← Back to Ideas
November 2025 • Strategic Roadmap

Future Work Roadmap: V8-V11+ Development

Modular feature additions from market context to full technical analysis. Each model adds ONE complexity layer, creating clear upsell paths from base event signals to elite multi-factor alpha engines.
🗺️ Product Roadmap 📊 Modular Architecture 💰 Pricing Strategy 📖 28 min read

Core Philosophy: Keep It Simple, Stupid

Key Principles:

  1. Each model adds ONE complexity layer
  2. Train models independently to measure alpha contribution
  3. Different models for different customer segments
  4. Clear upsell path: V7 → V8 → V9 → V10 → V11
  5. All models share core architecture (signal generation + deterministic portfolio)

Strategic Opportunities: Beyond US Markets

SEDAR Canadian Market 🇨🇦 HIGH PRIORITY

Status: Blue Ocean Opportunity - Zero Competition

Quick Summary:

  • Apply V7 architecture to Canadian market (SEDAR filings)
  • Zero competition - No LLM signals for TSX/TSXV
  • Already have SEDAR data infrastructure
  • Unique value: Mining/resource event intelligence (NI 43-101, drill results)
  • Start with dual-listed companies (use US stock prices we already have)
  • $3M ARR potential (50 customers @ $5K/month)

Key Insights:

  • ✅ All SEDAR filings have English versions (no translation needed)
  • ✅ Many TSX companies dual-listed on US exchanges (~500-800 stocks, 80% of market cap)
  • ✅ 95% code reuse from V7
  • ✅ First mover advantage in untapped market

Model Version Roadmap

V7 - Event Signals ✅ CURRENT

Status: In Training

Features:

  • Past events with tiered selection (0-14d ALL, 15-60d conf≥7, 61-180d conf≥8 or IDF≥10)
  • Recency-weighted event sorting
  • Basic market regime (bull/bear/neutral)
  • Filing metadata (CIK, date, type)

Output: BUY/SELL/SKIP + conviction (0.0-1.0)

Customer Segment: Event-driven investors, catalyst traders
Pricing Strategy: Base tier - "Event-Driven Signals"

V8 - Event + Market Context Signals PLANNED

Status: Next Priority (if insider/agreement not ready)

New Features Added:

  • VIX (volatility regime at filing date)
  • Sector performance vs SPY (relative strength)
  • Market breadth (% stocks above 200-day MA)
  • Credit spreads (high-yield vs treasury spread)
  • Interest rate environment (10-year yield, Fed funds rate)
  • Market momentum (SPY 30/60/90 day returns)

Why Separate Model:

  • Some clients want pure event-driven (no market timing dependency)
  • Market context requires different data infrastructure
  • Adds cost (market data feeds)
  • Different customer need: macro-aware vs event-pure
Customer Segment: Macro-aware fundamental investors, multi-strategy funds
Pricing Strategy: Mid tier - "Event + Market Context Signals" (2x base price)

Data Sources to Build:

  • VIX historical data (Yahoo Finance, CBOE API)
  • SPY and sector ETF prices (already have infrastructure)
  • Treasury yields (FRED API - Federal Reserve Economic Data)
  • Credit spreads (HYG, TLT ETF data)
  • Market breadth (NYSE advance/decline, need new source)

V9 - Event + Fundamental Signals PLANNED

New Features Added:

  • Valuation ratios: P/E, P/B, EV/EBITDA, P/S
  • Leverage metrics: Debt/Equity, Interest Coverage, Net Debt/EBITDA
  • Growth rates: Revenue growth YoY, Earnings growth YoY
  • Profitability: Gross margin, Operating margin, ROE, ROA
  • Quality metrics: Free cash flow yield, Asset turnover
  • Relative valuation: vs sector median, vs historical average

Why Separate Model:

  • Value investors want this, growth investors don't care
  • Fundamentals require parsing financial statements (complexity)
  • Some funds use only qualitative catalysts (events) not quantitative
  • Can charge premium for fundamental analysis layer
Customer Segment: Value investors, fundamental analysts, quantitative value funds
Pricing Strategy: Premium tier - "Event + Fundamental Signals" (3x base price)

Integration Challenges:

  • XBRL parsing is complex (standardization issues)
  • Different fiscal year ends (calendar vs non-calendar)
  • TTM calculations require previous quarters
  • Some companies don't report all metrics

V10 - Event + Technical Signals PLANNED

New Features Added:

  • Momentum indicators: RSI (14), MACD, Rate of Change
  • Trend indicators: Price vs 50/200 day MA, moving average crossovers
  • Volume indicators: Volume trend, OBV, accumulation/distribution
  • Volatility indicators: Historical volatility, Bollinger Bands, ATR
  • Price action: Support/resistance breaks, 52-week highs/lows
  • Chart patterns: Breakouts, consolidations (if detectable)

Why Separate Model:

  • Technical traders are a specific customer segment
  • Requires extensive price/volume history
  • Many fundamental investors think TA is voodoo
  • Adds significant data infrastructure cost
Customer Segment: Technical traders, momentum investors, quantitative TA funds
Pricing Strategy: Premium tier - "Event + Technical Signals" (3x base price)

V11 - Elite All-Features Model FUTURE

Status: After V7-V10 proven

Features (The "Kitchen Sink"):

  • ✅ Events (V7)
  • ✅ Market context (V8)
  • ✅ Fundamentals (V9)
  • ✅ Technicals (V10)
  • ✅ Insider intelligence (V7a/V12)
  • ✅ Agreement intelligence (V7b/V13)
  • New: Alternative data, news sentiment, social media
Customer Segment: Hedge funds, institutional investors, enterprise clients
Pricing Strategy: Enterprise tier - "Complete Alpha Engine" (5x base price)

Trade-offs:

  • Higher accuracy (hopefully!)
  • Harder to explain which signals drove decision
  • More expensive to run (data costs, compute)
  • Requires all data pipelines working

Special Purpose Models

V7a - Event + Enhanced Insider Intelligence

Status: Depends on insider system completion

New Features Added:

  • Clustering detection: Multiple insiders buying same timeframe
  • Magnitude analysis: Trade size vs typical holdings
  • Role hierarchy: CEO/CFO trades weighted higher than directors
  • Timing patterns: Buys before catalysts, post-blackout buying
  • Filing velocity: Form 4 filing delay as urgency signal
  • 10% owner tracking: Activist and institutional position changes
  • Repeated buying: Same insider accumulating over time

Why Valuable:

  • Insiders know more than we do (legal front-running)
  • Clustering = high conviction across management
  • Large unusual trades = material information
  • Timing relative to events = predictive
Customer Segment: Insider trading specialists, event-driven funds
Pricing: Premium add-on (+$Y/month to base)

V7b - Event + Agreement Intelligence

Status: Depends on agreement extraction completion

New Features Added:

  • Agreement type classification: Licensing, supply, distribution, JV, M&A, employment
  • Strategic importance scoring: Exclusive rights, duration, scope
  • Revenue signals: Minimum commitments, milestone payments, royalty rates
  • Competitive moat indicators: IP licensing, long-term locks, exclusivity
  • Risk factors: Termination clauses, contingencies, penalty provisions
  • Counterparty analysis: Size/quality of partner (Fortune 500 vs startup)

Why Valuable:

  • Agreements create future revenue visibility
  • Exclusive agreements = competitive moats
  • Large commitments = material revenue impact
  • Strategic partnerships = validation
Customer Segment: Corporate development analysts, M&A-focused funds, contract intelligence users
Pricing: Premium add-on (+$Z/month to base)

Architecture Decision: Modular Add-Ons (Recommended)

The Approach:

Base: V7 - Event Signals

Add-ons customers can choose:

  • +Insider Intelligence ($+Y/month)
  • +Agreement Intelligence ($+Z/month)
  • +Market Context ($+W/month)
  • +Fundamentals ($+X/month)

Train model variants:

  • V7
  • V7 + Insider
  • V7 + Agreements
  • V7 + Insider + Agreements
  • V7 + Insider + Agreements + Market
  • etc...

Pros:

  • ✅ Maximum customer flexibility
  • ✅ Can charge for each feature
  • ✅ A/B test feature value
  • ✅ Customers pay for what they need

Cons:

  • ❌ Need to train multiple variants
  • ❌ More complex infrastructure
  • ❌ Version management complexity

Sales & Pricing Strategy

Tier 1: Base Event Signals ($X/month)

Target: 1000 customers
Pitch: "Pure event-driven alpha from SEC filings"

Tier 2: Event + Specialized Feature ($2X/month)

Options:

Target: 300 customers per variant
Pitch: "Enhanced with [insider intelligence / agreement analysis / macro awareness]"

Tier 3: Event + Multiple Features ($3X/month)

Popular combos:

Target: 100 customers
Pitch: "Multi-factor model combining [X, Y, Z]"

Tier 4: Elite Everything ($5X/month)

Features: All of the above
Target: 20 enterprise customers
Pitch: "Complete Alpha Engine - our best model"

Implementation Prioritization

Phase 1: Validate V7 (Current)

Success criteria: V7 generates alpha above benchmark

Phase 2: Choose Next Feature (Decision Tree)

If insider system is ready:

If agreement extraction is ready:

If neither ready:

Phase 3: Combine Winners

If both insider and agreements add alpha:

Phase 4: Build Out Tiers

After proving modular approach works:

  1. Build V8 (Market Context)
  2. Build V9 (Fundamentals)
  3. Build V10 (Technicals)
  4. Build V11 (Everything)

Each time:

Success Metrics

V7 (Events Only):

V7+ with Additional Features:

Product Success:

Next Actions

  1. Complete V7 training and evaluation (in progress)
  2. Measure V7 baseline performance on test set
  3. Decide which feature to add next based on:
    • What's ready (insider vs agreements vs market)
    • What's easiest (market context = easiest)
    • What adds most value (TBD from testing)
  4. Build data pipeline for chosen feature
  5. Train V7+feature model and measure improvement
  6. Iterate based on results

Strategic discussions between Kee Kimbrell and Claude Code

From Ten-K Wizard heritage to Kaleidoscope's future