# Designing Out Mistakes: The Shingo Approach Newsy: What is Good Design? Published: 2026-02-12T23:00:00.215Z Source: https://newsyai.com/edition/9a302678-a252-404d-8240-76d820e414d0 Archive: https://newsyai.com/newsy/5c850c31-aa90-4fca-8708-9148e1d10645.md --- # What Is Good Design? ## Today’s Lens: Shigeo Shingo — Make Errors Harmless > “Mistakes are inevitable, but the defects that arise from them are not.” — Shigeo Shingo Shingo’s deep take: good design doesn’t beg humans to be careful; it reshapes the system so ordinary slips can’t cascade into harm. This is the heart of poka‑yoke (“error‑proofing”): move quality upstream, design in constraints and confirmations, and catch issues at the source so they never become user‑visible failures. In practice, that means preferring structure over supervision—engineering flows that make the right action the easy action, and the wrong action impossible or self‑revealing. ([shingo.org](https://shingo.org/mistake-proofing-mistakes/?utm_source=openai)) ### A short expansion - Treat every frequent mistake as a design brief. Ask: what guardrail would either prevent this error or surface it immediately with a clear recovery? Build “shutdown” (hard stops) where stakes are high and “warning” (soft alerts) where learning matters. In digital products, that translates to strong defaults, typed inputs, irreversible‑action friction, and instant validation—quality you don’t “inspect in” later, but “design in” now. ([lean.org](https://www.lean.org/lexicon-terms/poka-yoke/?utm_source=openai)) ## Sources - Shingo Institute on Shingo’s maxim and the spirit of mistake‑proofing. ([shingo.org](https://shingo.org/mistake-proofing-mistakes/?utm_source=openai)) - Lean Enterprise Institute lexicon entry on poka‑yoke (origins, shutdown vs. warning). ([lean.org](https://www.lean.org/lexicon-terms/poka-yoke/?utm_source=openai)) - ProQuest article citing Shingo (1986) on preventing errors from becoming defects. ([proquest.com](https://www.proquest.com/scholarly-journals/process-improvement-poka-yoke/docview/218438294/se-2?utm_source=openai))