prototyping-when-and-what

Prototype three categories: risky assumptions you're unsure will work, disputed decisions where stakeholders disagree, and high-investment choices where being wrong is costly. Skip well-understood patterns and decisions you'd make the same way regardless of the outcome.

Match fidelity to the question: sketches or wireframes for structure, mockups for visual direction, clickable prototypes for flow testing, code prototypes for performance, and production betas for adoption at scale. Pick the cheapest fidelity that answers the question.

Stop when the prototype's question is answered, iterations show diminishing returns, or the next iteration would essentially be the production build. Define explicit "good enough to ship" criteria before starting, and ship when they're met.

Polished prototypes create unrealistic expectations. Label fidelity explicitly before demos, name what isn't built yet (e.g., placeholder visuals), and set timeline expectations from prototype to production.

Avoid prototyping without a specific question, endless iteration past diminishing returns, building prototypes to production code quality, showing polished prototypes without framing, and fidelity creep where a low-fi prototype gradually becomes as expensive as the real product.