How this site measures capital-light growth
Last updated October 22, 2026. Thresholds below are the ones the code uses today.
Capital-light growth asks how much new capital a business had to put in to grow its profit. Warren Buffett's example is See's Candies, which needed about $0.42 of capital for each $1 of added profit. This site measures the same ratio from each company's 10-K filings, and its default list needs profit at least doubled, a 25% or better return on capital and no more than about $3.33 of capital per $1 of new profit.
What is the formula?
- Operating profit: pre-tax profit before interest, averaged over three fiscal years at each end of the window so one odd year does not decide the answer.
- Capital needed: operating working capital plus net property, plant and equipment. The full-cost view adds goodwill and acquired intangibles.
- Profit added = profit now − profit then. Capital added = capital now − capital then.
- Capital per $1 = capital added ÷ profit added. Lower is better, and zero or less means no new capital at all.
- Incremental return = profit added ÷ capital added, as a percent: the same idea turned upside down.
The windows are 10 and 15 years, and a comparison needs at least 8 years of data.
What do the bands and reference points mean?
| Band | Incremental return | Reading |
|---|---|---|
| See's-like | 100% or more | Each new dollar of profit needed under a dollar of capital |
| FlightSafety-like | 30% to 100% | Good, not extraordinary: roughly 3 dollars of capital per dollar of new profit |
| Capital-hungry | Under 30% | Growth needed a lot of new capital, like a typical business |
The reference points are from Buffett's 2007 letter: See's needed $32 million of capital for $77 million more profit ($0.42 per $1), FlightSafety $509 million for $159 million ($3.20), and a typical growing business $400 million for the same $77 million ($5.19). Read what the See's Candies test is.
How does it become Compounders at a Fair Price?
The Compounders at a Fair Price page runs every tracked stock through four lenses and lists how many it clears:
| Lens | What it asks | Bar |
|---|---|---|
| Capital-light growth | Did profit grow on little new capital? | Profit at least doubled over 10 years, a return on capital of 25% or more and no more than $3.33 of capital per $1 of new profit |
| Profit is real | Is profit backed by cash? | No flag on the earnings quality checks |
| Cash to owners | Does it pay owners? | A net shareholder yield of 1% a year or more, 3-year average |
| Fair price | Is the price reasonable for its growth? | A PEG of 1.5 or under |
Where does the data come from, and what is left out?
Operating profit, working capital, property and equipment and goodwill come as filed from each company's 10-K through the SEC's company-facts data on SEC EDGAR. The cash and earnings-quality lenses use this site's other 10-K datasets, and P/E and PEG come from the weekly screener run.
Banks, insurers, lenders and REITs are left out because their capital and cash flow are not comparable, and companies with fewer than 8 years of filings cannot be measured. They show n/a on the lens instead of a failing mark.
What are the limits?
A balance sheet cannot show everything a business invests in. Research, brands and advertising never appear as capital, so software and consumer-brand companies look lighter than they are, and content costs at Netflix, Disney and Warner Bros. Discovery are not counted, which understates the capital they need. Cyclical industries, including energy, chemicals, metals and chips, plus machinery and car makers, depend on where in the cycle the window starts and ends, and the page says so on their rows.
Frequently asked questions
What does capital-light mean?
A business that can grow profit without putting much new capital into working capital, equipment or acquisitions. The test is how many dollars of capital it added for each dollar of profit it added.
Why a 25% return floor?
Without a floor the ratio rewards a large, low-return capital base that merely held still, such as a big oil company whose profit rebounded. FlightSafety, which Buffett calls good but not extraordinary, earned about 25% before tax.
Why use three-year averages at each end?
A single year can be distorted by a write-down or a boom. Averaging three years at each end of the window makes the comparison about the business, not about one unusual year.