salesforce-pipeline-forecasting-reference.md
Salesforce Pipeline & Forecasting Reference
Opportunity Stage Architecture
Stage Design Principles
- Each stage must have a verifiable exit criterion — not "rep feels confident"
- Stages represent buyer milestones, not seller activities
- Keep to 5-8 stages
- Close probability should increase monotonically
- Consistent across teams unless truly different motions (use Record Types)
Recommended B2B SaaS Stage Model
Common SaaS defaults. Replace with your historical stage-to-close conversion rates within 90 days of implementation.
| Stage | Probability | Forecast Category | Exit Criteria |
|---|---|---|---|
| Discovery | 10% | Pipeline | Meeting completed; pain confirmed; ≥1 stakeholder identified |
| Qualification | 20% | Pipeline | Budget confirmed; decision process mapped; timeline established; champion identified |
| Solution Design | 40% | Pipeline | Technical requirements documented; demo/POC delivered; success criteria agreed |
| Proposal | 60% | Best Case | Proposal delivered; pricing discussed; no unresolved objections |
| Negotiation | 75% | Commit | Verbal agreement; contract in legal review; procurement engaged |
| Closed Won | 100% | Closed | Signed contract; PO received |
| Closed Lost | 0% | Omitted | Documented loss reason (required) |
Stage Variations by Motion
- PLG/Self-Serve: Trial → Activated → Converted → Closed Won
- Enterprise: Add "Security Review" and "Procurement" between Negotiation and Closed Won
- Channel/Partner: Add "Partner Qualified" after Discovery
- Renewal: Renewal Upcoming → Renewal Sent → Negotiation → Renewed/Churned
- Expansion: Expansion Identified → Qualification → Proposal → Negotiation → Closed Won/Lost
Validation Rules
| Rule | Logic | Purpose |
|---|---|---|
| No stage skipping | Only advance by 1 step (or Closed Lost from any) | Prevents jumping to Closed Won |
| Amount at Proposal | Stage ≥ Proposal → Amount ≠ null | Pipeline value accuracy |
| Close Date sanity | Stage < Negotiation → Close Date > TODAY + 14 | Realistic forecasting |
| Contact Role required | Stage ≥ Qualification → ≥1 OpportunityContactRole | Multi-threading documentation |
| Loss reason required | Closed Lost → Loss_Reason__c ≠ null | Win/loss analysis |
| Next Step required | Open stages → Next_Step__c ≠ null | Action orientation |
Forecast Categories
| Category | Definition | Minimum Criteria |
|---|---|---|
| Pipeline | Early-stage, not forecast-ready | Active opportunity, not yet qualified |
| Best Case | Could close this period but uncertain | Qualified, close plan exists, timeline fits |
| Commit | Will close this period | Verbal yes from economic buyer, contract in legal, no competing evaluations |
| Closed | Done | Signed contract |
| Omitted | Excluded from forecast | Closed Lost, pushed, or disqualified |
The Commit rule: If a deal has been in Commit for >30 days without advancing, it is not a Commit. Downgrade to Best Case.
Collaborative Forecasting
Configuration
| Setting | Options |
|---|---|
| Forecast type | Opportunity Amount, Opportunity Product, Opportunity Split, Product Family |
| Hierarchy | Maps to role hierarchy (Rep → Manager → Director → VP → CRO) |
| Period | Monthly or Quarterly (match business cadence) |
| Adjustments | Manager override up/down; tracked and auditable |
| Split forecasting | Each person sees their split amount, not full opportunity value |
When to Use Each Forecast Type
| Type | Best For |
|---|---|
| Opportunity Amount | Standard; single product/price per deal |
| Opportunity Product | Multiple products per deal; product-level forecast needed |
| Opportunity Split | Team selling; multiple reps share credit |
| Product Family | Forecasting by product category |
Opportunity Splits
Enable team selling attribution without double-counting.
Split types:
- Revenue Split: Must total 100% across team members; feeds individual quota attainment
- Overlay Split: Can exceed 100%; tracks influence without affecting primary quota
Common configurations:
- AE 70% + SE 30% (co-sell)
- AE 80% + SDR 20% (sourcing credit)
- AE 100% + Product Specialist 20% overlay
Pipeline Inspection Tool
Salesforce's built-in enhanced pipeline view:
Features:
- Opportunity list with Einstein Score, Activity count, Close Date changes
- Week-over-week pipeline value changes
- AI-driven insights (Einstein enabled)
- Inline editing for quick updates during reviews
Einstein Opportunity Score (paid add-on):
- 1-99 score predicting close likelihood
- High (67-99), Medium (34-66), Low (1-33)
- Based on: historical win patterns, activity, progression speed, engagement
- Updates weekly
Meeting usage:
- Filter by Commit + Best Case for forecast calls
- Sort by Einstein Score descending (focus on risk)
- Use change column to spot slipped deals
- Inline edit Close Date and Next Step live
CRM Analytics (Einstein Analytics / Tableau CRM)
When Native Reports Aren't Enough
| Need | Native Reports | CRM Analytics |
|---|---|---|
| Simple pipeline views | ✅ | Overkill |
| Cross-object joins (3+) | ❌ Limited | ✅ Full SAQL |
| Cohort analysis | ❌ Manual | ✅ Built-in |
| Trend with snapshots | ❌ | ✅ Snapshot datasets |
| AI predictions | ❌ | ✅ Einstein Discovery |
Key CRM Analytics Dashboards
- Pipeline Trending: Weekly snapshot; visualise growth, conversion, leakage
- Cohort Performance: Group deals by creation month; track conversion per cohort
- Rep Performance Matrix: Activity volume vs win rate scatter plot
- Forecast Accuracy History: Actual vs forecast by period; identify bias patterns
Forecast Accuracy
Measurement
Forecast Accuracy = 1 - |Actual - Forecast| / Target
| Accuracy | Rating | Action |
|---|---|---|
| ≥95% | Excellent | Maintain |
| 85-94% | Good | Minor calibration |
| 70-84% | Acceptable | Review commit criteria |
| 50-69% | Poor | Process issue |
| <50% | Broken | Fix data quality first |
Benchmark: Only 25% of B2B companies forecast within 5% accuracy.
Snapshot Tracking
Custom object Forecast_Snapshot__c:
- Period__c, Snapshot_Date__c, Rep__c (User lookup)
- Commit_Value__c, Best_Case_Value__c, Pipeline_Value__c
- Actual_Closed__c (populated after period ends)
- Accuracy__c (formula)
Populate via Scheduled Flow at each forecast cadence.
Pipeline Hygiene Automation
Monthly Scrub Flow (Scheduled, first Monday)
Query open Opportunities:
Flag 1: Close Date in past AND not Closed
→ Push Close Date to end of current month
→ Task: "Update close date"
Flag 2: No activity >30 days AND Stage ≥ Qualification
→ Set Pipeline_Health = "At Risk"
→ Notify owner + manager
Flag 3: Same stage >2x average days
→ Set Pipeline_Health = "Stalled"
→ Task: "Review or close"
Flag 4: Amount = 0 AND Stage ≥ Proposal
→ Notify: "Missing amount"
Pipeline Quality Score
Formula field Pipeline_Quality_Score__c (max 100):
- +20: Amount > 0
- +20: Close Date in future
- +15: ≥1 Contact Role
- +15: Next Step populated, date in future
- +10: Activity in last 14 days
- +10: Loss Reason (if Closed Lost)
- +10: All stage-required fields populated
Dashboard: Average score by team/rep. Below 60 = data quality problem.
n> For a research-backed Pipeline Quality Score based on Gong and Ebsta deal health dimensions, see the pipeline-visibility skill.