Reading Financial Statements ⑤ Four-Dimension Analysis: Growth, Profitability, Safety & Efficiency
Financial statement analysis through four dimensions: growth, profitability, safety, and efficiency, with DuPont analysis and industry-specific KPIs.
- A single metric can deceive; you only see the full picture when you cross-check four dimensions: growth, profitability, safety, and efficiency. Each dimension divides line items across the three statements, turning static figures into dynamic ones.
- DuPont analysis breaks ROE into three parts: net profit margin × asset turnover × equity multiplier. The same ROE may come from high margins (NVIDIA), high turnover (Costco), or high leverage (banks) — the underlying driver is entirely different.
- Safety asks "will it go under?": the current ratio, quick ratio, interest coverage, and operating cash flow to net income form four lines of defense.
- Revenue and EPS are not universal yardsticks. Every industry has its own key performance indicators (KPIs): retail watches same-store sales, beverages watch volume, banks watch NIM and the efficiency ratio — comparison only becomes meaningful when you use the right KPI.
- Choosing which metrics to use is itself a point of view. You must first understand a company's story before you know which yardstick to measure it with.
This is the fifth installment of the "Reading Financial Statements & the Notes — A Primer" series. In the first four pieces we dissected the three statements one by one; but being able to read a single statement is not yet analysis. Real analysis takes line items across the three statements and divides them against each other, building financial ratios that diagnose a company's health, then diagnoses from four dimensions at once. The metrics are countless as the stars; this piece hands you the most practical "four-dimension + DuPont + KPI" framework.
0. First, a method: a single number says nothing — you must compare
Is a "55% gross margin" high or low? Is a "20% ROE" good? Looking at one number alone tells you nothing — financial figures are almost always relative; they only speak once you compare them. And comparison runs in two directions, both of which you must use in any dimension of analysis:
| Direction of comparison | Compared against | Purpose | How to use it, what to look for |
|---|---|---|---|
| Vertical (across-period / trend) comparison | The company itself, across periods (quarter vs. quarter, year vs. year) |
Reveals trend and momentum: is it getting better or worse? | Pull the same line item or ratio into a multi-year series and look at the direction. If the three margins decline year after year, chase down the cause; if days inventory outstanding quietly lengthens, that is an early signal of softening demand or channel stuffing. |
| Horizontal (peer / cross-sectional) comparison | Industry competitors, at the same point in time | Reveals relative standing: strong or weak within the peer group? | Use the same yardstick (the industry's KPI) to compare against rivals. A gross margin below peers means cost control or pricing power needs work; an ROE above peers requires a DuPont breakdown to see whether it is genuine skill or just high leverage. |
You must always use both directions; drop one and you will misjudge: do only vertical comparison and the whole industry may be improving while you credit the company; do only horizontal comparison and a rival may have a one-off factor that leads you to misread the long-term trend. Vertical shows the trend, horizontal shows the standing; cross the two and the diagnosis becomes three-dimensional.
You cannot weigh an elephant against an ant — and differences in company size are no different. So in practice you first do two things:
- Common-size analysis (also called percentage statements): convert every income-statement line into a percentage of revenue, and every balance-sheet line into a percentage of total assets. That way an $8 billion-revenue company and an $800 billion one can have their structures placed on the same basis — both vertical and horizontal comparison rely on this to be fair.
- Normalize first: before comparing, correct for accounting standards (IFRS vs. US GAAP), inventory methods (LIFO/FIFO), and one-off items; otherwise you are comparing "accounting differences," not "operating differences" (a point that recurred throughout the first four pieces).
Remember the order: normalize first → then convert to common-size → only then do vertical and horizontal comparison. Skip the first two steps and any high-low you measure is likely false.
1. Growth: is the growth real? Is its source healthy?
The first growth number is year-over-year revenue growth (YoY), but the rate alone is not enough — you must ask three things:
- Are the "assumptions" behind the growth reasonable? Smartphone demand, EV penetration, AI data-center capex... the sources cited may differ, but the logic must hold up; unrealistic hand-waving is worth no more than a glance.
- Is the industry trend rising or falling? Doing the right thing in the right industry earns bonus points (e.g., AI, where cloud penetration is still low); in the wrong industry, all your effort yields half the result.
- Is the growth "volume" or "price"? Are profits keeping pace? Revenue surging while profits lag may mean revenue is being padded by undercutting on price to win orders.
Two contrasts from this series: NVIDIA grew revenue +65% in FY2026 and Data Center +68%, an extreme case of "volume and price rising together" amid the AI wave; Monster grew revenue +10.7%, but broken down that is sales volume +13.3%, unit price −1.6% — healthy volume growth, but currently lacking pricing power. Both are growth, yet the quality is worlds apart — and that difference only becomes visible once you break revenue apart.
2. Profitability: the three margins, ROE/ROA/ROIC, and the DuPont breakdown
The three margins: gross margin, operating margin, net margin
The three margins are the figures investors watch most closely (recap from Part 2): gross margin reflects "materials," operating margin reflects "labor and management," net margin reflects "charges (interest, tax, non-operating items)." Horizontal comparison (against peers) reveals relative cost control; vertical comparison (the company's own history) reveals which item is driving the profit trend. If any one of the three margins trends down, be sure to find out why.
ROE / ROA / ROIC: how much equity, assets, and invested capital each earn
| Metric | Formula | What it shows |
|---|---|---|
| ROE return on equity | Net income ÷ shareholders' equity | How much each dollar of shareholder capital earns back (Buffett's favorite) |
| ROA return on assets | Net income ÷ total assets | The efficiency of all assets (including those bought with debt) |
| ROIC return on invested capital | After-tax operating profit ÷ (equity + interest-bearing debt) | The return on the capital truly invested, after stripping out short-term operating liabilities |
ProfitVision LAB's moat filter, following growth-investing convention, sets a threshold of ROE ≥ 17%. All five of our companies clear it — but the "way" they clear it is entirely different, and that is what DuPont analysis breaks apart for you.
DuPont analysis: breaking ROE into three drivers
(profitability) × (operating efficiency) × (financial leverage)
The same high ROE can come from three completely different engines. Place NVIDIA and Costco side by side and this becomes obvious at a glance:
| Company | Net margin | Asset turnover | Equity multiplier | ≈ ROE | Driver |
|---|---|---|---|---|---|
| NVIDIA | ~55.6% | ~1.05 | ~1.3 | ~76% | Profitability-driven (pricing power) |
| Costco | ~2.9% | ~3.6 | ~2.6 | ~27% | Turnover-driven (thin margins, high volume) |
| TSMC | 45.1% | ~0.48 | ~1.5 | ~32% | High profitability, low turnover (asset-heavy) |
| JPMorgan (bank) | High | Very low | ~12 (very high leverage) | ROE 17% / ROTCE 20% | Leverage-driven (regulated) |
This table is the heart of the whole piece: Costco's net margin is only 2.9%, yet on 3.6x asset turnover it still produces a 27% ROE — this is what "thin margins, high volume" looks like inside the DuPont formula. NVIDIA is the opposite: ordinary turnover, but a 55% net margin pulls ROE straight up to 76%. A bank's high ROE comes mainly from an extremely high equity multiplier (leverage), which is exactly why analyzing a bank means looking at de-leveraged ROTCE and keeping a close eye on the CET1 capital ratio — high leverage is a double-edged sword, and SVB in Part 6 will let you watch it turn on its wielder.
DuPont also reminds us: directly pitting two companies' ROEs against each other is often a trap. The difference may come from accounting standards (IFRS revaluation inflates equity → depresses ROE; the US GAAP cost model → higher ROE), inventory methods (LIFO vs. FIFO affect net margin), or leverage structure (the equity multiplier). In CFA practice, before any peer comparison you first "normalize" these differences to a common basis. Metrics aren't for ranking high versus low; they're for decomposing the drivers and finding a company's profit engine.
3. Safety: will this company go under?
Profitability asks "does it earn?"; safety asks "will it go under?" Four lines of defense:
| Metric | Formula | Interpretation |
|---|---|---|
| Current ratio | Current assets ÷ current liabilities | > 1 is preferable; measures short-term solvency |
| Quick ratio | (Current assets − inventory − prepaids) ÷ current liabilities | Stricter; strips out the hardest-to-liquidate inventory |
| Interest coverage | EBIT ÷ interest expense | How many times over earnings can cover interest; higher is safer |
| Debt-to-equity (D/E) | Total liabilities ÷ shareholders' equity | The higher, the deeper the reliance on debt and the more volatile earnings |
Add Part 4's operating cash flow to net income ratio (≥ 0.8) and you have a complete safety check-up. One practical reminder: a current ratio below 1 is not necessarily bad — it depends on the composition of current liabilities. Some companies, for instance, pack large amounts of "deferred revenue" into current liabilities (cruise lines' prepaid tickets, Costco's prepaid membership fees); such "liabilities" are really revenue not yet recognized, not debt to be repaid, and they distort the ratio. When reading a ratio, always trace it back to the underlying line items.
There is no one universal safety threshold. Capital-intensive industries (TSMC, cruise lines, telecom) are inherently higher in debt-to-equity because they fund plants with long-term debt — this is reasonable "matching long with long." What is truly dangerous is "funding long with short" (using debt maturing within a year to build a plant that takes ten years to pay back). Monster carries almost zero debt, scoring full marks on safety but also meaning it isn't using leverage to amplify returns. Safety has no standard answer, only "suited or not to this business and this stage of life."
4. Efficiency: how fast cash and goods turn over inside the company
Efficiency analysis is Part 3's three "turnover days" (DSO, DIO, DPO) and the "cash conversion cycle" (CCC). One key point to recap: divide a static balance-sheet item by a dynamic income-statement flow and you measure operating efficiency.
Costco is the benchmark on the efficiency dimension — a negative cash conversion cycle, running the business on suppliers' money. Growing same-store sales while still keeping inventory turnover fast means it isn't padding revenue by stuffing the channel but genuinely selling goods quickly. The trend in efficiency metrics matters more than any single-period figure: days inventory outstanding suddenly lengthening, or days sales outstanding quietly rising, is often an early signal of softening demand or channel stuffing.
5. Industry KPIs: revenue and EPS are not universal yardsticks
This is the part of the whole analysis that retail investors most often overlook. Every industry has its own key performance indicators (KPIs), usually buried in the annual report and earnings calls, and they are what analysts truly watch. Only by using industry-specific KPIs for peer comparison does the comparison become reasonable and meaningful:
| Industry | Key KPI | Example from this series |
|---|---|---|
| Retail / membership | Same-store sales (SSS), membership renewal rate | Costco SSS +5.9% (ex-fuel/FX +7.6%), renewal rate 92.3%; membership fees ~51% of operating income |
| Beverages | Sales volume, depletion | Monster energy-drink shipment volume +13.3% (the main growth driver) |
| Semiconductors | Mix by process node, utilization, platform mix | TSMC advanced nodes (≤7nm) 74%, N3 24%; NVIDIA Data Center mix ~90% |
| Banks | NIM, efficiency ratio, ROTCE, CET1, net charge-off rate | JPMorgan efficiency ratio ~52% (lowest among peers), ROTCE 20%, CET1 14.5% |
| Manufacturing / defense | Backlog | Beyond revenue, undelivered orders foreshadow future revenue |
Read this table and you understand why "valuing Tesla on EPS" would be laughable — for a company in its growth phase, the market looks at penetration and unit economics, not current earnings per share. Costco's membership fees making up as much as 51% of operating income is an especially crucial KPI: it reveals Costco's true business model — the merchandise business runs at roughly breakeven to serve members, and the real profit comes from that membership card. An insight like this is something revenue and EPS alone will never show.
Metrics come in countless varieties; you cannot — and need not — compute them all. The real craft is selecting, based on the company's and the industry's characteristics, the few metrics that reveal the key story. And the selection itself signals that you already have a view on this company — you must first grasp its story before you know which yardstick to use. This is what I mean by: I teach you how to think, not just what to do.
- Segment information: a beautiful set of overall margins may be propped up by one high-margin segment. Only by breaking out segments do you see the true source of growth and profit (NVIDIA's Data Center, Monster's energy drinks vs. its declining alcohol brands).
- Non-recurring items: one-off disposal gains, impairments, and litigation settlements distort a single period's ROE / net margin. Strip them out of "sustainable earnings" before you compute a normalized metric.
- The definition basis of a KPI: does same-store sales include fuel and FX? How is the renewal rate defined? Different bases differ greatly (Costco discloses two versions of SSS). Align definitions before comparing.
- Accounting policies (standard differences): inventory method, depreciation lives, lease treatment — must-check items when normalizing for peer comparison.
Four-dimension analysis lets you read the health of a "normal" company. But for some companies, the metrics on the face of the statements will never reveal the problem — because they bury the landmines in notes no one reads. In the final installment, Tail Risk & Earnings Calls, we look at seven gory blow-up cases: how the devil hides in the details, and how to dig out of an earnings-call transcript the truths a company would rather you not know.
FAQ
What is DuPont analysis good for?
It breaks ROE into net margin, asset turnover, and the equity multiplier, letting you see the "driver" behind a high ROE: high profitability (NVIDIA), high turnover (Costco's thin-margin, high-volume model), or high leverage (banks). The same 20% ROE can sit on wildly different fundamentals — DuPont is the tool that helps you see through to that.
Why can't revenue and EPS value every company?
Because each industry's value drivers differ. Retail watches same-store sales and membership renewal, beverages watch volume, semiconductors watch process mix and utilization, banks watch NIM and capital ratios. Use the wrong yardstick and you misjudge — like "valuing a loss-making growth stock on EPS." Only industry-specific KPIs enable a meaningful peer comparison.
Is there a uniform passing standard for safety metrics?
No. Capital-intensive industries (semiconductors, cruise lines, telecom) are inherently high in debt ratio, which is reasonable "matching long with long"; what is truly dangerous is "funding long with short." A current ratio below 1 is not necessarily bad either — it depends on the composition of current liabilities (e.g., prepaid membership fees or prepaid cruise tickets are really revenue not yet recognized, not debt to be repaid). Safety must be judged together with industry characteristics.
All content is for research and educational purposes only and does not constitute investment advice. ProfitVision LAB is not a licensed investment adviser in Taiwan; all content reflects personal research and teaching based on public information, with no guarantee of accuracy or completeness. Companies and figures mentioned are teaching examples only, based on official filings, and do not constitute a buy/sell recommendation for any security. Investing involves risk; assess and bear responsibility yourself.
ProfitVision LAB · Shiba the Disciplined | I teach you how to think, not just what to do
Comments ()