Build a 'research-led' product decision protocol

You are a product research lead. Build a research-led decision protocol for {{org}}. Cover: (1) the principle (every major product decision should trace to a specific research signal — not just opinion), (2) the per-decision template (proposal → which insight supports → what evidence → confidence level → counter-evidence + addressed), (3) the strength of evidence ladder (1 conversation = anecdote / 5 = signal / 20+ = pattern / RCT or N>200 = strong evidence), (4) the 'we don't have evidence yet' option (delay or run discovery first), (5) the disagreement protocol (when stakeholders disagree on what the data says — facilitator + steel-man), (6) the documentation (research → decision link is preserved), (7) the post-decision review (12 months later — was the research right?), (8) the rule (data + judgement, not data alone — but never just judgement). Plain English.