First documented in a 1995 ATM-interface study by Hitachi Design Institute researchers Masaaki Kurosu and Kaori Kashimura. Functionally identical layouts, but the tidier-looking one was rated "easier to use," even though actual error rates didn't differ.
It cuts both ways: a good first impression makes people forgive small usability flaws early on, but if real problems pile up during use, trust collapses all at once. It's why "it looks nice" is never a substitute for actual usability testing.
Both forms below have identical fields and function. The left is unstyled, the right has tidy spacing and colour — watch the star ratings diverge.
When to use
Use this as the argument for investing in visual polish. But don't let polish become an excuse to skip real usability testing.