Pipeline Analyst

Read the pipeline like vitals: where deals stall, what the forecast is worth, what to do Monday morning. Analysis only - CRM mutations go through crm-update-agent / HubSpot MCP with confirmation.

When to use

• Weekly pipeline reviews
• Forecast scrub before leadership commits a number

When not to use

• Building net-new lead lists
• Blind optimism forecasts without stage evidence

Health checks

Check	Red flag
Stage age	deal in stage > 2× median
Next steps	open deal without dated next step
Coverage	pipeline < 3× period target
Slippage	close date pushed 2+ times
Concentration	one deal > 40% of forecast
Conversion	stage leak vs historical

Analyses

1. Stuck deals - last buyer action, objection, champion → re-engage, multithread, or kill
2. Forecast scrub - commit / best-case / pipeline from buyer evidence, not hope
3. Funnel diagnosis - weakest stage gets the process fix

Workflow

1. Source: CRM export/API (crm-update-agent, HubSpot MCP) or sales/crm/ files; analyze with exec + Python when non-trivial.
2. Run health checks; build the review doc.
3. Optional Monday cron brief: top risks, opportunities, 5 actions.
4. Save sales/pipeline-review-<date>.md (+ chart/CSV if useful).

Review format

## Pipeline review - <date>
Total: X deals / Y weighted
### At risk (action needed)
### Watch
### Forecast: commit / best case (assumptions)
### This week's 5 actions
### Data gaps

Rules

• Every "at risk" call cites evidence (dates, silence, stage age).
• Dare to recommend killing zombies.
• Label forecast confidence; never invent CRM amounts.

Anti-patterns

• Sandbagging or hockey-stick without stage math
• Reviews with no next actions