Competitive advantage: what actually protects profits

The problem a moat solves

Start with the force that competitive advantage exists to resist.

Suppose an industry earns unusually high returns on the capital invested in it. That fact is public, because listed companies publish their results. Capital is mobile, entrepreneurs are numerous, and existing players in adjacent industries are always looking for somewhere better to deploy money. So capital flows in, capacity expands, someone competes on price to fill it, and returns fall towards the level available elsewhere.

This is not a theory anyone needs to believe on faith. It is the ordinary behaviour of markets, and it is why competition authorities spend their time analysing market power, barriers to entry and the costs customers face when switching supplierSource^source. Those are precisely the structural features that determine whether the flow of competing capital actually reaches a company's customers.

A moat is whatever interrupts that process. It is not a quality of the product. It is a feature of the situation that makes it uneconomic, slow or impossible for a competent, well-funded competitor to take the business away.

The practical test is a thought experiment. Imagine a serious rival with plenty of money, competent management and a genuine intent to take these customers. What exactly stops them, and how long would it take? If the answer is "nothing in particular, they would just have to try harder", there is no moat, however good the company is.

Six things people mistake for a moat

A better product

Products are copied. In most industries the copying takes months, not decades. A superior product earns excess returns for as long as it takes competitors to match it, which is a lead, not a barrier. Leads are valuable and they are not durable by themselves.

A large market share

Share is a result, not a cause. It becomes a cause only where scale itself lowers unit costs or improves the offering, and in many industries it does neither. A large share in a market with no scale economics and no switching costs is simply a bigger target.

Brand recognition

Recognition is not pricing power. Plenty of instantly recognisable brands compete on discount because customers know them and still buy on price. The version of brand that constitutes a moat is narrower. It is the ability to charge more than a functionally equivalent alternative and keep the customer. That shows up in sustained gross margins above the sector, not in awareness surveys.

Worth noticing here that internally generated brands are generally not recognised as assets in the financial statementsSource^source. A company that spent decades building a reputation frequently shows almost nothing for it in book value, which is one reason book-value comparisons across brand-driven and asset-driven businesses mislead so consistently.

Being first

First movers get a head start and often a costly education for their successors, who arrive later with better information and lower costs. Being first only converts into an advantage if it is used to build something structural in the meantime, such as a network, a cost position or a set of switching costs.

Excellent management

Management quality is real and it matters. It is also mobile, mortal and replaceable. An advantage that depends on a particular person is a dependency, not a barrier. The relevant question is whether the good decisions built structure that survives the decision maker.

Size without density

Scale advantages are usually local rather than global. A distribution business with 40 per cent of one city is often in a far stronger position than one with 4 per cent of a continent, because the cost advantage comes from route density, not from headcount. When someone cites scale, ask, scale relative to what geography and what fixed cost.

The sources that survive scrutiny

Cost advantage that a rival cannot replicate

Not "we are efficient", which any competitor can also become, but a structural cost position. Privileged access to a resource, a location that cannot be duplicated, a process protected by patent, or genuine scale economics in a market too small to support two players at that scale.

The test is whether an equally competent competitor could reach the same cost with the same effort. If yes, this is operational excellence rather than a moat. Operational excellence is worth having and it depreciates without constant maintenance.

Switching costs

The most underrated source, because it is undramatic. Switching costs exist when leaving is expensive in money, time, risk or disruption. Software that a company's processes are built around. A supplier whose component is certified into a customer's regulated product. A bank holding a business's payment infrastructure and payroll.

Switching costs are visible in behaviour rather than in claims. High customer retention, contracts that renew without competitive tender, and the ability to raise prices modestly without losing accounts are all observable indicators.

Network effects

The product becomes more valuable to each user as more users join. Genuine network effects are powerful and rarer than the term suggests. Three questions separate the real ones from the claimed ones.

• **Is the network local or global?** A marketplace for a service delivered in person has a network effect within each city and none between cities. Being dominant in one city gives no protection in the next, which is why such businesses must win each market separately.
• **Can users belong to several networks at once?** If customers can easily use two or three competing platforms simultaneously, the effect is much weaker than the user count implies.
• **Does the value plateau?** Beyond a certain density, additional users may add little. Once a rival reaches that threshold, the advantage of being much larger largely disappears.

Intangible assets with legal or perceptual force

Three sub-types with different durations.

Patents grant exclusive rights for a limited period in exchange for publicly disclosing the inventionSource^source. That is a moat with an expiry date printed on it, and the honest analysis asks what happens the year after.

Regulatory licences, permits and approvals can be extremely durable where the number issued is limited. This is common in banking, telecommunications, utilities and certain professional services. The corresponding risk is that the same authority that created the advantage can change it.

Brand in the narrow pricing-power sense described above, which is the slowest to build and, when genuine, among the slowest to erode.

Efficient scale

A market large enough to support one or two operators profitably and no more. A pipeline between two points, a regional airport, a specialised industrial facility serving a limited catchment. The barrier is arithmetic. A second entrant would split the volume and destroy the economics for both, so rational competitors do not enter.

The vulnerability is that the barrier depends on the market staying small. If demand grows enough to support a second operator, the protection disappears exactly when the business looks most attractive.

Testing a claimed moat with numbers

Narrative is cheap. These checks use published statements and take longer than a screen but less than an afternoon.

**Returns on capital, sustained.** Compute the return on invested capital over eight to ten years rather than one. A moat should show up as returns persistently above the cost of that capital across a full cycle. One good year is noise. A decade is evidence. Also check whether returns are rising, flat or drifting down, because a narrowing spread is the signature of a moat eroding.

**Pricing behaviour under stress.** Find a period when input costs rose. Did gross margin hold? A company that passes cost increases through without losing volume has pricing power. A company whose margin compressed absorbed the cost because it could not pass it on, whatever its brand recognition.

**Share stability, not share size.** Look at market share over five to ten years. Stable or slowly rising share in a competitive market indicates structure. Share that oscillates with promotional activity indicates a commodity relationship with customers.

**Retention and repeat behaviour.** Where disclosed, customer retention, renewal rates and revenue from existing customers are the most direct evidence of switching costs.

**Reinvestment requirements.** Compare capital expenditure with depreciation over a decade. A business that must spend heavily just to hold position is defending its ground rather than occupying protected ground, and the difference shows up in what is left for owners.

Every one of these tests looks backwards. They tell you that an advantage existed. They do not tell you it will persist, and the entire value of a moat lies in the years that have not happened yet.

Worked example, identical margins, opposite situations

Two hypothetical retailers. Both report gross margin of 34 per cent, operating margin of 7 per cent and return on invested capital of 18 per cent last year.

**Retailer A** operates in a dense cluster of neighbourhoods where it holds a large share, sharing one distribution centre across all stores. Its delivery cost per store is roughly half what a new entrant would face until that entrant built comparable density. Over the last decade its share in that cluster has moved between 31 and 35 per cent, its return on capital has stayed between 15 and 20 per cent every year, and gross margin held steady through a period of rising supplier costs.

**Retailer B** operates a similar number of stores spread thinly across a wide region, with no density anywhere. Its 18 per cent return last year followed three years at 9, 11 and 12 per cent, improving mainly because a competitor closed stores nearby. Gross margin fell by two points during the same cost inflation that Retailer A absorbed without loss.

The current-year ratios are indistinguishable. The ten-year pattern says one company has a structural cost position tied to geography and the other had a good year because a rival withdrew. When that rival's space is taken by someone else, Retailer B's numbers revert. Retailer A's do not, unless something changes the underlying density economics.

This is why single-year screening on profitability finds both companies and cannot rank them. The moat question is answered by the shape of a decade, not the level of a year.

How moats erode

Assume erosion by default and look for the mechanism. There are four.

1. **Technology changes the basis of competition.** The advantage was real and the thing it protected stopped mattering. This is the most common and the hardest to see from inside the company.
2. **Regulation changes.** Where a licence, a restriction or a standard created the protection, a rule change can remove it faster than any competitor could. This cuts both directions, and it is a specific risk to name in any research on regulated businesses.
3. **Distribution shifts.** A company whose advantage lay in controlling access to customers loses it when customers reach the category through a different channel. The product may be unchanged and the moat gone.
4. **Self-inflicted.** Management harvests the advantage. Prices are pushed until customers start looking, quality is cut to protect margin, the network is monetised until participants leave. Extracting value from a moat and maintaining it are in direct tension, and the extraction shows in the numbers first as expanding margins, which is why margin expansion is not automatically good news.

A seven-question interrogation

Ask these of any company where a competitive advantage is claimed, including by yourself.

1. **What specifically stops a well-funded competitor from taking these customers?** Answer in one concrete sentence. Vagueness here is the finding.
2. **How long would replication take, and what would it cost?** If a rival could match the position in two years for a sum they can afford, the advantage is a lead.
3. **Does the evidence appear in a decade of numbers?** Returns on capital, margin behaviour under cost pressure, and share stability.
4. **Who actually decides the purchase, and what would make them change?** Sometimes the buyer is not the user, and the relationship you assumed exists does not.
5. **Is the advantage bounded, and by what?** Geography, category, contract length, patent expiry, licence renewal. Almost every real moat has a perimeter, and knowing where it sits is more useful than believing there is none.
6. **What would erode it, and would you see that in published information before it showed up in earnings?** Name the leading indicator you would watch.
7. **Is the company currently harvesting it or reinvesting in it?** Rapid margin expansion with flat volumes and rising customer complaints is harvesting.

If most answers come from company presentations rather than from independent evidence and published statements, treat the moat as unproven rather than absent. Unproven is a legitimate research conclusion.

What a moat does not do for you

It does not tell you what the shares are worth. A durable advantage is a fact about the business; price is a separate fact, and a widely recognised moat is usually reflected in what people are willing to pay for it. Identifying a great business and buying it at any price are different activities with different outcomes.

It does not guarantee growth. Some of the most protected businesses in existence operate in markets that do not grow at all. Protection determines whether profits persist, not whether they expand.

It does not protect against bad capital allocation. A company can earn superb returns in its core operation and destroy the proceeds on acquisitions outside it. The moat protects the core; nothing protects the cash once it leaves.

It does not survive without maintenance. Networks need participants, brands need to keep their promise, cost advantages need reinvestment, licences need compliance.

And it does not make the business immune to accidents, governance failures, or a management team that decides to become something else.

Using the idea honestly

The most useful version of moat analysis is not a label applied to a company. It is a written sentence naming the specific mechanism, the evidence for it in ten years of numbers, the perimeter it operates within, and the thing that would break it.

Written that way, the analysis becomes checkable. Six months later you can return to the sentence and ask whether the evidence still holds and whether the breaking condition has moved closer. That is a considerably more valuable habit than sorting companies into moat and no-moat, because businesses do not stay in those categories, and the interesting moment is always the transition.