How Did Meituan Waimai Win Originally? Can It Win Again with Its Initial Advantages?
Great summary.
Let's revisit Wang Huiwen's explanation of food delivery:
1. The economies of scale in food delivery are also quite limited. Having 100 riders around you versus 1,000 riders doesn't necessarily mean significantly faster delivery. Generally, consumers have a psychological expectation for delivery speed; once that threshold is met, going faster doesn't add much value. So after a certain scale is reached, user experience and costs won't see further improvements—hence the C-curve.
2. Looking at food delivery again, despite Meituan working so hard and struggling so much today, Ele.me still exists. This is because the economies of scale in food delivery aren't strong enough. In reality, the products of Ele.me and Meituan are very homogenized. The fact that two players remain in the industry indicates that the economies of scale in this business aren't strong.
3. Often, technological changes drive shifts in cost, experience, and possibility. A key reason the food delivery business grew so large is the普及 of smartphones. Since Apple launched the iPhone, smartphone costs have continuously dropped. Today, smartphones cost around 600 RMB, allowing delivery riders to use them. If riders had to use iPhones, they couldn't afford them. Similarly, if e-bikes were expensive, riders couldn't afford them. If merchants' order-management software ran on computers, the cost would be high, but phones are convenient and cheap. The rise of Douyin (TikTok) also stems from cheap data traffic.
4. Ele.me entered the market through campus delivery, and Meituan initially targeted the same campus market. At that time, the white-collar market was served by Daojia Meishi (Home Gourmet). Meituan didn't enter the white-collar segment initially because white-collar workers have high time-sensitivity expectations—they eat and then head to meetings. In contrast, campus users are often playing games and don't mind waiting.
In the campus market, students are densely concentrated, and merchant deliveries are simpler. Hiring part-time students to deliver cut costs to just 1 RMB per order, compared to 7 RMB per order in the white-collar market. At this stage, building an in-house delivery fleet wasn't viable—it places heavy demands on organization and would slow things down. After establishing operations this way, Meituan expanded into the white-collar market and built its own delivery team the following year.
5. This is also why Meituan didn't build its own delivery team initially. For a latecomer, building a delivery team requires capital and organizational capability, which means seeking external financing. Investors demand ROI, and the ROI in the campus market was high early on. You need a roadmap—a comprehensive, complex strategic plan from the start that affects many cost and revenue decisions.
6. During membership promotion campaigns, Meituan's strategic investment team members would all switch to using Ele.me to order food. These people were highly educated, well-paid, ordered Starbucks daily, and were considered premium users with high LTV. But once the membership promotion ended, they stopped ordering altogether.
The question here is: after free delivery fees are introduced, are consumers buying the same things on food delivery platforms as they did before? In fact, these consumers weren't buying Starbucks on food delivery platforms to begin with. Once delivery fees were waived, they started ordering Starbucks through the platform. When subsidies disappeared, they stopped and reverted to buying Starbucks offline.
7. Take this example: food delivery is a heavily subsidized business. If you look at it regionally, the most price-sensitive areas are Guomao and Zhongguancun—high-income populations who know both platforms are subsidizing. If they eat a more expensive meal at noon, they feel it makes them look unintelligent, so they try to maximize platform value by comparing prices. These are the users who squeeze the most value out of platforms.
8. Segmenting and operating by customer tier is tricky because many things go against intuition. If you don't truly segment and do the math, you'll think certain users are great. A typical example is Douban: such a great product but unprofitable. The moment they run any ads, people complain that "abei" betrayed his original ideals. This problem exists in almost all businesses. Nearly all Chinese internet companies subsidize users—but are you subsidizing the wrong ones? Just like Meituan's strategic investment users, who looked premium but had crystal-clear calculations about their own time costs and platform pricing. If subsidies ever stopped, they'd have no habitual dependency and would simply switch to another platform. Only companies like JD.com or Costco can serve this type of user well—extremely efficiency-oriented, having squeezed efficiency out of their own operations.
9. When your store has many product categories, your operations become extremely complex. Food delivery is the same: when you have many merchants, consumers, and riders in a given area, operations become very complex. To improve this, you need a dedicated team to analyze business performance. At the time, Meituan didn't have a sizable team yet. Lao Wang wanted to see which company had such a team and tried to recruit them over. He first looked at internet companies but found no satisfactory talent supply. Then he asked HR to find similar roles in offline retail organizations—but again, no talent was available. So Lao Wang went to the heads of offline retail companies and asked why they didn't have such critical positions. They said if there was such a need, they'd turn to consulting firms like MBB. So Lao Wang recruited several people from consulting firms and established Meituan's business analytics team.
So why do retail companies, which seem to need business analytics teams the most, actually lack such personnel? Lao Wang investigated what other roles they should have but didn't, and discovered many retail companies lack product and R&D teams—while Walmart did. After thinking about it for a long time, he concluded it was because their pace of change is slow. If industry change is slow, keeping business analytics and R&D staff on payroll leads to underutilization, and such teams are expensive. So when industry change is slow, the ROI of maintaining such a team is very low—in fact, it's more cost-effective for companies to skip hiring internally and instead pay consulting fees when needed. The same applies to software: most retail companies don't develop their own software or maintain a software R&D team; instead, their software is built by SAP or similar vendors. This results in slow iteration. But when an industry undergoes sudden change, companies that haven't cultivated such talent internally can't deeply understand their own business logic, and external solutions can't keep pace with industry rhythm—making it easy for the company to be disrupted. This is a severe consequence of slow industry change.
10. Remember that at that time, Meituan and Ele.me were different. Ele.me had already built up its offline operations, helping merchants manage food delivery orders, so the users it brought online already had food delivery demand. Ele.me opened in 12 cities, while Meituan simultaneously opened in over 20 cities—including some Ele.me considered off-limits. The real situation wasn't that subsidies were insufficient; rather, those campuses hadn't been cultivated and had no real demand. When many merchants were added at that point, each merchant got very few orders. At that time, the platform didn't provide delivery services—only merchants could deliver. Most merchants running brick-and-mortar stores likely had no spare staff to handle deliveries, so they prioritized serving offline customers, because offline foot traffic was larger and merchants could better perceive consumer needs. This created a vicious cycle: few consumers led to few merchant orders, which led to poor service, which made consumers think food delivery was unusable. At this point, employees would suggest adding subsidies—but that raises the question of whether subsidized users are real users.
So at one point, a city manager in Chengdu opened operations at a single campus, listed only 8 merchants, then ran aggressive promotions on campus—distributing flyers and running discounts—and told the merchants there would be many orders the next day, so they should hire extra staff. Thanks to the heavy promotion, order volume rose. With only 8 merchants, orders were spread thin across each, and merchants realized that food delivery demand was strong. At the end of the day, they identified which merchants had poor delivery performance and told them if they didn't improve the next day, they'd be taken offline. Since 8 merchants were enough to meet demand, removing one could be instantly replaced by another, so merchants immediately ramped up delivery staff and improved delivery experience. Next, Meituan reduced subsidies on the consumer side. While some consumers would inevitably leave, those who stayed did so because of the improved service experience.
Iterating one step further: as order volume and consumer experience improved, a few more merchants were added—but not too fast. Letting merchant count, order volume, consumer experience, and delivery efficiency form a virtuous cycle is what made the business turn. So recognizing whether the bottleneck in this business stage is insufficient demand or insufficient supply—and knowing how to drive demand when it's lacking, or boost supply when that's the constraint—is crucial. Get it backwards and you'll mess it up.