Investing Classics: Retail & Consumer CasesFour Consumer Franchises — and the Lessons They Teach

Costco's counterintuitive move: it caps the markup on any item at about 14% (most retailers run 25–50%). Overall gross margin is therefore only ~11%; the stores barely profit — the real earnings come from annual membership fees. This means the company's incentives are fully aligned with the customer: the more it saves you, the more members renew. Renewal rates hold steady near 90%, essentially a sticky "prepaid subscription." This is the hardest kind of "cost advantage" moat to copy — because it requires the firm to voluntarily forgo short-term margin.

Costco opened its first warehouse in Seattle in 1983 and merged with Price Club in 1993. Over the decades countless retailers tried to imitate "membership + low price" and nearly all failed — because the moment shareholders demand higher margin, the flywheel stops. Munger watched it from the board for over twenty years and repeatedly said its culture (extreme frugality + honesty toward customers) is the true barrier.

The flywheel's price is fragile membership value: the moment customers feel "the fee isn't worth it," the model stalls. Thin margins also leave no room for error. The more practical limit — a business as good as Costco is almost never "cheap": on P/E it always looks expensive, so value investors often miss it for years. The membership model is also highly dependent on specific categories and density; not every retailer can copy it.

• Are this business's incentives aligned with the customer, or does it profit from information asymmetry?
• Is its low price a temporary promotion, or a structural cost advantage?
• Is there sticky recurring revenue (membership/subscription/renewal rate)?
• If a "cheap price" never comes, will my discipline make me miss it forever?

Buffett avoided tech stocks his whole life, yet from 2016 he loaded up on Apple, which at one point in 2023 was nearly half of Berkshire's equity portfolio. The turning point wasn't that he suddenly understood chips, but that he changed the frame: not a tech bet, but a "product consumers can't live without + ecosystem lock-in." His reasoning was plain — people would sooner give up their second car than their iPhone; the switching costs woven through hardware, software and services make stickiness far exceed ordinary hardware. This is the circle of competence in action: the boundary can expand, but use a framework you truly understand (brand, stickiness, pricing power).

Berkshire first built its position in early 2016, with a cumulative cost basis of roughly $31 billion and a stake near 5.9%. By 2023 it was worth over $170 billion at one point — among the largest single-investment returns of his career. Notably, he sharply trimmed it again in 2024 — partly on valuation and tax considerations, a reminder that even the best business has "price discipline."

The most honest part of this lesson is Buffett's misses: he publicly admitted whiffing on Amazon and Google because he failed to find a framework he understood in time. "Reframing" is double-edged — it can correctly expand your circle, or supply an excuse for "I don't really understand it but want in." Pushing Apple to nearly half the portfolio is itself huge concentration risk; and if the "consumer brand" frame one day fails because innovation stalls, the narrative turns against you.

• Am I seeing it through a framework I truly understand, or using one to mask that I don't?
• Is its "stickiness" backed by a concrete mechanism (ecosystem/data/habit)?
• Has this position's concentration exceeded the single-point risk I can bear?
• However good the business, have I set a "price ceiling" for it?

In November 1963 the "salad oil scandal" erupted: De Angelis's firm used water-filled tanks to obtain fraudulent warehouse receipts, and an American Express subsidiary had vouched for salad oil that simply didn't exist, facing potential payouts of up to about $150 million. The stock nearly halved. The young Buffett walked into Omaha's restaurants and merchants and observed whether people still used their Amex cards. The answer: they did, as usual. He concluded the loss was one-time and capped, while the two-sided cardholder-and-merchant trust network underpinning the company's value was untouched.

Buffett's partnership then invested about $13 million, roughly 40% of its assets at the time — one of the most concentrated bets of his career. As the payouts settled and the brand ran as usual, the stock rose sharply over the following years, delivering about $20 million in profit to the partners. Mr. Market, in panic, had treated a one-time bill as a death sentence for the business.

This principle's most dangerous trap is survivorship bias: Amex survived, but far more "scandal stocks" never came back — the collapses of Enron, Wirecard and Lehman weren't Mr. Market overreacting; the franchise itself was dead. Telling a "quantifiable one-time loss" from a "permanent collapse of trust" is extremely hard. The Amex bet worked because the loss had a clear ceiling and the core network was observably intact — both were required.

• Is the loss one-time and capped, or will it keep eroding cash flow?
• Did customer/merchant behavior change — or only the stock price?
• Is the damage surface reputation, or the core trust it charges for?
• Am I bottom-fishing a franchise, or catching a still-falling knife?

In 1972 Berkshire bought See's for $25 million, while its net tangible assets were only about $8 million — by Graham's cigar-butt standard, a crazy "three times too expensive." But Munger convinced Buffett: See's brand lets it raise prices every year without losing customers, expanding profit with almost no added capital. That earning power "above tangible assets" is economic goodwill, especially valuable in an inflationary era — because it needs no constant capital injection to maintain real profit. The deal pushed Buffett from the cigar-butt bargain hunter toward "pay a fair price for quality."

By 2007 See's had contributed over $1.35 billion in cumulative pre-tax profit to Berkshire, on less than $40 million of additional capital over the decades. Buffett said the acquisition "changed our investment philosophy." By contrast, many "cheap" cigar butts he bought early on returned far less — the difference being that See's earnings could keep growing, rather than enjoying a one-time discount.

"Pay for quality" easily slides into "pay any price for quality." The 1972–74 "Nifty Fifty" are the lesson: a batch of acknowledged quality growth stocks, pushed to extreme valuations, halved in the ensuing bear market. The economic-goodwill logic holds only for asset-light, pricing-power businesses; for capital-intensive or commodity firms, it lures you into overpaying. Ironically See's itself had a ceiling — candy is a regional business, and Buffett could never reinvest its cash back into equally high-return "more See's."

• Does its profit growth require constant added capital, or come almost "for free"?
• Does it have a history of raising prices without losing customers?
• Is the premium I'm paying for "quality" a fair price — or any price?
• Can the cash it earns be reinvested at equally high returns?