At Il Giornale, Howard Schultz borrowed heavily from Italian espresso bars. Customers stood at the counter. Opera played, and the menu was in Italian. Those details expressed his idea of what Americans would enjoy. They also gave customers something definite to respond to. A hypothesis had become a business that could teach him where he was wrong.
Richard Rumelt treats strategy as that kind of commitment. Choose a reasoned explanation of what might work, then expose it to results that could contradict it. Waiting for proof would prevent many useful experiments. Protecting an idea from contrary evidence would prevent learning from them.
Schultz had noticed an anomaly while visiting Milan: expensive coffee and sociable espresso bars attracted a broad market, while high-quality dark-roasted coffee occupied a small American niche. The difference raised a possibility. It did not establish that Americans wanted an exact copy of Italian habits. A single shop required serious money, but it made the idea testable without the enormous investment some ventures needed.
At Il Giornale, Schultz and his team changed the experience as they learned. The Italian menu and opera disappeared. Chairs accommodated people who wanted to sit; paper cups later met demand for takeout. The business that eventually adopted the Starbucks name developed through these adjustments. Its success does not turn Schultz's first idea into something he knew would work.
Experience also produced knowledge competitors could not simply buy from a textbook: which customers wanted what, how operations worked, and why changes helped or failed. Roasting, branding, and serving coffee within one business let connected activities change together. Rumelt sees a reason for ownership where that coordination and learning matter. Owning another operation is often wasteful when an outside supplier already provides what is needed.
Give your judgment something to answer to
Even with useful feedback, a thinker can stop learning. The first plausible explanation brings relief from a messy problem. Once there is a course to defend, searching for another diagnosis feels like giving up progress. Rumelt's remedy starts with recovering attention from daily demands. Write down important things you can actually do, connecting immediate work with the larger purpose it should serve.
Then work backward from a proposed action to the difficulty it is meant to solve. Does that diagnosis account for the other facts? Attack its assumptions and reconsider the evidence until a substantially different response becomes possible. Slight variations on the original proposal can preserve its central mistake. Imagining criticism from people whose judgment you know well can help expose a weakness; their imagined approval supplies no evidence.
Intuition still earns a place in familiar decisions. It needs more scrutiny when judging probabilities, competence, or causation. Judging a proposal by its resemblance to a success story can push aside broader evidence about how often that kind of outcome occurs. Looking outward supplies evidence that a vivid local story can conceal, without proving that every case must follow the average.
Write a definite diagnosis or prediction before a meeting. Afterward, compare the discussion and subsequent events with that record. Without it, having considered several possibilities makes the eventual answer feel like something you predicted. The record makes practice possible by showing which judgment was yours before other people's confidence helped settle it.
When success creates its own reassurance
A crowd can make the same mistake collectively. People follow a price or a popular decision, assuming someone else checked the fundamentals. Their participation strengthens the signal that attracted them. Agreement then contains less independent information than it appears to contain. Rumelt asks for an assessment outside that circle, including relevant experience in other settings. Independence means examining evidence, not opposing whatever most people believe.
In his account of the financial crisis, credit helps explain how reassuring results can build danger. Easier borrowing supports purchases and rising property prices. Those higher prices increase the collateral available to secure further loans, supporting more borrowing and purchases. The apparent confirmation partly comes from the process being evaluated.
Falling prices can reverse the interaction. Heavily indebted owners have little remaining value after debt to absorb losses. Sales push prices down further, prompting more selling, tighter credit, and defaults. A stretch of rising prices has not tested that reversal. The sequence explains one way borrowing amplifies a boom and its collapse; it is not a complete explanation of every crisis.
Rumelt calls confidence based on recent calm the smooth-sailing fallacy. A design flaw may remain dormant until conditions change. Complexity compounds the problem when people cannot understand how the system's parts could fail together. Assessing safety requires considering unwelcome conditions and consequences, including combinations absent from recent experience. Declaring such questions disloyal disables the analysis.
Incentives matter alongside beliefs. Someone who keeps the gains while passing losses to others has more reason to create risk. Fees and compensation can reward continuation even when the arrangement is becoming dangerous. Neither enthusiasm nor a quiet record resolves that conflict.
A definite decision gives learning somewhere to begin. Keep track of what supports it, what would contradict it, and how the system might behave beyond favorable conditions. These practices cannot find every danger or guarantee a successful strategy. They let commitment remain accountable to evidence as circumstances change.
Three things to keep
Commit to a hypothesis that observable results could challenge.
Record judgments before discussion, then compare them with what happens.
Test reassuring results against independent evidence, adverse conditions, and incentives.