The stock research checklist (with a worked example)

The case for researching a company from a checklist rather than from intuition is not a matter of taste; it is one of the better-documented effects in decision research. When a 19-item surgical checklist was introduced across eight hospitals in the landmark study by Haynes et al. (NEJM, 2009), inpatient complications fell from 11.0% to 7.0% and deaths from 1.5% to 0.8% — not because surgeons lacked knowledge, but because under time pressure even experts skip steps, and a written list makes the skipped step visible. Company research has the same failure mode. Nobody forgets that cash flow matters; people forget to check it this time, on this company, because the story arrived first and the numbers were only ever going to be confirmation. A checklist does not make anyone smarter. It makes skipping visible — and it leaves a written record of what you knew when.

What follows is the 12-point checklist we use as the manual baseline for structured company research: four blocks — business, financials, valuation context, risk — three checks each, every check phrased as a question to answer from a named, free source. Then the whole thing is run end to end on one real filing, every figure cited, so you can see what the output of a checklist pass actually looks like. One thing it is not: a decision procedure. The checklist tells you what to examine — it has nothing to say about what anyone should do afterwards, and that is a feature, not a gap.

The four blocks, and why these twelve

Block I — Business (checks 1–3)

Check 1, the revenue engine: what actually generates the revenue, in what mix? Item 1 of the 10-K plus the revenue tables answer this in one sitting — see the 30-minute 10-K method for the fast path through the filing. Check 2, concentration and dependence: named customers above 10% of revenue (a required disclosure), supplier and manufacturing dependence, and geographic mix from the segment note. Check 3, the pricing-power proxy: gross margin level and its three-year direction. Margin expanding without an acquisition or an accounting change is one of the few externally visible traces of pricing power; margin eroding is a question the MD&A must answer. All three come from the 10-K, free on EDGAR.

Block II — Financials (checks 4–6)

Check 4, growth in context: the three-year revenue trend next to the drivers management states in the MD&A — volume, price, mix, currency, acquisition. A growth number without its stated driver is not yet information. Check 5, earnings to cash: net income beside operating cash flow. A persistent gap in either direction has an accounting explanation, and the cash-flow statement names it. Check 6, balance-sheet capacity: cash and marketable securities against total debt, with one glance at the maturity schedule in the debt footnote. The question is not a ratio threshold; it is whether the balance sheet could carry a bad year without forcing decisions.

Block III — Valuation context (checks 7–9)

Check 7, per-share arithmetic: the diluted share count across three years, plus repurchases and dividends from the cash-flow statement. Shrinking share count converts flat earnings into growing earnings per share — you want to know how much of the per-share record is that conversion. Check 8, the multiple, dated: price divided by trailing diluted earnings per share, written down with the date and the price source. Not because one multiple settles anything, but because an undated multiple in your notes is worthless later. Check 9, embedded expectations: put the multiple beside the growth the company has actually delivered. The gap between the two is not an answer — it is the precise question the rest of your research has to address. This block is deliberately called valuation context: it produces comparisons and questions, never a fair-value verdict.

Block IV — Risk (checks 10–12)

Check 10, the specific risks: the non-boilerplate entries in Item 1A — the ones naming a supplier, a jurisdiction, a lawsuit — and the diff against last year's list. Check 11, accounting quality: auditor name and tenure, critical audit matters, any restatement or material-weakness language, the related-party footnote, and whether revenue-recognition policy changed. The eight-pattern grid in the 10-K guide covers where each one lives. Check 12, falsifiers: before closing the file, write down the two or three dated, checkable observations that would change the picture — the analytical pre-commitment that makes next year's review honest instead of confirmatory.

The worked example: Apple's FY2025 10-K, all twelve checks

Sources for everything below: Apple's Form 10-K for fiscal 2025 (year ended 27 September 2025), filed 31 October 2025, accession 0000320193-25-000079 on EDGAR, plus one dated market price. This is illustrative analysis of a public filing — a demonstration of the method, not an assessment of the company for any investor.

Block I. Check 1: total net sales $416.2 billion, of which iPhone $209.6 billion (50%) and Services $109.2 billion (26%) — two engines, one device-led, one margin-led. Pass: the mix is knowable and documented. Check 2: no customer above 10% of revenue, but two concentrations stand out — manufacturing (outsourcing partners concentrated in a few Asian countries, per Item 1A) and geography: Greater China revenue fell 4% to $64.4 billion after falling 8% the year before, the only segment declining two years running. Flag — a documented trend to keep examining. Check 3: total gross margin went 44.1% → 46.2% → 46.9% over three years, with services at 75.4% against products at 36.8% — the mix shift itself is the margin story. Pass.

Block II. Check 4: revenue grew 6% after 2% the prior year, and the MD&A names the drivers (Services growth from advertising, the App Store and cloud services; iPhone up 4%). Neutral: mid-single-digit growth, honestly explained. Check 5: operating cash flow of $111.5 billion against net income of $112.0 billion — a ratio of essentially 1.0, with no accrual gap to explain. Pass. Check 6: cash and marketable securities of $132.4 billion ($35.9 billion cash, $18.8 billion current and $77.7 billion non-current securities) against $90.7 billion of term debt and $8.0 billion of commercial paper — net cash of roughly $33.8 billion. Pass.

Block III. Check 7: diluted shares went 15.81 → 15.41 → 15.00 billion over three fiscal years (−5.1%), behind $89.3 billion of repurchases and $15.4 billion of dividends in fiscal 2025 alone; diluted EPS of $7.46 grew faster than net income for exactly this reason. Pass: the per-share mechanics are explicit. Check 8: at the $327.74 close of 21 July 2026, price ÷ trailing diluted EPS = 327.74 / 7.46 ≈ 44× — written down with its date, which is the entire point of the check. Neutral: a multiple is context, not a conclusion. Check 9: 44× trailing earnings sits beside 6% delivered revenue growth. That pairing is the open question of this file — what would have to be true for the arithmetic to work — and the checklist's job ends with stating it precisely. Flag — as a question, not a judgment.

Block IV. Check 10: the company-specific entries in Item 1A include single-source components, concentration of manufacturing, and App Store regulatory exposure across several jurisdictions. Flag: specific, named, trackable. Check 11: Ernst & Young, auditor since 2009; no material weakness disclosed; one critical audit matter — uncertain tax positions, with gross unrecognized tax benefits of $23.2 billion — and the tax footnote also explains why net income jumped 19.5% on 6% revenue growth (fiscal 2024 carried the EU State Aid charge; the effective tax rate fell from 24.1% to 15.6%). Pass, with the tax file noted. Check 12: three dated falsifiers recorded — Services growth slipping below 10%, Greater China extending its decline into a third year, and the resolution of the uncertain tax positions. Neutral: this check is only ever “done” or “not done”.

When the checklist says “no view”

Six passes, three neutrals, three flags — and note what the scorecard does not produce: a score, a verdict, or an instruction. That is the discipline working as designed. A checklist pass has three honest outcomes. Sometimes every question has an answer and the flags resolve into explanations — a documented understanding. Sometimes the flags stay open because public sources cannot answer them — a documented reason to stop, which is a perfectly good output. And sometimes the answers are unknowable in principle — the file then closes with “no view”, and writing those two words, dated, is more valuable than a forced opinion. The checklist's product is always the same thing: a dated record of what you examined, what you found, and what would change it.

From manual checklist to structured research

Run this checklist on three companies and a pattern emerges: blocks I, II and IV are largely extraction — the same fields, from the same filings, every time — while the judgment lives in the questions the extraction surfaces. That split is exactly how Acutic's research methodology is organised: the business, financials and risk blocks map onto the fundamentals and risk pillars of the scoring model, computed the same way for every company (the construction is documented in what a stock quality score measures), and the screening tools exist so the extraction layer runs across a whole universe instead of one file at a time. What software deliberately does not carry is checks 9 and 12 — embedded expectations and falsifiers are judgment work, and they stay with the reader. For the filing-reading layer that feeds blocks I and II, start with 10-K vs 10-Q.

Further reading: How to read a 10-K in 30 minutes — the fast path through the checklist's main source document — and 10-K vs 10-Q — which filing answers which check. Create free account.

Acutic provides investment research and educational analysis under MAR Art. 20 / § 85 WpHG. Acutic does not provide investment advice (Anlageberatung per § 1 Abs. 1a S. 2 Nr. 1a KWG / Art. 4(1)(4) MiFID II), portfolio management, or any other licensed investment service. No content in this article constitutes a personal recommendation.