Asking the Right Questions

'Why are sales down?' is a business question, not a data question. The analyst's first job is translating between them.

A good data question is specific, measurable, and answerable:

• 'Why are sales down?' → 'Which regions and product lines drove the decline — and did order count or order size drop?'
• 'Is the marketing working?' → 'How does 90-day retention differ between customers acquired by channel?'
• 'Who will churn?' → 'Which behaviors in the past 30 days separate churners from stayers?'

Each rewrite names the columns, the comparison, and the cut. That's what makes it computable.