Usability testing is the practice of observing real users complete tasks in a product or prototype to identify friction and failures before shipping at scale. It can be run moderated (a researcher guides the session live) or unmoderated (the user records themselves through a remote-testing platform), in person or remote, on the live product, a prototype, or a competitor's product. It is the highest-signal-per-dollar method in the user-research toolkit and the one most consistently skipped by founders who assume their product is obvious.
The canonical reference, from Jakob Nielsen's research at the Nielsen Norman Group (1990s onward): 5 users surface roughly 85 percent of usability issues on a given interface, after which r...