Analysis Process

Two systematic frameworks for investment research

Feynman-Inspired Deep Dive
Bottom-up, disease/drug landscape analysis. Biology first, then drugs, trials, competitive landscape, valuation. The ultimate test: can you take any expert statement about this disease, fully understand every word and its implications, AND explain it to someone with no science background without losing a single nuance?
Explanation Standards (Apply to All Steps)
Three Layers — Expert, Everyday, Analogy
1. Expert: Full technical terminology, receptor biology, signaling pathways. Sufficient for real-time conversation with a KOL or scientist.

2. Everyday Language: Same depth, same nuances, same number of distinct points — but in plain words. NOT a summary. NOT analogies. If the expert layer makes 6 points, the everyday layer covers the same 6. Could someone reading ONLY this layer understand ALL implications?

3. Analogy (optional, last): Only after layers 1 and 2 are solid. Must be vivid, logical, memorable. Bad analogies are worse than none. Analogies reinforce — they never substitute.
Full Functional Treatment of Every Entity
When introducing any biological entity, cover all 11 points: (1) What is it literally? (2) What does it do? (3) Where is it? (4) How is it activated? (5) Why does it matter for this disease? (6) What happens if you block it? (7) What drugs target it? (8) What does targeting predict for clinical profile? (9) What does actual data show? (10) Is the drug oral or injectable, and why? (11) How does it compare to other targets?
The 10 Steps
Step 0
What Do We Already Know?
Review related deep dives, biology, and knowledge that already exist. What biological foundations are already laid? What diseases, targets, drugs share biology? Start from existing knowledge — never restart from zero on covered biology.
Step 1
Define the Real Question
Define the key question BEFORE any analysis. Frame at three levels: the science question, the investment question, and the elevator pitch. What sub-questions must the deep dive answer?
Step 2
Exhaustive Jargon Map
Define EVERY non-everyday term. Not just clinical headlines — all biological entities, cell types, enzymes, signaling molecules, clinical measurements. Expert definition AND everyday language for each. Living section: update as later steps introduce new terms.
Step 3a
Healthy Biology (First Principles)
Understand the biological system on its own terms BEFORE disease or drugs. Build step by step, sequentially. Three Layers + Full Functional Treatment throughout. Goal: disease mechanisms and drug targets become OBVIOUS rather than memorized.
Step 3b
Disease Biology & Target Landscape
What goes wrong in disease, traced through the healthy pathway. Map ALL druggable intervention points. For each: what would blocking address, what would it miss, predicted safety, which drugs target it, predicted vs actual clinical profile.
Step 3c
Connections & Mental Model
Map how this deep dive connects to all prior knowledge. Cross-disease biology, overlapping targets, cross-disease drug positioning. Every new deep dive should make previous deep dives richer.
Step 3d
Bottom Line Distillation
Distill 3a-3c into streamlined bottom-line that PROVES understanding. Not a shortcut — the OUTPUT of deep understanding. Everyday version must have equal depth to expert version.
Step 4
Clinical Trial Analysis
Analyze key trial data from scratch. For each trial: design, endpoints (expert + everyday), results, what numbers mean. Flag cross-trial comparison pitfalls. Apply Theory vs. Data Reconciliation — does data match Step 3b predictions? Mismatches are key insights.
Step 5
Competitive Landscape
Comprehensive map: all drugs by mechanism and format. 5 required deliverables: efficacy table, dosing/convenience table, commercial landscape, catalyst map, doctor decision framework. Market structure analysis: replacement vs split vs switch vs expansion.
Step 6
Valuation
Translate biology/trial insights into valuation. Market size, penetration, probability of success, pricing — all clearly justified. All assumptions explicit. Could someone without a finance background follow the model?
Step 7
Scenario Model
Bull/Base/Bear scenarios with probabilities. Expected value calculation. Identify the 2-3 factors that genuinely move the needle. Sensitivity analysis: what changes the thesis most?
Step 8
Downstream Impact Chain
Per-scenario downstream impacts on other companies/drugs in portfolio. Second-order effects (payer behavior, pipeline reprioritization, biosimilar dynamics). Trading playbook: if X happens, do Y.
Step 9
Post-Mortem & Calibration
Track predictions vs outcomes. When wrong: biology gap or market irrationality? Update framework based on lessons learned. Calibrate confidence levels for future predictions.
Process frameworks from Beechwood Capital investment methodology