Memory
What we keep
Design + The Brain · LEARN
Working memory is brutally small. Miller's famous 1956 paper put the limit near seven chunks; later work by Cowan revised it closer to four. Either way, a design that asks people to hold many things in mind at once is a design that fails. Recognition is far easier than recall — which is why visible options beat hidden commands, and why icons need labels.
Long-term memory is built through meaning, emotion, and repetition. Distinctive, coherent visual identities are memorable precisely because they reduce the number of things to remember to one strong, consistent impression. Chunking information, using familiar patterns, and offloading memory into the environment — labels, breadcrumbs, progress indicators — are all ways of respecting the brain's limits.
The practical rule: never make the user remember anything the screen could simply show.
The sources
Miller (1956) — The Magical Number Seven, Plus or Minus Two
Visit site ↗The landmark paper on the limits of working memory that still shapes how designers chunk information.
Recognition vs. recall — Nielsen Norman Group
Visit site ↗Why showing options outperforms asking users to remember them, one of the core usability heuristics.
Cowan (2001) — The magical number 4 in short-term memory
Visit site ↗A reconsideration of Miller's limit suggesting working memory holds closer to four chunks — even less room than designers assume.