Starbucks

Starbucks buys green coffee at commodity prices and sells it, transformed, at a multiple no commodity supports on its own. The gap is paid for by the room. Schultz's insight after Milan in 1983 was that the American market had a home and a workplace and nothing in between, and that a company willing to underwrite the rent on that gap — the "third place" — could charge for the coffee as if it were something else.

Structurally the company runs two store systems. Company-operated stores carry the lease, the payroll and the full revenue line; licensed stores sit inside airports, grocery chains and hotels, where a partner takes the operating risk and Starbucks takes a royalty plus the product sale. Company-operated stores are the large majority of revenue and roughly half the store count — a deliberate choice to keep control of the experience, and the reason the P&L looks nothing like McDonald's.

The live problem is that the third place and throughput are now in conflict. Mobile order-and-pay solved a queueing problem and created a worse one: a store designed as a living room became a pickup counter with congestion at the handoff plane, seats occupied by nobody, and baristas working a ticket queue they cannot see the end of. That is not a technology failure. It is a company optimising the metric it can measure — transactions per hour — against the asset it actually sells, which is the feeling of the room.

The moat has three layers and they are weakening at different speeds. The first is real estate: decades of site selection means Starbucks holds corners a competitor cannot buy, and density means a customer passes three of them on a normal commute. That layer is durable and hard to attack.

The second is the habit loop, formalised in Rewards. A stored-value balance plus a personalised order is a small but real switching cost — the customer has money on deposit and a drink the system already knows. Starbucks holds hundreds of millions of dollars of customer prepayments at any moment, an interest-free float that also functions as a lock-in.

The third layer is the one under pressure: the premium itself. When the room stops being worth waiting in, a $6 latte competes on price and speed against Dunkin', McDonald's McCafé, drive-thru chains like Dutch Bros, and in China against Luckin, which sells a comparable cup for a fraction of the price and won the market on speed and discounting. Starbucks' China business went from growth engine to strategic problem in under five years, and the company has since sold control of the retail operation there to a local partner. A moat built on being the nicer option does not survive the option ceasing to be nicer.

For a decade the allocation policy was simple and, for a while, correct: open stores, buy back stock, raise the dividend. It returned enormous cash to shareholders and it hollowed out the balance sheet, taking book equity negative. The buybacks were executed at prices that assumed the growth algorithm was permanent.

The more consequential misallocation was operational rather than financial. Capital and management attention went into digital ordering and throughput — the things that made the next quarter's transaction count look better — while the physical asset that justified the price premium was allowed to degrade into a pickup counter. Niccol has reversed the priority: labour hours restored, menu simplified, ceramic mugs and seating back, condiment bars returned, and hundreds of underperforming stores closed. That is the right diagnosis, and it is expensive.

The China decision is the cleanest recent call. Rather than defend an unwinnable price war with owned capital, Starbucks took a minority-plus-royalty structure with a local partner — converting an operating problem into an annuity. Whether that reads as discipline or retreat depends entirely on what Luckin does next.