Recognition vs recall

Recognising something you're shown is far easier than recalling it from a blank memory. Good interfaces show the options instead of making you remember them.

The demo

A 20-second memory test. You will see six things for five seconds. Then we check how many you can recall from a blank page, against how many you can recognise from a line-up.

What this demo shows (text version)

You study six items briefly, then first try to list them from memory (free recall) and then pick them out of a set of twelve (recognition). Almost everyone scores higher on recognition than on recall: pulling something from a blank memory is hard, but spotting it in a line-up is easy. That gap is why usable interfaces surface options to choose from (menus, toolbars, thumbnails, autocomplete) rather than asking you to remember and type the exact thing.

You recognised more than you could recall. That gap is the whole case for menus, icons, and autocomplete over a blank command line: a good interface lets you point at the answer instead of summoning it.

This is the strongest argument against interfaces that hide everything behind a hamburger or a command palette in the name of looking clean. Hiding options does not reduce effort; it just moves the work from recognition, which is easy, to recall, which is hard.

Search is the exception that proves the rule. A blank search box is recall, but autocomplete quietly turns it back into recognition the moment you start typing. The good ones meet you halfway.