Li Xiang’s Endgame Thinking: Scale Assumptions, the 20,000-Person Trap, and Three Fundamentals

The post before last covered the decline of Chinese retail giant Gome. The previous post covered how century-old Japanese manufacturer Ajinomoto transformed toward the semiconductor industry. This post looks at how Li Auto founder Li Xiang uses endgame thinking to understand the entrepreneurial journey. Recently, Li Auto founder Li Xiang gave an interview to LatePost. It addressed many key questions about the company. On product positioning, for example, Li said future products would remain centered on family users. Even in the RMB 200,000–300,000 price range, features serving the family experience—four screens, six spacious seats, and navigation-assisted driving—would remain standard. On market expectations, Li said many people expected numerous automakers to coexist in the final market structure. He disagreed. He expected the industry eventually to resemble smartphones, with four or five companies holding more than 80% of the market. Under that assumption, Li Auto's approach would differ from NIO's and XPeng's. NIO played a steady, conservative hand: modest expectations but reliable execution. XPeng took a leapfrogging approach, with each model entering an entirely new market. Li Auto would bet only on breakout products. Once it committed to a product, that product had to hold an overwhelming advantage over others in its segment. The interview contained many more insights specific to new-energy vehicles. It also offered several perspectives useful to people in other industries. First, how do you develop products through endgame thinking? The term has been popular over the past two years. As the saying goes, do not take the first step without seeing a path to the heavens. Before beginning something, envision what it will look like when it succeeds. But the broad idea of an endgame contains many variables: technology, users, and markets ten years from now. Which should come first? Li believes scale is especially important. Before starting, decide how large the business should become. This advice originally came from Meituan's Wang Xing and Wang Huiwen: discuss scale first, then changes in technology, demand, business, and competition. As complexity scientist Geoffrey West has said, all changes ultimately manifest as changes in scale. The growth of life, for example, is essentially an increase in cells; a company's growth is essentially an increase in users. For Li Auto, adopting the perspective of scale answers many questions. First, should it develop platforms and prepare its supply chain in advance? Absolutely. Like Apple and Huawei, it should not develop each final product independently. It should combine hardware and software platforms as though assembling building blocks. Platforms are the only way to solve the capacity problem. Wang Xing once did the arithmetic for Li: expanding production too early might waste one or two hundred million on factories. But if demand grows and supply cannot keep up, the losses could run into billions or tens of billions. Second, within this platform approach, determine what must be developed in-house. For Li Auto, batteries and panels become more efficient and less costly as they become more standardized, so it leaves them alone. Intelligent sensors have an established supply chain, but internal development could reduce costs, so the decision remains flexible. Components that differentiate the product—such as very large head-up displays (HUDs), air suspension, and some core chips—must be developed in-house. That is the first takeaway: developing products from an endgame perspective. The second is how to avoid the 20,000-person trap. Many companies, particularly internet companies, have long spoken of a 20,000-person trap: problems can emerge once headcount passes that threshold. When Huawei first exceeded 20,000 employees in 2001, for instance, Ren Zhengfei wrote “Huawei's Winter,” sharply criticizing its problems. When Alibaba exceeded 20,000 employees in 2009, its founding team collectively resigned and reapplied for their positions. In many companies' eyes, 20,000 people marks a qualitative change that requires an upgrade in organizational management. For Li Auto, that upgrade centers on building a matrix organization. Dedao instructors Liu Run and Ning Xiangdong have explained its precise definition; their lessons can be found on Dedao. The core of a matrix organization is interlocking objectives. As a company grows, it becomes a system with multiple objectives. An innovation team might develop a new product, while sales prefers established products with mature markets, strong recognition, and better earnings. Although they belong to the same company, their goals differ. A matrix organization essentially links those goals together. Salespeople, for example, are assessed both by their sales manager and by the innovation team. The latter can give new products greater weight: selling one new product might count as selling ten established products. Many departments then have both vertical and horizontal performance measures. The overall structure resembles an intersecting matrix. The benefit is that every objective receives attention, leaving no obvious organizational weakness. To borrow Wang Huiwen's analogy, a team with one distinctive specialty can play in a league. But a World Cup finalist cannot be such a team: it has to be strong in every area. That is the second takeaway: avoiding the 20,000-person trap by interlocking objectives through a matrix organization. The third is how to keep developing talent. This year, 55% of Li Auto's new hires came through campus recruiting, making a dedicated development system important. Employees are expected to master three fundamentals. First, solve language problems. Learn to communicate with others and treat communication as important work. At a particularly difficult point, Li made a list of companies that might be interested in them. He drew by hand on an A4 sheet what Li Auto could offer those companies and carried it with him. With that sheet, he approached them one by one. That round of conversations ultimately secured Wang Xing's investment. Second, solve mathematics problems. Support growth with a rational, measurable numerical system. Li Auto uses the formula: product value = product appeal × product quality ÷ product price. Product appeal has more than a hundred defined dimensions, and both emotional and rational factors can be assessed quantitatively. Only with that degree of rigor can a product be carried forward. Third, solve philosophy problems. Keep asking why to establish the right direction. Two principles must guide that inquiry, drawn from Russell's wisdom and fundamental to Li's understanding of philosophy: be grounded in facts and seek truth rationally; practice inclusive cooperation so that the world can function constructively. One final observation: if a philosopher were to serve as a mentor to today's founders, Russell might be a good choice. Why? In The Life of Russell, he is quoted as saying that three passions governed his life: First, a commitment to love. Second, a commitment to knowledge. Third, compassion for human suffering. Doesn't that sound remarkably like what we often say about products loving their users, building learning organizations, and developing products around users' pain points? #DedaoHeadlines20221221