Most investors can tell you whether a company is growing. Far fewer can tell you whether that growth is worth anything. Return on capital employed is the number that answers the second question, and it is the single most useful figure on a report card that most people never look at.
What it actually measures
Every business needs money tied up in it to operate — factories, inventory, software, working capital. That is capital employed. ROCE asks a simple question about it: for every dollar locked up in this business, how many cents of operating profit come back each year?
A company with 25% ROCE earns 25 cents a year on every dollar invested in it. A company at 4% earns four. Both might be growing revenue. Only one is turning that growth into something worth owning.
Why growth alone tells you nothing
Consider two companies that both grow revenue 20% a year. The first does it without needing much new capital — it already has the factory, the software, the brand. The second has to buy a new plant for every increment of sales. Ten years later the first has compounded shareholder money at 20%. The second has spent a decade running hard to stand still.
This is why growth and quality are different questions. Revenue growth tells you the business is getting bigger. ROCE tells you whether getting bigger makes shareholders richer — and those two things come apart more often than most people expect:
- High growth, high ROCE — the compounder. Rare and usually expensive.
- High growth, low ROCE — growth bought with capital. Airlines, most retail expansion.
- Low growth, high ROCE — the cash cow. Boring, and frequently underpriced.
- Low growth, low ROCE — value trap territory, whatever the P/E says.
ROCE vs ROE: why we prefer the first
Return on equity answers a similar question but measures profit against shareholder equity alone, not against all the capital in the business. That difference matters because debt shrinks equity. Load a mediocre business with borrowing and its ROE climbs, not because the business got better but because the denominator got smaller.
ROCE counts debt and equity together, so leverage can't flatter it. A company can't borrow its way to a good ROCE — it has to actually earn more on the money. That's why it's the returns metric Investily scores on, and why a stock can post a respectable ROE and still lose points on the returns pillar.
The trap: a falling ROCE isn't always bad news
ROCE is a ratio, and ratios move for two reasons. Profits can fall — or capital can rise. A company that makes a large acquisition adds the entire purchase to capital employed on day one, while the earnings from it arrive over years. Its ROCE drops immediately, and nothing about the underlying business has changed.
So when you see returns fall, check which half moved. Compare the growth rate of book value against the growth rate of revenue. If capital is compounding faster than sales, the ROCE decline is a denominator story, and the question becomes whether the capital was well spent. If profits are genuinely shrinking, that's a different and more serious problem.
What counts as a good number
As a rough frame, sustained ROCE above 20% is the mark of a business with real pricing power or a structural advantage. Between 10% and 20% is a solid, ordinary company. Below 10% and you should want a specific reason — early in a build-out, recovering from a downturn, or a capital-heavy industry where that is simply the ceiling.
The word doing the work there is sustained. One good year proves nothing; a business can post a strong ROCE on a one-off gain or a cyclical peak. What you want is a decade of it, or at minimum a multi-year trend that holds. A single-year figure is a snapshot, and snapshots of ratios lie more often than almost any other number on a report card.
The bottom line
ROCE is how you tell a great business from a merely large one. It resists the leverage tricks that flatter ROE, it exposes growth that costs more than it earns, and it moves slowly enough that a good number is hard to fake for long. If you only add one metric to how you read a company, add this one — and read it across years, not quarters.
For a live example of how far apart two companies in the same industry can sit on this measure, see Nvidia against AMD, where the gap runs from 82% to under 4%.