Real-Time Dashboard & Reports Playbook

Introduction

Most dashboards are built for presentations, not decisions. They reflect what happened last month, not what's happening now. This playbook builds dashboards that drive action.

Section 1: Defining Metrics That Matter

• Start with decisions, not data: What decisions does this dashboard support?
• Limit to 5-7 primary KPIs per dashboard — more creates noise.
• Distinguish lagging indicators (revenue) from leading indicators (pipeline velocity).
• Every metric should have an owner and an acceptable range.

Section 2: Data Architecture for Real-Time Reporting

• Step 1: Identify source systems and refresh frequency requirements.
• Step 2: Build a centralised data layer (warehouse or lakehouse) — avoid direct tool connections.
• Step 3: Define transformation logic in one place using dbt or similar.
• Step 4: Connect BI tool (Looker, Metabase, Power BI) to the clean data layer.

Section 3: Dashboard Design Principles

• One screen, one story — don't mix operational and strategic metrics.
• Use colour consistently: green = good, red = attention needed.
• Show trend context alongside current value (vs last week, vs target).
• Build for the end user — ops dashboards look different from exec dashboards.

Key Conclusion

“A great dashboard changes behaviour. If nobody acts on what they see, the dashboard has failed — regardless of how beautiful it looks.”