How to Analyze a Company: The Workflow Famous Investors Actually Use

Every working investor — Buffett, Lynch, Munger, Smith, Sleep, Marks — uses a different temperament and a different vocabulary, but they all converge on the same ten checks. The order they apply them in is not the order the 10-K presents them in. Survival comes first, cash comes second, and growth comes last. This guide walks through the five-step workflow that order produces, the seven famous-investor playbooks that overlay on top of it, and a 300-word worked example against Coca-Cola (NYSE: KO).

There is a folder on my desktop called Lessons Paid For. The oldest file in it is a 2023 trade that lost me about three thousand francs in my first semester at ETH because I'd looked at exactly one number — quarterly revenue growth — and ignored everything else. The balance sheet would have told me the company was funding its growth with short-term debt. The cash flow statement would have told me operating cash had been negative for six straight quarters. The income statement was the only place the story still looked clean. It was the only place I bothered to read.

I read the statements in the right order now. So does everyone whose track record I respect. The order is not an aesthetic preference — it is a sequence of survival tests, each of which has to pass before the next one matters. Growth without solvency is a footnote in someone else's case study. Margins without cash conversion are an accounting opinion. Multiples without a durable business underneath them are a sentiment indicator.

This guide is the workflow. Five steps in the order they should be executed, seven famous-investor playbooks for context, the ten-item checklist they all converge on, and a worked example using a real company we've already published a full deep dive on. Pair this piece with the companion guide on reading the balance sheet line by line, and you have the full toolkit for a single afternoon of work per company.

---

The order that actually matters

The 10-K opens with the income statement because the SEC requires it. That is the wrong order for a working investor. A company with a broken balance sheet does not deserve a margin discussion. A company with manufactured cash flow does not deserve a multiple. Read in the right sequence, the second question becomes worth asking only if the first one has been answered.

The sequence is survival → truth → growth → discipline → durability:

1. Balance sheet — can this business survive a bad year?
2. Cash flow statement — is reported profit backed by cash, or is it an estimate?
3. Income statement — is growth real, and where does it actually come from?
4. Capital allocation — is every retained dollar worth more than a dollar?
5. Qualitative layer — is the moat durable, and is management trustworthy?

The lesson is not that Wirecard was unique. The lesson is that the income statement is the one statement management has the most discretion over, and the balance sheet is the one with the least. Read the constrained one first.

---

Step 1 — The balance sheet (the survival check)

The balance sheet is a single-day photograph. It answers one question: what does this business own, what does it owe, and is the difference compounding? If you can answer that in three lines and a ratio, the rest of the analysis has somewhere to land.

The companion piece — How to Read a Balance Sheet: The Lines That Actually Matter — walks through every line item. For the workflow, five inputs do almost all of the work:

• Cash plus short-term investments versus total financial debt. A coverage ratio above ~1.0 means the company can retire all financial debt from existing cash. Below ~0.5 means the business depends on capital-markets access to roll the balance sheet. In stress windows, that access disappears first.
• Long-term debt versus annual net income. The working rule is that a quality business can extinguish long-term debt with three to four years of earnings. Above that, the business is a leveraged bet on rates and growth holding.
• Goodwill as a share of total assets. Goodwill is not productive capital. It is the cumulative premium paid above book value for acquisitions. When goodwill is over 30-40% of assets, the equity base was bought, not built. One impairment can erase years of reported earnings overnight — Kraft Heinz wrote off $15B of goodwill on a Friday morning in February 2019 and the stock lost a quarter of its value before lunch.
• Retained earnings trajectory. A multi-year compounding retained-earnings line is the cleanest statistical signature of a business that internally funds itself. A flat or shrinking line means dividends and buybacks are funded by issuance or refinancing, not by operations.
• Current ratio. Current assets over current liabilities. Above 1.5 is comfortable. Below 1.0 is a deliberate choice by mature businesses with predictable cash (utilities, consumer staples doing aggressive buybacks) or an early warning for everyone else.

A company that passes the five-line check earns the right to be discussed on growth, margin, and capital allocation. A company that fails it does not get the conversation.

Microsoft is the textbook fortress balance sheet — Terry Smith owns it in size for exactly this reason. Net cash, compounding equity base, zero issuance. The Fundsmith partnership letters describe a balance sheet like this in one phrase: the kind you can put away and not worry about.

---

Step 2 — The cash flow statement (the truth check)

The income statement is what management says happened. The cash flow statement is what the bank account shows. The gap between the two is the area where every accounting scandal of the past twenty years has lived. Read them together, in order.

The diagnostic metrics:

• FCF / Net Income ratio over a full cycle (5-7 years). A ratio above 1.0 means reported earnings are backed by cash, often more than backed. A ratio chronically below 0.8 is a working-capital problem, an aggressive accruals stance, or a CapEx-intensive business reporting more income than it can return to shareholders.
• CapEx intensity (CapEx / Revenue). Below 3% is asset-light. 3-8% is moderate. Above 10% is heavy industry, telecoms, or owned-retail. The lower the CapEx intensity, the more of operating cash converts into FCF — but a too-low number on a growing business eventually means under-investment that catches up later.
• Working-capital direction. Inventory and receivables growing faster than revenue is a yellow flag — either demand is softening and management hasn't admitted it, or the company is loosening credit to push product. Either way, the next two quarters' cash flow will tell you which.
• Share-based compensation as a percent of revenue. This is real economic dilution, even though it sits above the operating line. Above 5% for a non-tech business or above 10% for a tech business is aggressive — the share count is being quietly inflated to compensate insiders, and reported EPS overstates per-share economics.
• Cash conversion cycle. DSO + DIO − DPO. Negative is the Amazon/Costco/Apple signature — suppliers finance the business. Strongly positive and rising means the business finances itself, and the cash is locked in working capital.

A working FCF / NI test is worth more than three quarters of earnings beats. Coca-Cola's ratio has tracked close to 1.0× across the last decade — a franchise model that routes some capital through bottler investments occasionally compresses it below that line, but the absolute FCF of roughly $9–10 billion per year is what has funded 62 consecutive years of dividend increases. When FCF and net income track this closely across a full cycle, the income statement is backed by a withdrawal slip.

---

Step 3 — The income statement (the growth + margin check)

Only after the balance sheet and the cash flow statement have earned it does the income statement become useful. Now we get to ask where the growth comes from and whether the margins are durable.

The five diagnostics:

• Revenue trajectory over 10 years. Identify the CAGR. Identify inflection years (any year with >30% change in the growth rate). Find the reason in the earnings transcript or the news — was it a product launch, a price reset, an acquisition, a regulatory event? Growth without a cause is a forecast, not a fact.
• Segment-level decomposition. Total revenue is an average across products and geographies. Segment growth tells you which engine is actually compounding. A 10% headline with one segment at 35% and another at -5% is a different business than a 10% headline that is uniformly 10%.
• Margin progression. Gross, operating, EBITDA, net — 10 years. A gross-margin walk is the cleanest signal of pricing power. An operating-margin walk net of brand investment is the signal of operating leverage. A net-margin walk against an unchanged tax rate is the signal of capital structure and FX.
• Operating leverage diagnosis. Did operating income grow faster than revenue? Slower? If faster, the business has fixed-cost leverage that scales. If slower, fixed costs are growing as fast as revenue and the business is functionally a take-rate model.
• Share-count direction. Net income per share is the only earnings metric that matters at the portfolio level. A growing share count silently dilutes every other improvement. A shrinking share count adds quietly to per-share earnings every quarter. Compare share count today to share count five years ago, and again to ten years ago.

A gross margin held above 60% through input-cost spikes, a global pandemic, and two inflationary cycles is the same diagnostic signal in a different form — the company is not buying its customers with price, it is keeping them with brand and habit. That is the diagnosis you get from reading the income statement in order — and you only get to make it after the balance sheet and the cash flow statement have said the company is allowed to be interesting.

---

Step 4 — Capital allocation (the discipline check)

The first three statements describe the business as it is. Capital allocation describes what management has decided to do with the cash the business has generated. It is the most underrated diagnostic in fundamental analysis, because it is the one place management's actual judgment is the data.

Five buckets absorb every dollar of operating cash:

1. Reinvestment in the existing business — CapEx, working capital, R&D capitalised through opex.
2. Acquisitions.
3. Share buybacks.
4. Dividends.
5. Cash retained on the balance sheet for optionality.

The Buffett test for retention is unforgiving: for every dollar of cash a company retains (does not pay out), it should produce at least a dollar of incremental market value over time. If five years of retention has not produced five years of market-value compounding, the company should be paying the cash out, not keeping it. Companies that fail this test for a decade are companies whose CEO is paying themselves with shareholder money.

The buyback test is harder. A buyback is only accretive if the company buys back stock below intrinsic value. A buyback executed at the cyclical top is a destruction of capital that hides inside per-share metrics. Look at the dollar-weighted average price the company paid in buybacks over the last five years versus the current price — if the company has been buying high and is now light on cash entering a downturn, that is management malpractice dressed as discipline.

Buffett, of course, holds the master class on capital allocation. Berkshire's record is most easily read as a 60-year capital-allocation track record where the underlying businesses are almost incidental to the compounded discipline of where to put the next dollar.

---

Step 5 — The qualitative layer (the durability check)

The numbers describe what has happened. The qualitative layer describes whether what has happened is repeatable. Five questions, briefly:

• Moat. What stops a competitor from doing what this company does? Brand, switching cost, network effect, scale, regulation, patent, embedded distribution. If you cannot name one of the seven in a single sentence, the moat is "first mover" — which is a moat for about 18 months.
• Management. Tenure, equity ownership relative to compensation, track record before this job. Founders behave differently from professional managers, and the cap-table structure tells you which one you are buying.
• Customer concentration. What share of revenue comes from the top 10 customers? Above 30% is concentrated; above 50% is a single-point-of-failure business that needs a discount in valuation.
• Cyclicality. Did revenue fall in 2008, 2020, 2022? By how much? Cyclical businesses are not bad businesses, but they should not be bought as compounders.
• Regulatory exposure. Is the business regulated, lobbied against, sanctioned, or politically visible? The cost of regulation is paid in operating margin, and the timing is set by people you do not work for.

If the company passes all five steps, you are now allowed to think about price. Price is the last input in the workflow, not the first.

---

The famous playbooks

Seven investors. Seven temperaments. The same workflow, weighted differently for each.

Warren Buffett

Signature metric: Owner earnings = Net income + D&A − maintenance CapEx − any required incremental working capital. Buffett wants the cash a business throws off after every dollar of maintenance reinvestment, because that is the cash he can actually take out of the business.

Operational test: ≥15% return on equity unleveraged, consistently, for at least a decade. A high-ROE business with low debt is the structural signature of a moat that has converted into pricing power.

Famous line: "The first rule of investing is don't lose money. The second rule is don't forget the first rule." Buffett spends almost all of his analytical time on the balance sheet because the balance sheet is the document that tells you whether you can lose money permanently.

Where this slots in the workflow: Buffett's primary lens is Step 1 (balance sheet) and Step 4 (capital allocation). He is famously uninterested in next-quarter earnings, and the historical Berkshire 13-Fs show concentrated bets on businesses with decades of clean balance sheets and disciplined capital return.

Charlie Munger

Signature method: Invert. Munger reads the company and then asks the opposite question. Not "why will this work?" but "what would have to be true for this to fail?" The list of failure modes is the actual risk register; the bull case is the residual.

Operational test: A latticework of mental models — Munger's published list runs to about 100 — applied across statistics, psychology, biology, physics, and history simultaneously. The point is not memorisation; it is the discipline of attacking a company from more than one direction.

Famous line: "All I want to know is where I'm going to die, so I'll never go there." Munger wants to identify the businesses that cannot fail, then wait for them to be cheap.

Where this slots in the workflow: Munger lives in Step 5 (the qualitative layer) and Step 1 (survival). Quality over price, always — and the inverted question runs across all five steps simultaneously.

Peter Lynch

Signature metric: PEG ratio = P/E ÷ earnings growth rate. PEG below 1 means the price is paying less per unit of growth than the growth itself is worth.

Operational test: Six categories — slow-grower, stalwart, fast-grower, cyclical, turnaround, asset play. Each category is bought for a different reason and held for a different reason. Mixing categories is the most common mistake. A stalwart is not a fast-grower; a turnaround is not a slow-grower.

Famous line: "Invest in what you know." Lynch's point is not anti-intellectual — it is that an investor's edge comes from understanding a business better than the marginal seller, and the easiest way to do that is to live with the product.

Where this slots in the workflow: Lynch's lens is Step 3 (the income statement, particularly the segment decomposition) and Step 5 (qualitative, particularly customer experience). The famous One Up On Wall Street worksheets are functionally Step 3 done qualitatively.

Benjamin Graham

Signature method: Margin of safety. Compute intrinsic value conservatively. Buy at a meaningful discount to that value. The discount absorbs the analyst's mistakes, the analyst's optimism, and the future's surprises.

Operational test (deep value): Net-current-asset value — buy the company below its working capital minus all liabilities. This is the cigar butt approach Graham practised in the 1930s and 40s. It is structurally rare in 2026 markets — the universe of net-net stocks is mostly micro-cap and Asian small-cap — but the principle of price is a separate question from quality survives intact.

Famous line: "In the short run, the market is a voting machine. In the long run, it is a weighing machine."

Where this slots in the workflow: Graham is Step 1 and Step 6 (price). He is the temperamental opposite of Buffett's mature framework — Graham takes business quality as a residual; Buffett takes it as the prime input. The synthesis of both is the modern fundamental discipline.

Terry Smith

Signature filter (Fundsmith): Three rules. Buy good companies. Don't overpay. Do nothing.

A good company in Smith's framework is defined operationally:

1. ROCE > 15% consistently. Return on capital employed is the cleanest measure of how much economic profit a business generates per unit of invested capital.
2. Organic revenue growth. Growth bought through acquisition is not the same as growth produced by the underlying franchise.
3. Low capital intensity. A business that requires constant heavy reinvestment to grow has a lower terminal return than a business that compounds capital lightly.

Famous line: "We don't trade. We buy and hold." Fundsmith's portfolio turnover is in the single digits per year. The point is not laziness — it is that the compounding is in the businesses, not in the trading.

Where this slots in the workflow: Smith's lens is Step 3 (margin, growth) and Step 4 (capital allocation). The ROCE > 15% filter is a structural quality check that pre-screens almost every candidate before the deep work begins.

Nick Sleep

Signature concept: Scale economies shared. A business that grows by giving back its scale gains to the customer — in the form of lower prices, better service, or wider selection — creates a flywheel that compounds for longer than businesses that capture their scale gains as margin. Costco shares; Amazon shares; the early-era Walmart shared. Most companies do not.

Operational test: Watch the relationship between unit cost and unit price over the cycle. A business sharing scale economies sees unit price falling alongside unit cost. A business capturing scale economies sees unit price flat while unit cost falls. The first looks worse on next-quarter margin and dramatically better on 10-year compounding.

Famous line (from the Nomad partnership letters): "What we are doing is hunting for businesses that will compound at a high rate for a very long time."

Where this slots in the workflow: Sleep is the deepest version of Step 5 (qualitative). The concept cannot be reverse-engineered from financial statements alone — it requires reading the long-form annual letters, the customer-experience research, and the internal incentive design.

Howard Marks

Signature method: Second-level thinking. First-level thinking is "the company will beat earnings, so the stock goes up." Second-level thinking is "everyone expects the company to beat earnings; the price already reflects that expectation; what surprise would actually move the price?" The Marks framework runs every thesis through the question what does the consensus already believe?

Operational test: Locate the position in the cycle. Marks famously divides every cycle into rational pessimism → irrational pessimism → rational optimism → irrational optimism. Returns are made at the top of the pessimism phase and lost at the top of the optimism phase. The hard work is identifying which one you are currently in.

Famous line: "Experience is what you got when you didn't get what you wanted."

Where this slots in the workflow: Marks is the price discipline (Step 6) and the macro overlay above all five steps. He is the cleanest reminder that a great company at a terrible price is a great way to lose money.

---

The convergent checklist

Seven investors. Seven temperaments. The same ten items, independently arrived at:

1. Compounding equity base. Retained earnings have grown for the last decade. No reliance on issuance.
2. FCF / Net income ≥ 1.0 across the cycle. Reported profit is backed by cash.
3. Low net debt. Cash plus short-term investments cover financial debt, or come close.
4. No goodwill bubble. Goodwill is below ~30% of assets, or backed by a track record of accretive M&A.
5. Share count flat or down. Buybacks net of issuance and SBC, over five years.
6. At least one segment with operating leverage. Mix shift produces gross-margin walk.
7. Durable margin. Gross margin holds in a downturn; operating margin recovers within one cycle.
8. Sane capital allocation. Buybacks below intrinsic value; reinvestment at high marginal return; M&A at sane multiples or none at all.
9. Founder or management skin-in-the-game. Insider ownership relative to compensation; long tenure; equity-aligned incentives.
10. No permanent capital loss risk. No single-customer, single-supplier, single-product, or single-regulator failure mode that could erase the equity.

A company that passes eight of ten is a compounder. A company that passes ten of ten is a coffee-can stock. A company that passes fewer than six is a trade, not an investment.

---

A worked example — Coca-Cola (NYSE: KO)

Coca-Cola is the canonical slow-growth compounder — a business almost every investor has an opinion on and almost every playbook weighs differently. The compressed workflow:

• Step 1 — Balance sheet. Roughly $10B of cash against ~$35B of long-term debt — a deliberate franchise structure that services fixed coupons from predictable gross margin. Goodwill and intangible assets (brand, trademarks, bottler rights) represent the majority of the asset base, reflecting six decades of brand acquisitions. Current ratio just above 1.0. Context pass: the leverage is intentional and covered multiple times by annual operating cash; this is a utility-like structure, not a Wirecard.
• Step 2 — Cash flow. FCF/NI ratio tracking close to 1.0× across the last decade, with periodic compression when bottler investments route capital through the balance sheet. Absolute FCF of roughly $9–10B per year has funded 62 consecutive annual dividend increases. CapEx under 2% of revenue. Pass.
• Step 3 — Income statement. Revenue roughly flat to low single digits annually over the past decade — volume stagnation offset by price and mix. Gross margin ~60%. Operating margin ~27%. Share count declining for more than 15 years. This is a margin stability and capital-return story, not a growth story. Pass on margin durability; flag on volume.
• Step 4 — Capital allocation. 62-year dividend growth streak. Consistent buybacks. Buffett's retained-earnings test passes without argument — his position, held since 1988, is the most public long-term endorsement of the allocation record. Pass.
• Step 5 — Qualitative. One of the most recognised brand franchises on earth, distributed in 200+ countries, with pricing power demonstrated through every inflationary cycle of the past 60 years. Structural headwinds: health-consciousness and sugar regulation are slow-moving but real. Pass on moat; headwinds to monitor.

Run the seven playbooks against Coca-Cola and the disagreements concentrate entirely on price, not on business quality. Buffett has owned it since 1988 and calls it one of the clearest examples of durable competitive advantage he has encountered. Lynch would classify it as a stalwart — not a fast-grower, not a turnaround — and size the position accordingly: a predictable earnings machine to hold through the cycle. Smith owns it in Fundsmith: ROCE well above 15%, organic pricing power embedded in the brand, capital intensity limited to the syrup and marketing layer. Graham would run the margin-of-safety calculation on normalised earnings and find the current multiple uncomfortably above his entry threshold; he would wait. Munger would invert: what permanently destroys Coca-Cola? A global sugar prohibition, a structural water crisis across equatorial markets, or a rapid cultural shift in what the drink signals — all three are visible long in advance. Sleep would note that Coca-Cola captures its scale economies in margin rather than sharing them with the customer, the inverse of the Costco model, and it works because the customer is paying for the brand signal rather than the price point. Marks would locate the valuation in the consumer-staples cycle and note that in risk-off environments the predictability premium inflates; in risk-on environments it compresses.

All seven are reading the same business through the same workflow. They are weighting different inputs because their portfolios are sized for different risk tolerances. The workflow does not change.

---

How to actually do this in 90 minutes

1. Five minutes — pull the latest 10-K (or 20-F) and the most recent quarterly transcript.
2. Fifteen minutes — Step 1, the balance sheet. Five lines, two ratios, one pass/fail.
3. Fifteen minutes — Step 2, the cash flow statement. FCF/NI ratio over five years, CapEx intensity, working-capital direction, SBC as percent of revenue.
4. Twenty minutes — Step 3, the income statement. Revenue trajectory, segment growth, margin walk, share count.
5. Ten minutes — Step 4, capital allocation. Five buckets, the retention test.
6. Fifteen minutes — Step 5, qualitative. Moat, management, concentration, cyclicality, regulation.
7. Ten minutes — apply two of the seven playbooks. Whichever two best match the business.

If, at the end of 90 minutes, the company passes eight of the ten convergent-checklist items and trades at a price that does not require heroic assumptions, you have a candidate. Most companies fail one of the first three steps. That is fine. The workflow is designed to fail companies fast.

Know what you own.

Was this research useful?

Submitting anonymously

Leave this empty

Send