← Back to Ideas
Example: Microsoft Corporation (MSFT)
87
94th Percentile
Strong operational health + insider confidence + expansion momentum
🎯 What is the Company Scoring System?
A multi-factor scoring model that distills complex company data into a single, actionable score (0-100).
Combines six key categories of information:
📊 Operational Health
±20 points
Expansions, suspensions, operational efficiency from SEC events
💰 Financial Strength
±15 points
Refinancing, credit ratings, covenant violations
🚀 Strategic Momentum
±15 points
Partnerships, acquisitions, strategic initiatives
🛡️ Governance Quality
±15 points
Insider buying/selling, auditor changes, board actions
📈 Growth Trajectory
±10 points
Transformer prediction (42.8% correlation)
⚠️ Risk Indicators
±10 points
Investigations, material weaknesses, red flags
📐 The Scoring Algorithm
The score starts at 50 (neutral) and adjusts based on company events, insider activity, and predictions:
Scoring Logic Overview
def calculate_score(cik, as_of_date):
score = 50 # Start neutral
# Get 90-day event window
events_90d = get_events(cik, days=90)
# 1. Operational Health (±20 points)
expansions = count_event_type(events_90d, 'expanded')
score += min(10, expansions * 3) # +3 per expansion, max +10
suspensions = count_event_type(events_90d, 'suspended')
score -= suspensions * 8 # -8 per suspension
# 2. Financial Strength (±15 points)
if has_event(events_90d, 'refinanced'):
score += 8 # Successful refinancing = strength
if has_event(events_90d, 'covenant_violation'):
score -= 12 # Covenant violation = major red flag
if has_event(events_90d, 'upgraded'):
score += 5 # Credit rating upgrade
# 3. Strategic Momentum (±15 points)
partnerships = count_event_type(events_90d, 'partnered')
score += min(10, partnerships * 4) # +4 per partnership, max +10
acquisitions = count_event_type(events_90d, 'acquired')
score += min(8, acquisitions * 3) # +3 per acquisition, max +8
# 4. Governance Quality (±15 points)
insider_features = get_insider_features(cik)
if insider_features['cluster_buying']:
score += 8 # Multiple insiders buying = confidence
if insider_features['c_suite_buying']:
score += 6 # C-suite buying = strong signal
if has_event(events_90d, 'dismissed_auditor'):
score -= 15 # CRITICAL RED FLAG
# 5. Growth Trajectory (±10 points)
transformer_prediction = get_transformer_prediction(cik)
score += transformer_prediction * 20 # Scale to ±10 points
# e.g., +15% prediction → +3 points
# -5% prediction → -1 point
# 6. Risk Indicators (±10 points)
if has_event(events_90d, 'investigated'):
score -= 6 # Regulatory investigation
if has_event(events_90d, 'material_weakness'):
score -= 10 # Material weakness in controls
# Clamp to 0-100 range
return max(0, min(100, score))
📊 What You Get: Complete Scoring Output
Example Output: Microsoft (MSFT)
{
"cik": "0000789019",
"ticker": "MSFT",
"as_of_date": "2024-08-01",
"composite_score": 87,
"percentile": 94,
"category_scores": {
"operational_health": 92, // 5 expansions in 90 days
"financial_strength": 89, // Moody's upgrade
"strategic_momentum": 85, // 3 major partnerships
"governance_quality": 81, // C-suite buying
"growth_trajectory": 88, // +14% transformer prediction
"risk_indicators": 90 // Minor EU investigation
},
"contributing_factors": {
"positive": [
"Expansion momentum (5 events in 90 days) +8 pts",
"Strong insider buying (cluster + C-suite) +6 pts",
"Credit rating upgrade (Moody's Aa2→Aa1) +5 pts",
"Partnership acceleration (OpenAI, Meta) +4 pts",
"Transformer prediction +14.2% +3 pts"
],
"negative": [
"Regulatory investigation (EU antitrust) -4 pts"
]
},
"peer_comparison": {
"sector": "Technology",
"sector_avg_score": 72,
"outperformance": "+15 points vs sector"
},
"score_changes": {
"1_month_delta": +7,
"3_month_delta": +12,
"6_month_delta": +18
}
}
🎯 Use Cases: How to Use the Score
1. Portfolio Screening
Goal: Find high-quality investment opportunities
Query: "Show me all companies with score >80 in Technology sector"
Result: Top 94th percentile companies with strong fundamentals
2. Risk Monitoring
Goal: Detect deteriorating positions early
Alert: "Alert when any portfolio holding drops >10 points"
Action: Review position, consider reducing exposure
3. Due Diligence
Goal: Validate investment thesis
Analysis: "Score improved from 65 → 87 over 6 months"
Evidence: Expansion momentum + insider confidence + transformer prediction
4. Sector Rotation
Goal: Identify strong sectors
Comparison: "Tech avg 72, Healthcare avg 65, Finance avg 58"
Decision: Overweight Technology based on relative strength
💡 Key Scoring Insights
Score Ranges and Interpretation
| Score Range |
Percentile |
Interpretation |
Recommended Action |
| 85-100 |
Top 10% |
Exceptional quality, strong momentum |
STRONG BUY / Overweight |
| 70-84 |
Top 25% |
High quality, positive indicators |
BUY / Market weight |
| 50-69 |
Middle 50% |
Average quality, mixed signals |
HOLD / Selective |
| 30-49 |
Bottom 25% |
Below average, warning signs |
REDUCE / Underweight |
| 0-29 |
Bottom 10% |
Poor quality, red flags |
SELL / Avoid |
Score Change Velocity
Pay attention to how fast the score is changing:
- Rapid improvement (+15+ in 3 months): Turnaround story, momentum building
- Steady improvement (+5-10 in 3 months): Healthy growth trajectory
- Stable (±5): Mature company, consistent quality
- Declining (-5-10 in 3 months): Deteriorating fundamentals, review position
- Rapid decline (-15+ in 3 months): Crisis developing, consider exit
💡 Combining Score with Price Action
The most powerful signals come from divergences:
- Score rising + Price flat: Market hasn't noticed yet → BUY opportunity
- Score rising + Price rising: Momentum in sync → HOLD/ADD
- Score flat + Price rising: Price getting ahead → CAUTION
- Score falling + Price flat: Deterioration not priced in → REDUCE
- Score falling + Price falling: Avoid catching falling knife → SELL
🏆 Competitive Advantages
vs Moody's / S&P Credit Ratings
Their Approach: Manual analysis, 30-90 day lag, credit focus only
Your Advantage:
- Real-time events (minutes vs months)
- Automated extraction (no analyst lag)
- Equity-focused (not just credit)
- Incorporates transformer predictions
vs FactSet / Bloomberg Scores
Their Approach: Fundamentals-based, backward-looking
Your Advantage:
- Event-driven (forward-looking)
- Insider activity integrated
- 42.8% correlation transformer
- Transparent methodology
✅ Unique: Multi-Source Fusion
Your Secret Sauce:
- SEC events (30 types, 11.9M examples)
- Insider filings (Forms 3/4/5/13D/13F)
- Transformer predictions (42.8% correlation)
- Risk indicators (red flags, investigations)
- All fused into single score
✅ Unique: Explainable AI
You Show Exactly Why:
- "+8 pts: 5 expansions in 90 days"
- "+6 pts: C-suite buying"
- "-4 pts: EU investigation"
- Every point is accounted for
- Not a black box
💰 Pricing Strategy
| Tier |
Coverage |
Update Frequency |
Features |
Price/Month |
| Tier 1 |
Top 1,000 companies |
Monthly updates |
Basic scoring |
$20,000 |
| Tier 2 |
Russell 3000 |
Weekly updates |
+ Historical trends |
$50,000 |
| Tier 3 |
Full universe |
Daily updates |
+ API access + Alerts |
$120,000 |
| Enterprise |
Custom universe |
Real-time |
+ Custom models + Integration |
$300,000 |
🎯 Target Customers
Primary Market
- Credit Analysts (Bond Desks): Monitor credit risk across portfolio
- Portfolio Managers: Screening tool for stock selection
- Risk Management Teams: Early warning system for deteriorating holdings
- Private Equity: Due diligence scoring for potential investments
Secondary Market
- Family Offices: Simplify complex analysis for investment committee
- Wealth Managers: Client reporting (easy to explain 0-100 score)
- Corporate Development: M&A target screening
- IR Teams: Competitive intelligence (how do peers score?)
📈 Revenue Potential
Conservative Projection
150-300 customers × $50K average = $8-15M ARR
High margin (70%+), sticky product, clear differentiation
🚀 Implementation Timeline
Development Roadmap
Week 1-4: Core scoring algorithm development
- Implement 6 category scoring logic
- Integrate with existing event database
- Connect to insider features API
- Wire up transformer predictions
Week 5-6: Historical backtesting & validation
- Calculate scores for all companies (2015-2025)
- Validate score → return correlation
- Adjust weights for optimal performance
Week 7-8: API & delivery infrastructure
- Build REST API
- Create daily batch scoring pipeline
- Setup alert system for score changes
Total: 6-8 weeks from idea to beta launch
✅ Why This Product Works
Clear Value Proposition
- Simplifies complexity: 100+ data points → single 0-100 score
- Actionable: Clear interpretation (85+ = strong buy, <30 = sell)
- Real-time: Updates within minutes of new SEC filings
- Explainable: See exactly what drives each point
- Proven: Built on 42.8% correlation transformer + validated events
- Scalable: Cover entire universe (not just top 500)
This is the product Raul should get most excited about. It's practical, explainable, and immediately useful.
🎯 Next Steps to Productize
- Validate scoring weights: Backtest on historical data to ensure scores actually predict returns
- Build pilot for Raul: Score his portfolio companies, show score changes over time
- Create demo dashboard: Interactive tool to explore scores by sector, compare peers
- Prepare case studies: "Companies with score >85 outperformed by +12% vs SPY"
- Package for beta: Select 10 pilot customers, offer 3-month free trial