optimal choices, they satisfice by clicking the first plausible option; and that they do not figure out how things work, they muddle through with a partial and often wrong model.
The practical guidance follows. Pages should have clear visual hierarchy, conventions should be followed rather than reinvented, clickable things should look clickable, pages should be broken into defined areas, and noise should be eliminated. Krug argues for cutting half the words on every page and then half again. Navigation should answer where am I, and every page needs a name matching what the user clicked to get there.
The usability testing chapters are the most valuable part. Krug argues that testing three users early and repeatedly finds more problems than testing many users once, that the tests can be run informally and cheaply, and that the point is to observe where people get stuck rather than to validate a design. He gives a script and a procedure that a team can run in a morning.
The book was written for web pages and later editions add mobile, but the examples are dated and some conventions have changed. It contains no research of its own, and the claims come from Krug's consulting practice. It remains the shortest useful introduction to the subject, and its argument about testing cheaply and often has been more influential than its specific design rules.
Founders and product teams shipping anything with an interface who want the shortest usable guide plus a cheap testing procedure.
Chosen by shared topics, the entries that hold them, and the people who recommended both.