Use a source hierarchy before analyzing the story
Screeners, charts, news, and research platforms are excellent discovery tools. They are weaker as final evidence when the question depends on a company’s reported numbers, contractual obligations, share count, risk disclosures, or accounting definitions.
A high growth rate can lead to a revenue-quality question. A low valuation multiple can lead to a balance-sheet or cyclicality question. The screen identifies where to look, not what conclusion to draw.
Build a dashboard that connects the business to the financial statements
| Layer | What to record | What can go wrong |
|---|---|---|
| Business model | Products, customers, geography, pricing, unit economics, competition | Revenue growth can hide weak economics or concentration. |
| Financial history | Revenue, margins, cash flow, capex, working capital, leverage | Earnings growth can diverge from cash generation. |
| Capital structure | Debt, cash, diluted shares, stock compensation, convertibles | Per-share value can change even when enterprise value is stable. |
| Expectations | Consensus growth, margin assumptions, guidance, valuation multiples | A good company can be a poor investment if expectations are already extreme. |
| Portfolio role | Position size, sector exposure, factor overlap, liquidity | A sound thesis can still create excessive concentration. |
Turn screening criteria into research questions
| Screen signal | Research question | Primary evidence |
|---|---|---|
| Fast revenue growth | Is growth broad, repeatable, and cash-generative? | 10-K/10-Q segment revenue, customer concentration, cash-flow statement |
| High margins | Are margins structural or temporarily boosted? | Gross margin history, operating expenses, stock compensation, restructuring |
| Low valuation multiple | What risk is the market pricing in? | Debt, cyclicality, customer loss, litigation, dilution, capital needs |
| Strong price momentum | What changed in the business versus only the share price? | Recent 8-Ks, earnings releases, guidance changes, filings |
| Attractive theme | How much of the economics actually come from the theme? | Segment revenue, end markets, capital intensity, competitive disclosures |
Do not copy a current stock list into a permanent learning article. Lists go stale quickly. The durable skill is knowing how to convert a filter into a due-diligence queue.
Use EDGAR as the backbone of the evidence trail
Search the company, confirm the filing type and period, and note amendments. A 10-K gives the annual business, risk, and audited financial picture; a 10-Q provides the continuing quarterly view; an 8-K reports certain material events more promptly. Read the filing relevant to the question rather than relying on a summary that may omit definitions or footnotes.
Confirm issuer name, ticker, filing type, filing date, and period covered.
Follow a reported metric into the financial statement, footnote, or segment disclosure that defines it.
Check prior periods and amendments so a one-quarter change is not mistaken for a durable trend.
Save the filing link, page or section, metric definition, and the assumption the evidence supports.
Write a thesis that can be tested before deciding what the stock is worth
A thesis should state what the company must achieve, why the market may be mispricing it, and what evidence would invalidate the view. Add a base case, a better case, and a worse case before attaching a valuation range.
| Thesis field | Example question |
|---|---|
| Driver | What revenue, margin, capital, or competitive variable matters most? |
| Evidence | Which filing or operating metric supports the view? |
| Expectation gap | What does the current price appear to assume? |
| Failure condition | What specific evidence would make the investor changes the investor's mind? |
| Review trigger | Which filing, event, or threshold deserves the next review? |
Move from setup to the company itself
Build the research file from primary evidence outward
Start with the issuer and filing record, then use secondary data to speed comparison rather than to replace the original evidence. A research note is stronger when every major claim can be traced to a filing, financial statement, company disclosure, or a clearly identified calculation.
- Define the question.Growth quality, margins, balance-sheet risk, valuation, or another testable issue.
- Collect the filings.Use the latest annual report, recent quarterly filings, material current reports, and proxy information relevant to the question.
- Build a source table.For every important number, record period, unit, source document, and whether it is reported, adjusted, or calculated.
- Write the counter-case.Record what evidence would make the thesis weaker instead of collecting only confirming facts.