Moats

A castle without a moat is just a building. A business without a moat is just a job.

Warren Buffett introduced this metaphor to investing in the 1980s, and it remains the single most useful framework for evaluating whether a company can sustain its profitability over decades rather than quarters. The moat is the structural advantage that prevents competitors from replicating your economics. Not your product. Not your team. Not your current margins. The durability of those margins — that’s the moat.

The distinction matters because profitability without a moat is temporary. Blackberry had 50% smartphone market share in the US in 2009 and operating margins above 20%. Five years later, the company was functionally irrelevant. The product was profitable. The profits were not durable. No moat. Coca-Cola, by contrast, has maintained gross margins above 60% for over forty consecutive years. The formula hasn’t changed. The competitive dynamics haven’t changed. The moat — an emotional brand attachment that no amount of capital can replicate — has held.

Buffett bought $1.3 billion of Coca-Cola stock in 1988 precisely because of this insight. He wasn’t paying for the syrup or the bottling network. He was paying for the fact that if you gave a competitor $100 billion and told them to displace Coca-Cola as the world’s leading soft drink, they couldn’t do it. The brand exists in the minds of billions of people. You can’t outspend that. You can’t out-engineer it. That is a moat.

The concept applies across every industry, but the sources of moats differ fundamentally. Pat Dorsey, former director of equity research at Morningstar, codified the taxonomy in “The Little Book That Builds Wealth” (2008): brand (Coca-Cola, Tiffany, Hermès), network effects (Visa, Facebook, the New York Stock Exchange), switching costs (SAP, Oracle, Bloomberg Terminal), cost advantages (Walmart, Costco, GEICO), and intangible assets including patents and regulatory licenses (Pfizer’s drug patents, utility monopolies, Disney’s intellectual property portfolio). Each source creates a different kind of barrier. Each has different durability characteristics.

Brand moats are psychological. Tiffany sells diamonds at a 10–15x markup over comparable stones from lesser-known jewelers — not because the diamonds are different, but because the blue box means something. Network-effect moats are mathematical. Each new Visa cardholder makes the network more valuable for every merchant, and each new merchant makes it more valuable for every cardholder. Visa’s operating margin has exceeded 60% for years because the network itself does the competitive work. Switching-cost moats are structural. Companies that run SAP’s enterprise software spend $50–100 million and 18–36 months implementing it; ripping it out costs just as much. The product doesn’t need to be loved. It needs to be embedded. Cost-advantage moats are operational. Walmart’s logistics network — over 150 distribution centers, a private fleet of 9,000 trucks, and satellite-linked inventory management since the 1980s — allows it to sell goods at prices competitors structurally cannot match.

The most durable businesses stack multiple moat types. Apple combines brand loyalty (consumers pay $1,000+ for phones that cost $400 to manufacture), switching costs (the iOS ecosystem traps photos, apps, messages, and habits), and network effects (iMessage’s blue-bubble social pressure among American teenagers). Disney layers intellectual property (Marvel, Star Wars, Pixar, princesses), brand power across generations, and distribution control through theme parks and Disney+. Each additional moat layer makes competitive entry exponentially harder, because an attacker would need to breach all of them simultaneously.

The critical insight that separates serious moat analysis from casual use of the term: a moat is not about having a competitive advantage. It’s about having one that persists. Groupon had a first-mover advantage in daily deals. It had no moat. Local restaurants could list on any competing deal platform with zero switching costs. Groupon’s revenue peaked at $3.2 billion in 2012 and declined to under $500 million by 2023. First-mover advantage without structural barriers is a head start in a race anyone can join.

The temporal dimension separates amateur moat analysis from professional practice. A moat’s value is a function of both width and duration — and duration is far harder to assess. Microsoft’s enterprise software moat has persisted for over three decades because the switching costs deepen every year as organizations build more workflows on top of the platform. By 2024, Microsoft 365 had over 400 million paid seats, each one adding another layer of organizational dependency that no competitor can unwind. Contrast that with Peloton, which briefly appeared to have a community-based moat during the 2020 pandemic lockdowns. Revenue surged to $4.1 billion in fiscal 2021. By fiscal 2024, it had dropped below $2.7 billion. The community engagement that felt like a moat was a circumstantial advantage that evaporated when circumstances changed. The test: would the competitive advantage survive a severe recession, a technology shift, or a change in consumer behavior? If the answer depends on conditions remaining stable, it’s not a moat. It’s a tailwind.

Hamilton Helmer formalized this in “7 Powers” (2016) by adding a critical requirement: a genuine source of competitive advantage must produce both a benefit (the company earns superior economics) and a barrier (competitors cannot replicate those economics through rational action). The dual test eliminates a surprising number of claimed moats. A food delivery company might have a benefit — faster delivery times from a denser driver network — but no barrier if drivers freely switch between platforms. A luxury car brand might have a benefit — higher willingness-to-pay from status signaling — and a genuine barrier in the century of brand-building required to replicate it. The benefit-barrier test forces precision where the moat metaphor alone permits vagueness.