Metrics worth building
- Lead response time: created timestamp → first qualifying activity by the assigned user
- Conversion time: date created → date closed, segmented by source and pipeline
- Source performance: volume, contact rate, viewing/meeting rate, win rate, revenue per source
- Agent conversion: normalized funnels per agent, not raw activity counts
- Lost reasons: mandatory-on-lose fields aggregated into strategy input
- Pipeline aging: time-in-stage flags for stuck deals
SQL-level patterns
Response-time and duration metrics need calculated fields: date differences between creation and first-activity timestamps, bucketed into bands (under 5 min, under 1 hour, same day, later). Define the metric precisely before writing SQL — 'response' must mean a specific, loggable event, or agents will dispute the dashboard instead of improving the number. We publish our definitions on the dashboard itself.
The second pattern is data readiness: BI Builder cannot compute what the CRM never captured. Most dashboard projects begin by making source, timestamps, and stage discipline reliable — see our data cleanup checklist.