How Buffett judges a moat, and how to check one yourself

Buffett's 2007 letter put it simply: a truly great business must have an enduring moat that protects excellent returns on invested capital. The word that does the work is enduring. Plenty of companies earn high returns for a few years. Few keep earning them once competitors notice.

Where moats come from

Most durable advantages fall into a handful of types. A brand that lets a company charge more, like Coca-Cola. Switching costs that make customers reluctant to leave, like core banking software. Network effects, where each new user makes the product better for the rest. A cost advantage from scale or process. And regulation or licences that limit who can compete.

Buffett adds a test that cuts through the theory. In 2010 he told the Financial Crisis Inquiry Commission that the single most important decision in evaluating a business is pricing power: whether it can raise prices without losing business to a competitor.

What the numbers should show

A real moat leaves fingerprints. Gross margins stay high and stable through good years and bad. Returns on capital stay well above the cost of capital for a decade, and not only in the last year or two. The company grows without needing to pour in proportionally more capital. If any of these is missing, the moat story needs better evidence.

That's why Replicly checks the moat two ways. Code measures margin stability, returns on equity and capital, and capital intensity across the history. Then Claude reads the annual report and identifies the moat type, quoting the filing for each claim, with a separate pass looking for threats.

What to look for in the 10-K

Read the business section for how the company describes its competition. Vague claims about being an industry leader mean little; specifics about contracts, retention, network size or cost per unit mean more. In risk factors, look for named competitors, pricing pressure and technology shifts. In management's discussion, watch how they explain margin changes. Price increases that stuck are good evidence. Margin gains that rely on cost cuts are weaker.

Then ask the uncomfortable question Buffett asks: what could make this advantage less valuable in ten years? If the truthful answer is technology or regulation you can't predict, the business may be outside your circle of competence, however good it looks today.