Insights
Policy health checks and recommendations
@guardrail-sim/insights
Policy insights, health checks, and best practice recommendations for B2B pricing governance.
Installation
Quick Start
RecommendationEngine
For more control, use the recommendation engine directly:
Insight Categories
Policy Health
Checks for structural issues with your policy:
| Check | Description |
|---|---|
noMarginFloor | Policy has no minimum margin rule |
noMaxDiscountCap | Policy has no maximum discount cap |
tooFewRules | Policy has fewer than 2 rules |
conflictingRules | Rules may contradict each other |
noPrioritySet | Rules lack priority ordering |
highMarginFloor | Margin floor may be too restrictive |
lowMarginFloor | Margin floor may be too permissive |
noVolumeConsideration | No volume-based rules |
Margin Protection
Checks for margin risk based on simulation results:
| Check | Description |
|---|---|
highApprovalRate | > 95% of discounts approved |
lowApprovalRate | < 50% of discounts approved |
marginFloorFrequentlyHit | Margin floor is the limiting factor > 30% |
averageMarginDeclining | Average margin trending down |
singleRuleDominates | One rule causing > 80% of rejections |
discountGap | Large gap between requested and max allowed |
Simulation Analysis
Insights from simulation coverage:
| Check | Description |
|---|---|
lowSimulationCoverage | < 100 scenarios tested |
segmentImbalance | Customer segments not evenly tested |
unusedRules | Rules never triggered in simulation |
volumeTierUnderutilized | Volume discounts rarely used |
highValueOrderRejection | Large orders being rejected |
noLimitingFactorVariety | Same rule always limits discounts |
Built-in Checklists
Pre-built checklists for common workflows:
Policy Setup Checklist
For new policy creation:
- Has unique policy ID
- Has descriptive name
- Has at least 2 rules
- Rules have priorities set
- Margin floor defined
- Maximum discount cap defined
Policy Review Checklist
For periodic policy audits:
- Margin floor is appropriate (10-20%)
- Discount cap is reasonable (< 50%)
- Volume tiers are utilized
- Customer segments covered
- No conflicting rules
Pre-Deployment Checklist
Before going live:
- Ran at least 100 simulations
- All rules triggered at least once
- No critical insights
- Approval rate is healthy (60-90%)
- Margin floor rarely hit (< 20%)
Insight Structure
Each insight follows this structure:
Custom Insight Packs
Create your own insight checks:
Report Structure
The analysis report includes:
Example Usage
See Also
- Getting Started - Basic setup
- Policy Concepts - Policy structure
- Examples - Real-world patterns