Beyond numbers — business quality and moats
The answer
A moat is a structural reason competitors cannot take a company's profits away. That is the standard term for a durable competitive advantage, and it is used here because everybody uses it.
A moat is a claim about the future, so it cannot be proved. It can only be tested, and the test is a number that would have fallen by now if the moat were not there.
Why this costs you money
Every expensive share is expensive because somebody believes the profits will last. The word used for that belief is "quality", and the price of quality is a high multiple.
If the belief is wrong, 2 things fall at once. Profit falls, because competitors have arrived. And the multiple falls, because the market stops paying for durability it no longer believes in.
That is the double loss. A company whose profit falls 30% and whose P/E falls from 55 to 25 has lost roughly 2 out of every 3 rupees of its value, and nothing dramatic happened. No fraud, no default, no scandal. Competitors simply turned up.
The mistake underneath it is almost always the same. Somebody sees a strong recent record — growing profit, high return on capital, a brand they recognise — and calls it a moat. Those are the results a moat produces. They are also the results of a temporary lead, a lucky decade, and a market nobody has bothered to enter yet.
A record is evidence. It is not a mechanism. Until you can name the reason a competitor cannot copy the business, you have found a company that has done well, not a moat.
How it works
There are 5 mechanisms that reliably keep competitors out. Almost every real moat is one of them, or a combination.
1. Intangible assets. Brands, patents and government licences. A brand is a moat only when it lets the company charge more than an identical unbranded product. A patent is a moat with an expiry date printed on it. A licence is a moat granted by a regulator and it can be withdrawn by the same regulator.
2. Switching costs. What the customer would lose by moving to a competitor: retraining staff, migrating data, re-certifying a component. Enterprise software and industrial components live here. This is the quietest moat, because the customer does not complain, they just stay.
3. Network effects. The product becomes more useful as more people use it. Exchanges, marketplaces and payment networks. A large user base alone is not a network effect. The question is whether user number 1,000,001 makes the product better for the existing million.
4. Cost advantage. The company produces the same thing for less, in a way competitors cannot copy: cheaper inputs from location or ownership, a process nobody else has, or scale that spreads fixed costs across more units. Cost advantages from scale are durable. Cost advantages from a temporarily cheap input are not.
5. Efficient scale. A market large enough to support 1 or 2 operators profitably and not 3. Pipelines, airports and regional infrastructure. A new entrant would destroy everybody's profits including its own, so nobody enters.
Pricing power is the test
Warren Buffett's phrasing is the most useful available: the single most important decision in evaluating a business is pricing power. If you can raise prices without losing customers to a competitor, you have a very good business.
Pricing power is what a moat produces, and unlike a moat it leaves a trace in the accounts. That is what makes it testable.
Gross margin is the cleanest place to look. It is revenue minus the direct cost of producing what was sold, divided by revenue.
Gross margin = (Revenue − Cost of goods sold) ÷ Revenue
Gross margin is not affected by most accounting choices. Depreciation policy, interest, tax and one-off gains do not change it. It answers one question: when input costs rose, could the company pass them on?
A moat that exists shows up as a gross margin that does not fall. Not a margin that rises. Holding a high margin through rising input costs and new competition is the observable footprint of pricing power.
Return on capital employed is the second place, measured over a full cycle. Broad studies of company returns find strong mean reversion: high returns decay toward average over 5 to 10 years as competitors arrive. A company that does not decay is the exception worth explaining.
Management, and what it actually means
"Good management" is usually a description of personality. The part that shows up in returns is narrower: what management does with the cash the business generates.
Four choices exist for every rupee of profit: reinvest, buy another company, pay a dividend, or buy back shares. A company earning 30% on capital should reinvest everything it can. A company earning 8% should return the money instead. Managers who reinvest at 8% because growth feels like progress destroy value every year while reporting larger revenue.
The check is available to any reader. Over 10 years, how much profit was retained, and how much did operating profit grow? A company that retained a great deal and grew very little allocated capital badly, whatever the annual report says.
What it tells you, and what it does not
A moat protects profitability. It does not create growth. A company can have an excellent moat around a market that is not getting any bigger. Pay a growth multiple for it and you will lose money while being right about the quality.
A moat does not survive a change in what the customer is buying. Every moat is built around a way of doing something. When the way changes, the moat protects a position that no longer matters.
And identifying a moat correctly does not make a share a good investment. Price still decides. A wide moat bought at 70 times earnings can produce a decade of poor returns while the business performs exactly as predicted.
Moats are also far easier to see backwards. Every case study of a great moat is written after the returns arrived. The confident moat calls that turned out wrong do not get written up.
The decision rule
Every moat claim gets a second question: what number would have already fallen if this moat were not real, and has it fallen?
If gross margin has held within a few percentage points for 5 years, including a year of rising input costs, the pricing power claim has survived a real test.
If gross margin has drifted down 1 to 2 points a year, the moat is being competed away, whatever the profit line says. Volume growth can hide margin erosion for years.
If gross margin held but return on capital fell, the company is buying its position — with advertising, discounts, or capital spending that does not earn a return. That is a position being bought again every year, not a moat.
Unless the company changed what it sells, or made a large acquisition, in which case the margin series is comparing 2 different businesses and you must rebuild it from the segment data.
Try this now
Five minutes. Do it on the holding you would most confidently describe as a quality business.
- Open the company's annual reports for the last 5 years. They are free on the company's investor relations page, and on the NSE, BSE or SEC website.
- From each year's profit and loss statement, take 2 numbers: revenue and cost of goods sold — sometimes shown as cost of materials consumed, or cost of revenue.
- For each year, calculate (revenue − cost of goods sold) ÷ revenue. Write the 5 percentages in a row, oldest first.
- Look at the shape of the row.
- Now write down, in 1 sentence, the mechanism you believe protects this company. It must be 1 of the 5 in this article: intangible asset, switching cost, network effect, cost advantage, or efficient scale. If you cannot name one, write "none".
What you should see. Three possible shapes, each meaning something different.
Flat or rising across 5 years. The pricing power is real and survived whatever those 5 years contained. If they included a year of rising input costs, this is strong evidence.
Falling steadily, 1 or 2 points a year. This is the finding worth having. Profit may still be growing, because volumes are growing, and the share price may be doing well. The margin says competitors are taking a little more every year. Companies rarely announce this. The row of 5 numbers shows it.
Jumping around with no pattern. Either the company sells commodities, where margin follows input prices and pricing power is not the right question, or the business mix changed. Neither supports a moat claim.
Then read your sentence from step 5 again. If you wrote "none", you were calling a good record a moat. That is the most common error in this subject, and 5 minutes found it.
Three real cases
1. Intel, 2018 to 2024 (United States) — a cost moat that stopped working For decades Intel's advantage was manufacturing: it made chips at smaller feature sizes than anyone, which meant better products at lower cost, funded by volumes nobody else had. That is a genuine cost advantage and it held for a very long time. From around 2018 Intel repeatedly delayed its next manufacturing processes while Taiwan Semiconductor Manufacturing Company advanced. The advantage inverted. Revenue fell from $70.8 billion in 2018 to $53.1 billion in 2024, and a net profit of $21.0 billion became a net loss of $19.2 billion. In August 2024 Intel announced 15,000 job cuts. On 8 November 2024 it was removed from the Dow Jones Industrial Average and replaced by Nvidia. The mechanism, not the brand, had been the moat. When the mechanism failed, the brand protected nothing.
2. Blockbuster, September 2010 (United States) — a moat around the wrong thing Blockbuster's advantage was physical: thousands of stores close to where people lived, and the scale to negotiate with studios. A competitor could not easily build the same store network. What it could not defend against was the customer no longer needing to travel to a store. Netflix began by posting discs and then moved to streaming. Blockbuster filed for Chapter 11 bankruptcy protection on 23 September 2010. The moat was never crossed. The market it protected disappeared.
3. Asian Paints and Birla Opus, from February 2024 (India) — a moat under live test Asian Paints has been India's largest paint company for decades, and its advantage has been widely described as distribution: relationships with a very large number of small dealers, a tinting machine in those shops, and a supply system that restocks them quickly. That is cost and scale combined, and it produced high margins and high returns on capital for a long time. In February 2024 Grasim Industries, part of the Aditya Birla group, launched a competing brand called Birla Opus, backed by a large capital investment, its own distribution, and the ability to fund losses for years. This is the rarest and most useful kind of case, because the result is not yet known. The number to watch is not market share and not the share price. It is gross margin. If the distribution moat is real, margins hold while a well-funded competitor spends. If it was mainly the absence of a serious competitor, margins fall.
The question that resolves it
A novice looks at a strong company and asks: is this a good business?
An expert looks at the same company and asks: why has nobody taken this away, and would that reason still work if a large, patient, well-funded competitor tried?
The first question is answered by the last 5 years of results. The second is answered by naming a mechanism, and most confident quality claims cannot name one.
What would make this wrong
If moats were identifiable in advance, then portfolios built by selecting wide-moat companies would reliably beat the market. Funds built on exactly this approach exist, and their records are mixed rather than decisive.
The honest limits are 3.
Hindsight is doing more work than anybody admits. Kodak, Nokia and Intel all looked like textbook moats at their peaks, and serious people described them that way at the time.
Gross margin is not meaningful for every company. Banks and financial companies have no gross margin in this sense; their equivalent is net interest margin, driven by interest rates as much as by pricing power. Service companies with few direct costs report very high gross margins that reveal little.
A stable margin can also mean a stagnant market. If no competitor arrived because the market is unattractive, the margin holds and the moat theory is untested. Stability is evidence only where somebody tried.
In India
Three moats have been unusually durable in India.
Distribution. Indian retail remains fragmented across a very large number of small outlets. Building a system that reaches them, keeps them stocked and gives them a reason to prefer your product takes decades and enormous working capital. That is much stronger here than in markets dominated by a few large retail chains.
Brand in low-cost repeat purchases. For an inexpensive product bought frequently, an Indian household often will not experiment, because the saving is small and a bad outcome is annoying.
Licences and approvals. Regulated industries carry approval-based barriers. These are real moats and also the most fragile kind, because policy changes them.
Two Indian cautions. Promoter control works in both directions. A promoter family with a long horizon can invest through a downturn in a way a professionally managed company cannot. The same control can move value out through royalty payments, related-party transactions and rent for promoter-owned assets. Read the related-party transactions note before attributing a high return to a moat.
Capital used to be scarce in India and is less scarce now. Many Indian moats were partly the result of nobody having enough money to attack them. A moat that was really a capital shortage is being tested across several Indian industries at once.
In the United States
Network effects and scale dominate. Payment networks, exchanges and marketplaces show the most durable advantages, because the mechanism strengthens itself.
Switching costs in enterprise software are formidable. A company running its accounting, payroll and inventory on one system will not move for a 20% price difference, because the cost of moving is measured in years of disruption. Gross margins in that industry are high and stable for exactly this reason.
Patents create moats with dates on them. Pharmaceutical companies earn very high margins on a protected product and lose most of that revenue within months of expiry, an event the industry calls a patent cliff. It is the clearest example of a moat with a known expiry date.
Two American cautions. Antitrust is a live risk on the strongest moats. A network effect powerful enough to be permanent attracts regulators, and US and European authorities have brought major cases against large platform companies.
Buybacks make quality look better than it is. Steady repurchases lift earnings per share and return on equity with no operating improvement. When reading a US company's 10-year record, look at revenue and operating profit before earnings per share.
Where they differ, and what that tells you
In India, the most common real moat is physical reach. In the United States, it is the network effect or the switching cost.
That difference is not cultural. It follows from the structure of each market. Indian retail is fragmented across millions of small shops, so getting a product in front of a customer is genuinely hard, and doing it well is durable. American retail is concentrated, so distribution is a negotiation with a small number of large buyers, and it is not a moat at all.
What that tells you is where to look for the mechanism. For an Indian company, ask how many outlets it reaches, how it services them, and how long a competitor would need to build the same. For an American company, ask what the customer would have to do to leave.
A second difference is about time. American moats are frequently attacked by technology, which arrives quickly. Indian moats have more often been attacked by capital, which arrives slowly and shows up first as a well-funded competitor spending on advertising and discounts. The technology attack is visible in a product. The capital attack is visible in a margin.
Carry this
- A moat is a mechanism, not a record. Name it in 1 sentence or you do not have one.
- A moat that exists shows up as a gross margin that does not fall.
- A wide moat at the wrong price still loses money. Quality is not a substitute for valuation.