CompanyGraph
An investor does not need to understand every business. The harder requirement is to know which parts of a business can be explained, tested, and updated—and which parts are being filled with confidence rather than evidence.
The circle is a boundary around a decision process
Warren Buffett popularized the phrase in Berkshire's shareholder letters as a way to describe businesses that an investor can understand well enough to evaluate. His 1999 letter says that the size of the circle matters less than knowing where its boundaries lie (Berkshire's 1999 letter). The idea is not an academic measurement scale. It is an investing rule about where a person can connect industry conditions, operations, economics, management decisions, and valuation without relying on unsupported assumptions.
A circle can be narrow and still useful. It can also expand through work: reading filings, visiting operations, testing a product, learning an industry, and receiving feedback from earlier judgments. The boundary should move when knowledge improves or when the business changes, not when an investor wants a larger opportunity set.
Familiarity is not competence
| What feels familiar | What it may actually provide | What still needs testing |
|---|---|---|
| Using the product | Experience of one user, price, and context | Unit economics, customer mix, quality, and competitive response |
| Recognizing the industry | Vocabulary and a rough mental model | Process details, contracts, regulation, and cash timing |
| Reading many reports | More recorded information | The ability to distinguish observation from management interpretation |
| Knowing a management team | Access, reputation, or narrative consistency | Whether decisions produce the promised operating result |
Competence is therefore specific to a question. Someone may understand a utility's regulated revenue but not its fuel hedge, or a software product but not its cloud-cost and renewal economics. “Technology” or “consumer” is usually too broad a boundary; the investor needs to identify the mechanism and the conditions that make it work.
Overconfidence can hide the boundary
Kruger and Dunning's experiments found that participants with weaker performance on several tasks were often less accurate in assessing their own performance (the 1999 study). The experiments concern grammar, logic, and humor tasks, not stock-picking. They support a general metacognitive warning: a lack of skill can also reduce the ability to notice one's own errors, but they do not provide a way to measure an investor's circle or prove that every novice is overconfident.
Financial markets add another difficulty. Prices, narratives, and analyst coverage can make an unfamiliar industry sound understandable before the investor has observed its decisive operating variables. The appropriate response is not to avoid all uncertainty; it is to label what is known, what is inferred, what is delegated to another expert, and what position size the uncertainty permits.
Berkshire illustrates selective competence, not universal expertise
Berkshire's 1999 letter explains that the company and its managers prefer businesses they can understand and warns against claiming predictive skill in industries they do not understand. The same letter describes the need to decide where to deploy excess funds and when a business should retain or return them (the full letter).
This is evidence of a stated decision policy, not proof that every Berkshire investment lies inside one objective circle or that the policy caused its returns. Berkshire also uses specialists in insurance, rail, energy, manufacturing, and investments. The case shows how competence can be organizational and delegated, provided the decision-maker understands what the specialist knows, what remains uncertain, and how the result will be monitored.
The boundary should include what you cannot observe
- Private contracts, customer churn, product defects, and supplier dependence may be decisive but only partially disclosed.
- Accounting numbers can be precise while the operational mechanism behind them remains uncertain.
- A past success may reflect a cycle, luck, or a one-time opportunity rather than repeatable expertise.
- Management may understand the business but face incentives or financing constraints that change what it can do.
- Delegating research does not transfer responsibility for deciding how much uncertainty the portfolio can bear.
A useful circle therefore has an outer boundary and a confidence gradient. The investor may understand the core economics, have partial knowledge of a new geography, and know almost nothing about a regulatory edge case. That does not require a binary “inside” or “outside” decision; it requires an explicit reduction in claim strength, time spent, or position size.
How to build and audit a circle
- Write the business mechanism in ordinary language before reading the valuation case.
- List the variables that can change revenue, margin, capital intensity, financing, and customer retention.
- Separate company-reported facts from your interpretations and identify the source for each important fact.
- Find a knowledgeable counterparty who could disagree for a specific operational reason.
- State the evidence that would make you reduce, exit, or defer the investment.
- Review past decisions for errors of mechanism, measurement, timing, and self-calibration rather than only for profit or loss.
The circle of competence is valuable because it links humility to a concrete research process. It is not an excuse to avoid unfamiliar businesses, a guarantee of good returns, or a moral judgment about who is intelligent. It is a boundary around the claims an investor is qualified to make and the money that should be placed behind them.