Idea generation begins with a question, not a ticker
A ticker becomes an idea only when there is a plausible difference between current expectations and future evidence. The source can be a company event, industry change, market screen, policy decision, data anomaly, or observed behavior, but the output should be a testable sentence.
Sources of fundamental ideas
- earnings or guidance changes that appear temporary or misunderstood;
- divergence between a company and its peers;
- pricing, capacity, or cost changes in an industry;
- management transitions and capital-allocation shifts;
- spin-offs, restructurings, recapitalizations, and asset sales;
- customer or supplier evidence that conflicts with reported expectations;
- balance-sheet stress or refinancing needs;
- accounting changes, estimate revisions, or cash-flow divergence;
- regulation, litigation, or reimbursement decisions;
- new products, technology shifts, or channel changes.
Screening and ranking
Screens can identify unusual valuation, growth, margin, momentum, leverage, revisions, short interest, or quality. A screen is a discovery tool, not a thesis. The analyst must learn why the company appears unusual and whether the data is comparable.
Rank ideas by magnitude of possible expectation error, researchability, catalyst fit, downside, liquidity, and portfolio relevance. An intellectually interesting company can still be a poor use of time if the evidence is inaccessible or the market may not resolve the debate within the fund’s horizon.
Event and forced-attention sources
Corporate actions force investors to re-underwrite securities. Spin-offs create new shareholder bases and standalone disclosures. Index additions and deletions change flows. Refinancings expose debt-market views. Activist filings, proxy contests, tender offers, and regulatory milestones can reveal information and create deadlines.
Alternative and public web data
Web traffic, app data, pricing, job postings, product reviews, geolocation, shipping, and transaction data can generate questions. The analyst must understand rights, coverage, sample bias, revisions, timestamp, and whether the data can be obtained consistently. A visually impressive dataset isn't necessarily representative or tradable.
The idea funnel
A disciplined funnel can have four stages:
- Observation: a fact, change, or anomaly.
- Hypothesis: a one-sentence explanation and security implication.
- Quick test: the smallest amount of work needed to reject or advance it.
- Full underwriting: model, evidence, valuation, catalyst, and risk.
Maintain rejected ideas with the reason for rejection. A future event may reopen them, and repeated rejection reasons can reveal that the funnel is biased toward ideas the strategy can't use.
Avoiding idea-generation traps
Price decline alone isn't a thesis. A low multiple can reflect a deteriorating business. A high short interest can mean a crowded short or a genuinely weak company. A popular narrative can contain a real trend while the security already discounts more than the trend can deliver.
The goal isn't constant novelty. It is a repeatable process for finding questions whose answers can change a portfolio decision.