This is an English adaptation of a FoodBud historical article originally published on March 8, 2022.
This article adapts ideas from Richard Rumelt’s Good Strategy/Bad Strategy, using Starbucks as a case study in how operators build valuable proprietary knowledge through focused experimentation.
Strategic insight is, in essence, the ability to be less short-sighted than others: to understand industry structure, trends, competitors, internal capabilities and resources, and to reduce bias in one’s own thinking.
For operators, the most valuable practical knowledge is not generic know-how that anyone can access. It is proprietary knowledge: what only your company can learn by operating, testing, observing customers, and adjusting in the field.
In 1983, Howard Schultz noticed an anomaly that later shaped a new business model. At the time, he was working in marketing and retail operations for Starbucks in Seattle, where the company mainly sold coffee beans.
On a business trip to Italy, Schultz experienced Italian espresso bars. He saw that cafés were popular, lively, and social. Each had its own character, but customers knew one another and had familiar relationships with the baristas.
Italy then had about 200,000 cafés. Milan alone had about 1,500, in a city comparable in size to Philadelphia.
From a retailer’s perspective, Schultz saw two things: high customer turnover and relatively high coffee prices. This was unlike his understanding of the U.S. coffee market.
In Seattle, espresso was a tiny market, mostly serving a small group of demanding customers, though that group was growing. Across Seattle and the broader United States, most people drank inexpensive, ordinary coffee, including affluent consumers. In Milan, expensive espresso had become a mass-market product.
Schultz also saw a second anomaly. In the U.S., fast food meant low-priced items in plastic packaging. In Milan, espresso was prepared quickly, sold at premium prices, and consumed in cafés that served as warm social spaces.
His strategic hypothesis was simple: if the Italian espresso experience could be replicated in the United States, mainstream consumers would embrace espresso.
After returning to Seattle, Schultz presented the idea to Starbucks’ two leaders. They allowed him a small space to make espresso, but they did not share his view of the opportunity. To them, Starbucks’ strength and mission were buying, processing, and selling coffee beans, not operating espresso cafés.
They also viewed espresso bars as a small, unremarkable market, associated with marginal customer groups rather than a mainstream opportunity.
Schultz’s first challenge was that his vision required a major shift in U.S. consumer taste and habits. What he saw in Milan was not just a business model. It reflected centuries of social development in Italy.
In the U.S., coffee had replaced tea and was consumed with meals or during breaks. In southern Europe, coffee had replaced wine in social settings, served in small quantities but with strong flavor.
His second challenge was that coffee, espresso, cafés, and espresso bars were not new. Millions of Americans had visited Italy and experienced espresso bars. Coffee-industry knowledge was not proprietary. If Schultz was going to create a new business, he needed to learn something others did not know or control a scarce resource.
At first, his proprietary knowledge was only an idea, a mood, and a feeling. Others who had experienced Italian espresso may not have drawn the same conclusion.
Fortunately, the idea could be tested without enormous capital. Opening one espresso bar required hundreds of thousands of dollars, not hundreds of millions or billions.
Schultz later left Starbucks and founded Il Giornale. The store directly imitated Italian espresso bars. Its 700-square-foot space used Italian-style décor, had no chairs, and customers stood while drinking coffee, as in Milan.
Had Schultz stayed rigidly attached to the original concept, Il Giornale might have remained a small espresso bar. Instead, he and his management team treated it as an experiment and watched customer reactions closely.
As the team gathered proprietary information, Schultz changed the operating model:
In international business terms, Schultz gradually localized the Italian espresso experience for American consumers.
In 1987, Schultz’s company fully acquired Starbucks and adopted the Starbucks name. The new company retained Starbucks’ original coffee-bean business while expanding espresso operations.
By 1990, the company was profitable. In 1992, Starbucks had 125 stores, 2,000 employees, and completed its IPO.
By 2001, Starbucks had become a U.S. icon, with 4,700 stores worldwide and revenue of $2.6 billion. Most of its revenue came from espresso-based beverages, described as handcrafted drinks, while the rest came from coffee beans, other food items, and licensing fees.
Only a few years earlier, a cup of coffee had cost 75 cents and was served in plastic cups. Starbucks helped normalize the sight of young professionals carrying and tasting $3 lattes.
Schultz began with a hypothesis: open an Italian espresso bar in Seattle. Testing showed the idea was viable, but the experiment also produced new information. The original hypothesis evolved through hundreds of trials into many new hypotheses shaped by business growth.
That learning process - observe an anomaly, form a hypothesis, test it, revise it, and test again - is scientific induction. For any successful operator, it is essential.
One reason Starbucks succeeded was that many customers were willing to pay premium prices for a handcrafted beverage experience in an urban oasis. But competition still matters: why did Starbucks keep competitors at bay for so many years?
In 2001, the author met Joe Santos in Paris. Santos taught strategy at INSEAD and had previously been CEO of Segafredo, a major Italian coffee company and supplier of roasted espresso beans to European hotels and espresso bars.
Santos explained that European coffee companies struggled to understand Starbucks and therefore struggled to respond competitively.
Segafredo supplied more than 50,000 cafés and restaurants each week. From that perspective, Starbucks still looked small. Large U.S. coffee companies such as Kraft, Sara Lee, and P&G focused on the mass coffee market.
Europeans had difficulty categorizing Starbucks. In Europe, cafés and restaurants were distinct. Starbucks was known as a coffee company, but it was actually a retailer. McDonald’s is also a retailer, Santos noted, but no one confuses McDonald’s with a beef company.
Europeans tended to see Starbucks as an American-style café, while Americans saw it as an Italian-style espresso bar. Yet real Italian espresso bars had no chairs, served pure espresso in small porcelain cups, did not offer takeaway coffee, had no tables, and did not brand the espresso by the café name because beans came from suppliers such as Segafredo.
Italian espresso bars were usually small family businesses, not units of a large chain.
Starbucks was different: customers could sit or take drinks away, coffee was served in paper cups, the menu included drinks unfamiliar or unappealing to many Europeans, the coffee was Starbucks-branded, and stores were company-owned. Most drinks involved milk. From a European view, Starbucks looked less like a coffee company than a milk company selling coffee-flavored milk.
Starbucks combined café, restaurant, chain retailing, and public-market financing, allowing it to expand far faster than many European players. By the time competitors understood the model, Starbucks had already scaled.
The deeper point is vertical integration. Starbucks made coffee, branded it with its own name, and served it through stores it operated. This was not designed to confuse competitors. It allowed Starbucks to adjust multiple parts of the business system at once and rapidly learn from those changes.
Vertical integration is not universally right. If a company can buy excellent products and services from external suppliers, it should not waste capital and management energy mastering production unnecessarily.
But when a strategy depends on adjusting several business links at the same time, especially when the interaction among those links generates important information, ownership and control of critical links become vital.
The article closes with a broader lesson from the author’s work with Hughes Electronics, whose businesses included communication satellites, missile systems, and other aerospace operations.
In a large production facility, the author observed a highly sophisticated communication satellite. The equipment integrated orbital mechanics, solar power, three-axis positioning, advanced computing, and electromagnetic receiving, amplification, and transmission technologies. In fixed orbit more than 22,500 miles above Earth, it could operate for decades without maintenance.
The author was helping Hughes managers develop strategy. Many were engineers promoted into management because of their ability to organize and direct technical work.
One experienced engineering manager, Barry, challenged the usefulness of strategy. Engineering, he argued, requires knowing what happens if one does A or B, then choosing the best path. Large aerospace systems cannot be built without careful planning, but strategy felt empty because it lacked clear theory.
The author’s response was to compare business strategy with science.
Scientific knowledge advances at the frontier of what is known. A good scientist forms a hypothesis about the unknown, tests it against established principles and past results, then tests it in reality. If scientists avoid the frontier and work only with certainty, they may be comfortable, but they will not create much new knowledge.
Good business strategy also operates at the boundary between the known and the unknown. Competition pushes companies to that frontier, where opportunities to outperform rivals can be found.
A new strategy should first be tested against established principles and accumulated business knowledge. If it passes that test, it must be applied in practice to see what happens.
Asking a strategist to produce a strategy that is guaranteed to work is like asking a scientist to produce a hypothesis guaranteed to be true. A good strategy, like a good scientific hypothesis, is grounded in knowledge and experience but still must be tested.
The key difference is that scientific knowledge is often broadly shared, while good strategy depends heavily on the distinctive wisdom a company has accumulated in its own business and industry.
For foodservice operators, the Starbucks case is a reminder that strategy is not only positioning or planning. It is a disciplined process of building proprietary operating knowledge through experiments that competitors cannot easily copy because they have not done the same learning.
Note: IPO and financial figures in this article are historical, drawn from the original article.