SKILL.md
Comparable Company Analysis
Institutional-grade peer benchmarking for US stocks. Compares financial metrics (margins, growth, ROE), valuation multiples (P/E, P/S, EV/EBITDA, EV/Revenue), and provides statistical context (median, percentiles, percentile ranks). Data from SEC EDGAR (financials) and Yahoo Finance (market data).
Setup
No dependencies required. All scripts use Python standard library only.
Workflow
Step 1 — Run the comps script
bash run.sh <SYMBOL1> <SYMBOL2> <SYMBOL3> ...Max 10 symbols. Returns JSON with:
- Financial metrics: revenue, margins (gross/operating/net), ROE, YoY growth
- Valuation multiples: P/E, P/S, EV/EBITDA, EV/Revenue
- Peer statistics: min, p25, median, p75, max for each metric
- Percentile ranks: each company's rank within the peer group (0=lowest, 100=highest)
For single-company deep dive, use fundamentals instead:
bash run.sh <SYMBOL> # single company returns full metricsStep 2 — Section 1: Comparison Table
| Company | Revenue | Gross Margin | Op Margin | Net Margin | ROE | Rev Growth | P/E | EV/EBITDA |
|---|---|---|---|---|---|---|---|---|
| AAPL | $416B | 46.9% | 32.0% | 26.9% | 157% | +6.4% | 33.0x | 25.1x |
| MSFT | $282B | 68.8% | 45.6% | 36.1% | 35% | +14.9% | 27.2x | 21.5x |
Include all available multiples. If EV/EBITDA is shown, note that EBITDA is approximated as operating income (D&A not available from SEC EDGAR).
Yahoo Finance unavailable: If P/E, P/S, EV/EBITDA, and EV/Revenue are null for all companies (Yahoo Finance endpoint failed), omit those columns from the table and add a note: "Valuation multiples unavailable — Yahoo Finance data could not be retrieved. Comparison is based on SEC EDGAR fundamentals only (margins, growth, ROE)." Rank by margins and growth instead.
Step 3 — Section 2: Statistical Summary
Use the peer_statistics from script output. Present as summary rows below the table:
| Stat | Gross Margin | Op Margin | Net Margin | ROE | Rev Growth | P/E | EV/EBITDA |
|---|---|---|---|---|---|---|---|
| Max | 68.8% | 45.6% | 36.1% | 157% | +14.9% | 33.0x | 25.1x |
| Median | 57.9% | 38.8% | 31.5% | 96% | +10.7% | 30.1x | 23.3x |
| Min | 46.9% | 32.0% | 26.9% | 35% | +6.4% | 27.2x | 21.5x |
If 5+ companies, include p25 and p75 rows — with fewer than 5 data points, percentiles are not statistically meaningful and can mislead (e.g. p25 of 3 values is just near the min). Only include columns in the statistics table that have data for at least 2 companies — omit any column that is entirely null.
Step 4 — Section 3: Investment Interpretation
Use the percentile_ranks from script output. For each company, highlight where it stands relative to peers:
AAPL: Margins below median (p27), cheapest valuation (P/E p0), slowest growth (p0)
MSFT: Highest margins (p100), fastest growth (p100), richer valuation (P/E p100)Focus on outliers — metrics where a company is notably above p75 or below p25 vs peers.
Then synthesize findings:
- Premium vs. discount: which companies trade at a premium/discount to peer median on multiples, and is it justified by growth or margins?
- Best positioned: which offers the best combination of growth + margins at reasonable valuation?
- Watch out: any company where valuation is high but fundamentals are deteriorating?
Formatting Rules
- Revenue/Net Income: B or M, e.g. "$416.2B"
- EPS: plain decimal, e.g. "$7.46"
- Margins, growth, ROE: one decimal percent, e.g. "26.9%"
- Multiples: one decimal, e.g. "28.4x"
- Percentile: integer, e.g. "p75"
Limitations
- US stocks only: SEC EDGAR covers US-listed equities
- Annual data: uses 10-K filings; quarterly shifts may not be reflected
- EV/EBITDA approximation: uses operating income as EBITDA proxy (D&A not in EDGAR XBRL)
- No analyst estimates: reported actuals only, not consensus forecasts