Cross 12% and 5.6 points go missing.
Which company changes every year 348 profitable consumer companies in five industries · 307 with two years · audit reports

01What was measured

From the audit reports of externally audited Korean companies, five industries were pulled — cosmetics, food, beverages, health functional foods and e-commerce — and for each company the operating profit line of the income statement, the operating cash flow (OCF) line of the cash-flow statement, and the working-capital movements (inventory, receivables) were read from the original documents. Twenty-seven companies whose industry code hides a different business (sales agents, multinational pharma subsidiaries, paint makers) were removed, and so was everything under 3 billion won of revenue — one entity with 0.1 billion of revenue printed an operating margin of 1,243% and a cash-flow margin of −1,122%, and that single row flipped the sign of a regression. The remaining 348 profitable companies are the population. Opening the prior year's report as well leaves 307 with two years.

The two ratios in this piece are defined as operating margin = operating profit ÷ revenue × 100 and cash-flow margin = OCF ÷ revenue × 100. Their difference is called the gap. Why these two: operating profit carries accounting judgment (depreciation, provisions, recognition timing); OCF is what actually entered and left the bank account from operations that year. Most of the difference is money tied up in working capital — sales not yet collected and stock not yet sold. The gap is computed per company and reported as a median.

Size first. Of the 348 profitable companies, 68 (20%) had negative operating cash flow in a year they booked a profit. One in five.

02It opens the moment you cross 12%

Four bands by operating margin.

In the three bands below 12%, the books and the bank account say the same number — the gap sits between −0.6 and +0.2. The moment a company crosses 12%, 5.6 points open up. That 5.6 points is larger than the entire operating margin of a company in the 0–3% band. The difference in median gap between the 12%+ group and everyone else is +6.0 points, with a bootstrap 95% interval of [+3.4, +8.5] — well clear of zero.

In the figure the distance between the two lines (2.8 points) is smaller than the gap (5.6) because they are computed differently. The lines are two band medians; the gap is the median of per-company differences. When the median of differences exceeds the difference of medians, it means the companies that open up, open up a lot — it is not many companies opening a little.

Splitting by industry gives the same picture. At 12% and above the median gap is +6.0 in health functional foods, +3.7 in food, +3.1 in beverages, +3.3 in e-commerce and +9.5 in cosmetics; below 5% all five industries sit between −4.3 and +1.4 — near zero or below it. Controlling for industry, size and growth in a regression, each point of operating margin brings 0.5 points of cash-flow margin (winsorized OLS, n=308, p<0.001). The other half is not in the bank that year. That coefficient is sensitive to extremes, though (one shell entity flipped its sign), so the spine of this piece rests on band medians and the bootstrap; the regression is a support, not a load-bearing wall.

03Four attempts to kill it

① A denominator illusion? Measured as a conversion rate (OCF ÷ operating profit), a company with more profit looks worse automatically. So it was re-measured against revenue — the gap in the table above. It opens only above 12%. Not an illusion.

② Industry mix? Re-measured inside each of the five industries. Same direction in all of them.

③ Growing pains? "Money got tied up in stock and receivables because the company was growing" is the standard explanation here. The outcome variable has to be split in two. The size of the gap is indeed widened by growth — companies growing sales by more than 20% show a median gap of +3.3 points, the other bands sit near zero. But whether cash goes negative has nothing to do with growth — 25% of shrinking companies, 21% of those growing more than 20%. In a logit regression, growth is not significant. So "fast growth widens the gap" is true and "fast growth drains cash" is false. They are different sentences. Nor is this the tautology "companies building stock are the growing ones" — of companies that built inventory the prior year, 30% shrank this year, versus 57% of those that reduced it. The builders are the growers, yes; that still has nothing to do with cash running dry.

④ A one-year snapshot? Working capital is a year-end balance and moves with a single push before closing. The prior-year reports were opened and the same thing re-measured — 307 companies. In the prior year too, the median-gap difference between the 12%+ group and the rest is +5.1 points (95% [+3.2, +6.4]). At the group level it opens in both years.

And one thing that survived the attempt to kill it. Lining up each company's two gaps side by side while doing ④, the correlation was −0.15. The companies that opened up this year are not the ones that opened up last year. That is the next section.

04A property of the group, not of the company

Companies were split at a two-year average operating margin of 12%, and for each one it was counted in how many of the two years the bank account was more than 3 points thinner than the books.

In the 12%+ group, four in five hit at least one year within two in which cash ran 3 points behind the books. But two years in a row is one in three. The two-year correlation of per-company gaps is −0.15 — knowing that a company opened up last year tells you nothing about this year. Opening up is a property of the group; which company gets that year cannot be predicted from last year's books, because the year of building stock and the year of receivables piling up rotate from company to company.

Profit-with-cash-outflow is mostly a one-year event too. Of companies with negative OCF this year, 37% were negative last year as well (base rate 15%) — more than twice the concentration, but 63% were fine a year earlier.

This is the real answer to "high-margin companies are good on cash." As a group, yes — a cash-flow margin of 14.6% beats every other band. But it runs more than 5 points behind the books every year, and a different set of companies produces that gap each year. Pick companies by margin (for credit, for trade terms, for investment) and a sizeable share of that group hits a cashless year every year, and last year's numbers do not show which ones.

05Why — advertising and other people's channels

So what makes the gap? Not growth, as §03 showed. Split by ad share.

Operating margin is the same in all four bands — every one sits around 6%. On the income statement alone, a heavy advertiser and a light one are the same company. Cash is different — companies spending 10%+ on advertising see only 58 of every 100 won of book profit reach the bank, and nearly four in ten (36%) had cash leave in a profitable year. In the regression, each point of ad share costs −0.18 points of cash-flow margin (p=0.003).

Overlay the channel and it sharpens.

Without advertising, the channel is irrelevant to cash (15% vs 17%). Advertise heavily and sell through other people's channels, and nearly half of profitable companies show a cash outflow, with a gap near 5 points. Advertising leaves the bank the day it runs; the sales it generates sit first in the settlement cycles of marketplaces and retailers (receivables) and in the pre-production for the campaign (inventory). The income statement has already counted those sales as "sold."

This is not why the 12%+ group opens up, though. The 32 companies that advertise 10%+ and sell through others have an ordinary median margin of 5.8%, and among the 83 companies above 12% this combination is 8 (10%) — the same share as overall (9%). Those 8 have the widest gap, +9.0 points, but the other 75 still open up by +4.5. Advertising and channel separate "profitable but cash went out"; the gap that bends at 12% stands without them. Why it bends at 12% is as far as public filings go — it takes order- and channel-level books to see.

Looking at where the money sits in the cash-flow statements of companies with a large gap, three in four (73%) lean to one side — inventory or receivables. The same "profitable but no cash," yet one is stock that did not sell and the other is sales not yet collected, and the remedies are opposites.

06What these numbers must not be used for

This is not causation. Heavy advertisers may simply be the businesses that carry stock by design (brand owners, manufacturers), and that is why the gap is large. The regression controlled for industry, size and growth, not for "is this a stock-carrying model." So this piece says "heavy advertisers do not turn profit into cash" (association), not "advertising eats cash," and the mechanism is claimed only as far as the ad-by-channel table supports it.

Only two years. The per-company correlation of −0.15 is measured on two years. With three or four, some companies may turn out to have it as a constitution. What can be said now is "on two years, last year does not predict this year."

The accounts scatter. Advertising is the advertising-expense line only. A company that files it under sales promotion is misclassified as a light advertiser — which biases the effect down, not up.

Some things died when the definitions were tightened. "Companies that cut advertising have the worst cash" was vivid as a narrative and vanished once margin, ad share and industry were controlled. "Margin recovered but profit fell" was an arithmetic error from not excluding prior-year loss-makers. Neither is in this piece.

FY2025, mostly separate-basis statements. Consolidated filers carry subsidiaries' inventory too. Externally audited companies only, so nothing under a few billion won of revenue. No company is named.

07So what changes

If the gap is a property of the group rather than a company's constitution, what you watch changes.

If margin is deciding something in your organization, these are the four questions.

08What this piece is actually saying

The group above a 12% operating margin runs more than 5 points thinner in cash than on paper every year, and a different set of companies produces that gap each year. One income statement shows only that a company is in that group; it does not show who gets the bad year. To see that, the cash-flow statement, the notes and the prior year's report have to be put on the same line for the same company, and then "which sale became cash" has to be split by order, channel and SKU. If Post 15 was "which order was incremental," this one is "which profit is cash."

Most of what we do looks like this. Far more often than the data being missing, the data exists but is not in a state where it can answer. Putting four statements on one line comes before the analysis, and it is usually the genuinely hard part.

Getting it into a state where it can answer is what we do.

(Note: Source — DART audit reports of externally audited Korean companies in five industries: cosmetics, food, beverages, health functional foods, e-commerce. FY2025, mostly separate-basis. After removing 27 companies whose industry code hides a different business (sales agents, multinational pharma, paint makers) and everything under 3 billion won of revenue, 348 profitable companies form the population; 307 with a prior-year report form the two-year set. Operating profit, operating cash flow and inventory/receivables movements parsed directly from the documents. Gap = operating margin − cash-flow margin, computed per company then the median. Bootstrap 4,000 draws. Regressions winsorized at 2–98%. Zero internal data.)