A framework for designing dashboards that actually drive decisions, arguing that 80% of successful data visualization work happens before any chart is drawn. Three upstream questions are proposed: context (what question is this answering), audience (who is reading it and how literate/accountable are they), and insight (what should change once the data lands). A case study walks through building a B2B SaaS talent-management dashboard, covering the choice of competency metrics over vanity metrics like time-spent, radar charts for multi-dimensional skill data, consistent color systems, and separate mental models for individual contributors versus managers. Reports improved engagement and reduced churn after redesign, while acknowledging isolating causal impact is hard.
