marie-kondolisted

# Marie Kondo — the art of tidying interfaces Marie Kondo's insight was never "own less." It was that **you cannot decide what to keep until you can see everything you have, all at once, in one category** — and that the decision should be *what to keep*, not *what to throw away*. That maps onto UI almost exactly, and it corrects the two things that go wrong when an AI (or a person) is told to "simplify": 1. They add rather than remove. People systematically overlook subtractive solutions even when subtraction is cheaper and better (Adams et al., *Nature* 2021 — only 41% subtracted when not cued). So "improve this design" reliably produces *more*. 2. When they finally do remove, they remove the wrong things — signifiers, contrast, labels, scent — and hide the rest in a drawer. Hidden navigation halves discoverability; weak signifiers cost 22% more task time and 25% more fixations (NN/g eyetracking, n=71). This skill is the counterweight to both. Full citations: `references/research.md`. ## Why it matters (the short version of the evidence) - Visual complexity is the single strongest predictor of whether a user likes a page, and it lands within **17–50 ms**, before they read a word. Complexity + colorfulness explain **48%** of the variance in appeal after a 500 ms glance (Reinecke et al., CHI 2013). The relationship is **inverse-linear with a weak U-shaped sub-component** (Miniukovich & De Angeli, CHI 2020) — high complexity is punished hard; low