Information and Data Collection
Sources and their hierarchy
Research starts with a documented data trail. Distinguish primary data you generate (channel checks, store visits, expert-network calls, proprietary surveys, management meetings) from secondary data already published (SEC filings, earnings transcripts, sell-side notes, industry-association statistics, government releases). Filings to recognize: 10-K/10-Q for financials and risk factors; 8-K for material events, often the earliest public signal; DEF 14A for compensation and dilution; Form 4 for insider transactions; Schedules 13D/13G for 5% holders; 13F for institutional positions. Segment disclosures, MD&A, and the GAAP-to-non-GAAP reconciliation usually carry more information than the headline EPS.
Regulatory guardrails
Regulation FD bars issuers from selectively disclosing material nonpublic information; it does not prevent an analyst from independently developing a view. Regulation AC requires an analyst to certify that a report reflects personal views and that no part of compensation is tied to a specific recommendation, plus attestation language during public appearances. FINRA rules require a Chinese wall between research and investment banking. MNPI obtained from any source - an expert network, a supplier, a departing employee - cannot be used or shared.
Traps
- Confusing Reg FD (an issuer obligation) with the analyst's own duty.
- Using restated or "latest revised" macro series, which injects look-ahead bias; prefer point-in-time data.
- Survivorship bias when building an index or peer set from today's survivors.
- Treating non-GAAP or pro forma figures as reported without checking what was excluded.
- Assuming alternative data is clean - vendors splice, backfill, and re-base panels.
Rule of thumb
When sources conflict, prefer the one closest to the filing, and date-stamp everything: a fact is usable only if you know when it was knowable.