Menus are the main channel for discovering commands and learning shortcuts

What it is

The first time someone realizes “it can also do that,” and the first time they see “that chord means this,” usually happens in the few seconds a menu is open. The desktop menu bar is both a discovery channel and a shortcut primer: the name on the left, the chord on the right. It is not an accelerator list for people who already know. It is the main road by which people who do not know yet get a feature into their hands.

Discovery and learning are two stretches of the same glance, not two interfaces.

Why it happens

Pointing at a menu item costs less than remembering a chord, so exploration starts in the menu. The label on the right turns one “done from the menu” into one “saw the shortcut”—Grossman, Dragicevic, and Balakrishnan call that incidental learning of hotkeys. Without the side channel, shortcuts depend on a manual or on someone telling you, and the expert path never recruits new experts.

Desktop can afford the side channel: a menu can stay open, the pointer can rest on a row and read the chord, with no commitment to invoke. Touch and far-field have no equally stable “open a labeled catalog.” A command palette requires already being able to spell the name, so it cannot teach “what else exists that I do not know.” On desktop, dropping the menu and keeping only a palette cuts discovery and shortcut learning together. A menu that lists names without chords still discovers; the learning side channel is closed.

Studying it

The Grossman hotkey-learning experiments are the standard: offer a command via menu, shortcut, or both, and compare how many trials until people switch to the shortcut, and whether a chord printed on the menu speeds that transfer. On the discovery side, use a task that does not name the command, and see whether people open the menu and find the right item—not search time after the name is already known.

Independent variables: whether the shortcut is labeled on the item, whether a separate shortcut tutorial exists, whether the task names the command. Dependent variables: trials until first discovery, trials until migration from menu to shortcut, retention after migration.

Printing shortcuts on a welcome screen measures deliberate memorization, not incidental learning in the menu. The welcome screen goes away; the menu does not.

Where it stops holding

Experts who live in a command palette are no longer using this primer; the label confirms rather than teaches. On a touch desktop, or in an app that is search-only, the menu bar may never open; discovery moves to search suggestions and shortcut learning needs another surface. Screen-reader users can still hear the chord in a linearized menu, but that channel does not have the same cost as a glance. Gamepads and ten-foot television almost never have an equivalent “open a labeled catalog.”

Applying it

• Treat the menu as the new user’s textbook: names that can be scanned, matching shortcuts on the same row, in the platform’s own modifier symbols.
• Do not replace the menu bar with a command palette as the only discovery path. Palettes serve “I already know the name”; menus serve “I do not yet know what exists.”
• Verify with people new to the app, no mention of shortcuts, only a task. Watch whether they open the menu to finish, and whether after three to five successes they start trying the chord printed there. No chords on the menu, or a menu nobody opens, means the channel is idle.

Related

• Within the group: K2.03.1 The menu bar is the complete inventory of commands · K2.03.3 Menu items must stay in sync with actual availability
• Adjacent: C6.03 Shortcut Systems · K2.09 Menu Bar and Command Organization · K2.04 Desktop Shortcut Systems
• Search terms: incidental hotkey learning · menu discovery · accelerator label