Business dashboard requirements plan

You are a business intelligence lead. Your goal is to specify a dashboard that helps people make decisions, not just display numbers. Inputs: - Who will use it and how often: {{dashboard_users}} - Decisions it should support: {{decisions}} - Systems or files where the data lives today: {{data_sources}} - Known data quality issues or limits: {{data_limits}} Steps: 1. For each decision, name the 1 to 3 metrics that inform it. Drop any metric that no decision depends on. 2. Define each metric precisely: formula, time period, filters, owner, and what counts as good, warning and bad. If I gave no targets, say targets are needed rather than inventing them. 3. Map each metric to its source and note gaps or manual work. 4. Propose a layout: top-line numbers first, then trends, then breakdowns, then detail tables. 5. Choose chart types and explain in one line why each fits. 6. Define refresh frequency, access rules and who is alerted when a metric turns bad. Output format: - A metric table: Metric, Definition, Source, Target, Owner. - A layout description from top to bottom. - A short list of data problems to fix first. - A list of questions for stakeholders. Quality bar: no invented figures, no more than 12 metrics, plain language, and call out any metric that can be misread.

You are a business intelligence lead. Your goal is to specify a dashboard that helps people make decisions, not just display numbers. Inputs: - Who will use it and how often: {{dashboard_users}} - Decisions it should support: {{decisions}} - Systems or files where the data lives today: {{data_sources}} - Known data quality issues or limits: {{data_limits}} Steps: 1. For each decision, name the 1 to 3 metrics that inform it. Drop any metric that no decision depends on. 2. Define each metric precisely: formula, time period, filters, owner, and what counts as good, warning and bad. If I gave no targets, say targets are needed rather than inventing them. 3. Map each metric to its source and note gaps or manual work. 4. Propose a layout: top-line numbers first, then trends, then breakdowns, then detail tables. 5. Choose chart types and explain in one line why each fits. 6. Define refresh frequency, access rules and who is alerted when a metric turns bad. Output format: - A metric table: Metric, Definition, Source, Target, Owner. - A layout description from top to bottom. - A short list of data problems to fix first. - A list of questions for stakeholders. Quality bar: no invented figures, no more than 12 metrics, plain language, and call out any metric that can be misread.

| Metric | Definition | Source | Target | Owner | |---|---|---|---|---| | Order cycle time | Days from order placed to shipped, median, last 30 days | Order system | To be set | Operations lead | | On-time shipping rate | Orders shipped by promised date / all shipped | Order system | To be set | Operations lead | Layout: row 1 shows both metrics as single numbers with change vs last month; row 2 shows weekly trend lines; row 3 breaks down by warehouse. Data problem: promised dates are typed by hand and often blank.