Data prompts

Numbers into clear decisions

15 prompts · Build

• Build the dashboard that answers the question your boss always asksA dashboard designed around a specific audience and decision — with the right metrics, the right visualisations, and nothing that's just there because it's available.
• Build the forecast your team can rally aroundA forecasting model design with assumptions documented, scenarios modelled, and confidence intervals that are honest about what you don't know — not just the number people want to hear.
• Choose the right chart for the point you're actually trying to makeA data visualisation brief that matches chart type to the relationship you're showing — and avoids the common choices that look impressive but confuse the audience.
• Define metrics that actually measure what you care aboutA KPI definition framework that connects each metric to a real business question — and surfaces the leading indicators that predict outcomes before they happen.
• Design the test so the result actually tells you something usefulAn experiment design framework covering hypothesis, variables, controls, sample, and analysis plan — before you run anything and discover it was inconclusive.
• Find out if the data you're relying on is actually cleanA data quality audit that checks completeness, accuracy, consistency, and recency — before you base a decision on numbers that are quietly wrong.
• Find the pattern in who's leaving before you lose more of themA churn analysis brief that segments churned customers, identifies leading indicators, and surfaces the intervention most likely to change the outcome.
• Get clear on the question before you touch the dataA data analysis brief that forces precision about what you're trying to learn — so the analysis produces an answer, not just an interesting pattern.
• Set up the test so you can trust the resultAn A/B test design with hypothesis, sample size, success criteria, and guardrail metrics — so you learn something real instead of running an experiment you can argue either way.
• Split your customers into groups that behave differently from each otherA customer segmentation model built around the dimensions — behavioural, needs-based, or value-based — that are most useful for the decisions you're trying to make.
• Stop arguing about what's driving growth and measure itAn attribution model design that matches your business model and marketing mix — so credit is allocated by evidence, not by the channel that argues loudest.
• Track the right numbers before you drown in dashboardsAn analytics strategy that starts with the decisions you need to make and works backwards to the data and instrumentation required to make them well.
• Turn a data dump into a story with a recommendation at the endA reporting narrative framework that leads with insight, supports it with data, and ends with a clear recommendation — the structure that makes data actually change decisions.
• Understand how your customers behave over time, not just at signupA cohort analysis setup that tracks retention, engagement, and monetisation by acquisition cohort — so you can see the trajectory before the average hides it.
• Write the query without spending two hours on Stack OverflowSQL query writing assistance for the specific data retrieval, transformation, or aggregation task you're stuck on — with explanations so you understand what it does.

All prompt categories · Prompt libraries you can download