This Issue at a Glance
Tony Xu, co-founder and CEO of DoorDash, shared his deep insights on one of the world's largest food delivery platforms during the program. Core thesis: DoorDash's essence lies in the "combination of mathematical and human problems" — logistics efficiency and human coordination must be solved simultaneously. The team's early practice of personally making deliveries to understand the physical details of the "atomic world" was key to building a sustainable business model.
Theme 1: Start with the "Human Problem," Then Solve the "Math Problem"
Tony Xu believes that the starting point of DoorDash's success is not algorithms, but a deep understanding of every link in the physical world.
"DoorDash really is a marriage between a math problem and a human problem."
- Understand the "atoms" first, then process the "bits": In the first two years, all team members (including engineers) personally delivered orders every day, breaking down a single delivery into 5-10 steps and further refining it to the lowest level of detail. This "ground-level perspective" allowed them to uncover variables that mathematical models could not foresee—such as fluctuations in restaurant kitchen capacity, the efficiency of different entrances, and delivery routes on the sixth floor of a shopping mall.
- Node density over population density: Tony points out that the key to delivery efficiency is not population density, but "node density"—the spatial distribution of stores, consumers, and delivery personnel. Most U.S. cities follow a "hub-and-spoke" model, where main streets and surrounding residential/office areas form efficient delivery units. Suburbs like Palo Alto are actually easier for fast delivery than San Francisco, because variables such as parking, entrances, and routes are more controllable.
- The math problem remains unsolved: Traditional "traveling salesman problems" or "multi-vehicle routing problems" assume that demand and supply are predictable. However, in DoorDash's system, when consumers place orders, whether delivery personnel accept them, and restaurant preparation times are all uncontrollable. Therefore, structured information must first be collected through human experience before applying mathematical optimization.
Theme 2: Counterintuitive Product Building — Prioritizing the Invisible "Backend"
Tony Xu emphasizes that DoorDash's most critical early decision was to prioritize building products for the delivery driver side and the merchant side, rather than the consumer side.
- Core Hypothesis Testing: The team's core hypothesis was — "If we can deliver from restaurants that have never offered delivery before, will consumers care?" To test this, the V1 product was nothing more than a Google Voice number that directly forwarded calls to the founder's phone.
- 90%+ of Engineering Resources Devoted to the Backend: In the first two years, the team spent the vast majority of engineering time developing the delivery driver app, logistics dispatch system, exception handling system, POS integration, merchant portal, and kitchen display system — all invisible to consumers. Tony argues that if the team had spent two years building the consumer-facing product but could not deliver reliably and accurately from the restaurants consumers wanted, the conclusion might have been "this business cannot work."
- Deliberate Slowing Down for High-Consequence Decisions: Tony describes this choice as a "high-consequence decision" — once the wrong path is chosen, it could lead to an incorrect business conclusion. Therefore, the team deliberately slowed down the development pace of the consumer-facing product, prioritizing the core variable of delivery reliability.
> Readers should note: This is a narrative from the perspective of a position holder. Tony uses "prioritizing the backend" to justify the rationality of his product strategy, but the success of this path also depends on the effective conversion of the subsequent consumer-facing product.
Theme 3: The Unexpected Benefit of Capital Constraints—Forcing Unit Economics
Tony Xu points out that DoorDash's early financing difficulties actually became an advantage, forcing the team to ensure positive unit economics for every order from the very beginning.
- Funding Challenges: DoorDash encountered difficulties in its seed round. Tony believes the reason was that investors lacked the information needed to make decisions—do consumers actually need delivery services? Can logistics issues be profitably solved? How will the market landscape evolve?
- Forced Mechanism of Capital Efficiency: With extremely limited funds (Tony notes that early capital was only 1/100th of peers), the team had to "obsessively" ensure positive unit economics for each order. As a result, for the first 5.5 years of DoorDash's existence, the company remained undercapitalized, but its bank account did not shrink rapidly.
- Unit Economics and Network Health Can Coexist: Tony refutes the view that "unit economics and network health are inherently in conflict," arguing that the real conflict is between "overall company profitability" and "network health," not unit economics itself. The key lies in understanding how much order volume is needed to achieve overall profitability—this is an investment decision.
| Dimension |
DoorDash's Early Approach |
Common Market Assumption |
| Capital Position |
Severely insufficient (seed round difficulties) |
Ample funding drives growth |
| Unit Economics |
Must be positive per order |
Losses acceptable for scale |
| Product Priority |
Backend (delivery personnel + merchants) |
Frontend (consumer side) |
| Core Hypothesis Validation |
Delivery reliability |
Consumer demand |
Theme 4: Leadership Philosophy — From "Observing the Self" to "System Design"
Tony Xu shared his unique perspective on leadership: culture is 80% "who you are" and 20% "who you want to become," and leadership should start from observing one's own strengths rather than copying an external template.
- Culture grows from stories: DoorDash did not write down its leadership principles until 2.5 years after its founding. Instead of setting principles first and then finding stories, the company first collected real stories from the preceding two years and then distilled principles from them. For example, the "customer obsession" principle originated from a massive system failure on September 28, 2013 — the team decided within 30 seconds to refund all orders (amounting to a double-digit percentage of the bank account balance), stayed up all night baking cookies, and delivered them to customers by 5 a.m.
- Evolves with the company's stage: In 2017, DoorDash added the principle of "making room at the table," committing to building a management team and a company as diverse as its customer base. This is the 20% "who you want to become" part — there were no ready-made stories, but deliberate construction was required.
- Three mechanisms for scaling: ① Written communication (meetings start with a document, a practice maintained for 5 years); ② Hiring for value alignment ("it's much easier to correct at the door than after the fact"); ③ Continuous reinforcement and reduction of gaps.
- Six traits of top talent: ① Ability to toggle between "high and low," diving into the lowest level of detail; ② Strong bias toward action; ③ Excellent hiring skills; ④ Deliberate improvement of 1% every day; ⑤ Possessing "followership" (leadership = how many people are willing to follow you, not how many report to you); ⑥ Ability to hold two opposing views simultaneously and constantly challenge one's own biases.
Theme 5: Digitization of Local Commerce – Still in the Very Early Stages
Tony Xu believes that, by any measure, the progress of the problem DoorDash is solving remains "very low."
- Food delivery penetration is only 10%+: Even with the pandemic's boost, food delivery as a share of total restaurant sales remains in the single digits, just above 10%. In comparison, ticket digitization has reached 40-50%, and e-commerce 20-25%.
- Full-scale online migration of local commerce has barely begun: Tony points out that we still do not know "where the last parking spot in a city is," "how much faster it is to make a salad than to bake a deep-dish pizza," or "which shelf position the apples are on in the sixth row"—once this information is digitized, it will spawn a host of new business models.
- Centralization vs. decentralization: Tony contrasts Amazon-style centralized warehousing and delivery with DoorDash-style local delivery as "centralization vs. decentralization." Consumers always want the widest selection, highest quality, and lowest price. Centralization allows control over the catalog but suffers from distance, while decentralization enables speed and low cost but is constrained by urban space. He believes the two models are converging—every street-corner store is simultaneously selling "experience" and "convenience."
- Impact on urban form: Tony argues that DoorDash will not change the fundamental logic of where people live, but rather eliminate "geographic privilege"—New Yorkers can access delivery services not because they are more special, but because they have better infrastructure. When this convenience spreads across the U.S., it will unlock pent-up demand.
Mentioned Positions
| Position |
Guest Stance |
Key Data |
| DoorDash |
Bullish (Founder’s Perspective) |
Food delivery penetration just over 10%; early-stage capital only 1/100 of peers; positive unit economics despite insufficient capital in the first 5.5 years |
| Shopify |
Neutral (Mentioned as Analogy) |
Used as an analogy for a "merchant website-building platform" |
Judgments Worth Remembering
1. “Node density over population density”——Tony Xu: The key to delivery efficiency is not urban population density, but the spatial distribution of stores, consumers, and couriers (node density). Suburbs like Palo Alto are actually easier for rapid delivery than San Francisco, because variables such as parking, entrances, and routes are more controllable.
2. “Build the back end first, then the front end”——Tony Xu: In its first two years, DoorDash allocated 90%+ of engineering resources to courier-side and merchant-side products (the parts invisible to consumers), rather than the consumer app. The core assumption was: if delivery cannot be reliably executed, no matter how good the consumer-facing side is, it is useless.
3. “Capital constraints were an accidental advantage”——Tony Xu: The difficulty of raising seed funding forced the team to ensure positive unit economics per order from day one. In the first 5.5 years, the amount of capital was only 1/100 of that of peers, but the bank account did not shrink quickly—capital efficiency became a core competitive advantage.
4. “Culture is 80% who you are and 20% who you want to become”——Tony Xu: DoorDash did not write down its leadership principles until 2.5 years after founding, and they were distilled from real stories of the first two years, not copied from external templates. The principle “Make room at the table,” added in 2017, represents the 20% “who you want to become” part.
5. “Six traits of great talent”——Tony Xu: ① Ability to dive deep into the lowest level of detail; ② Strong bias toward action; ③ Excellent hiring ability; ④ Improve 1% every day; ⑤ Possess followership (leadership = how many people are willing to follow you); ⑥ Ability to hold two opposing views simultaneously and constantly challenge one’s own biases.
6. “The digitization progress bar for local commerce is almost at zero”——Tony Xu: Food delivery penetration has just passed 10% (compared to 40-50% for ticketing and 20-25% for e-commerce); the full online transformation of local commerce “hasn’t even shown a dot on the progress bar.” We still don’t know “where the last parking spot is” or “how much faster making a salad is than baking a pizza.”
7. “Centralization and decentralization are converging”——Tony Xu: Amazon-style centralized warehousing and DoorDash-style local delivery are not opposites. Every street-corner store is simultaneously selling “experience” and “convenience,” and the two models will gradually converge.
8. “There is no final whistle in building a company”——Tony Xu: Unlike sports or academics, building a company has no definition of “ending.” The best success is a company that never ends—this requires the team to appreciate the journey itself, rather than chasing some final destination.