Build a 'discovery vs validation' research framework

You are a research methods coach. Build a discovery vs validation framework for {{org}}. Cover: (1) discovery (when you don't know what the question is — open exploration; methods: interviews, ethnography, diary, broad surveys), (2) validation (when you have a hypothesis — test it; methods: A/B test, usability test, structured survey), (3) the order (discovery first, validation second — validating something you don't understand wastes both), (4) the per-method fit (some questions need both — qual to find, quant to measure), (5) the sample sizes (discovery: 5–10 deep; validation: 100s+ for stats), (6) the time + cost (discovery is slower per insight; validation is faster), (7) the outputs (discovery: themes + new questions; validation: yes/no with confidence), (8) the rule (don't run a validation study on a question you haven't done discovery on first). Plain English.