46. So-Young's 金星 in His Own Words: 6 Years as a Nasdaq-Listed Company, from Peak to Bottom, and My Fourth Start

Summary

• So-Young went through a full China-concept-stock drawdown cycle: its shares fell from a post-IPO high of $22 to $0.50 at the end of 2022, a 97% collapse, and at one point received a Nasdaq delisting warning—“only last night did the stock return to where it was four years ago.” 金星 has not sold a single share since the IPO, and actually raised his stake from 16% to 25%; the pain was not about monetization but about dignity—“for the boss of a listed company, the stock price is a matter of self-respect—male self-respect.”
• In reviewing the failure, 金星 does not blame regulators or Big Tech: “The fundamental problem is that we simply weren’t good enough”—the platform model could not solve consumers’ biggest pain point, the medical-delivery process. So-Young’s post-IPO overexposure, including a nearly RMB70M Focus Media campaign, amplified the category’s visibility while putting the company more directly in the path of tighter regulation and Big Tech competition; the biggest strategic mistake was copying Meituan: “You don’t beat Meituan by becoming another Meituan; if you copy it, you’ll always be an inferior Meituan.”
• So-Young has fully pivoted to offline medical-aesthetics chains while still defining itself as an “industrial platform,” decomposing the sector into doctors, premises, equipment and products for coordinated deployment, much as Didi separated drivers from cars. Korean light-medical-aesthetics chains charge an average of KRW100,000, or about RMB550, yet achieve a 25% net margin, versus roughly 5% at well-run Chinese institutions; So-Young’s sales productivity per square meter is 3x that of ordinary institutions, and mature stores have net margins above 10%.
• The 1,000-store plan is engineered with extreme restraint: So-Young first used the virtual chain “新氧优享,” with nearly 200 shop-in-shop locations, to build its middle platform; after opening its first store in 2023, it waited 10 months before opening the second to validate the model, then added stores 2 through 20 in another 10 months. It opened 20 stores last year and targets 30 this year; franchising will begin with 3 fully managed stores—“I have the financial wherewithal, but I would feel it was unsafe.”
• The upstream bet is substantial: So-Young paid RMB791M for an 84.9% stake in 奇致激光, secured commercialization rights to 4 major injectable products in 2023, with registration approvals expected in 2026-2027, invested more than RMB1B over 4 years, and saw cash fall from a peak of RMB2.8B to RMB1.1B. He believes the upstream can be “Xiaomi-fied”—M22’s annual sales of 600 units is already a historic peak, with high prices and low volume reinforcing each other; his vision includes giving away equipment and charging by usage or revenue share to unlock more than 10,000 small and midsize institutions.
• The industry is undergoing a major supply-side shakeout: the top 10 medical-aesthetics groups have increased their share from less than 10% 5 years ago to 25%, while the small and midsize clinics that account for 95% of the market are closing in large numbers—but the industry itself has not shrunk. The host cited China’s penetration rate rising from 4% to 6%, versus 22% in South Korea; 金星 said upstream companies had grown from 100-200 to at least 1,000, with new entrants that have R&D capabilities becoming forces for price cuts—he sees the mass-market shift in medical aesthetics as unstoppable, and says China most lacks a Sam’s Club or Fat Dong Lai-style model of “high quality at low prices.”
• This is also a case study in founder psychology: the true darkest hour was not the stock price but So-Young’s app being taken down without warning for 5 full months—“not once did anyone give me a written explanation of why.” “Did I ever think about lying flat? For a moment, yes,” but his rule was “to leave as a winner”; his old boast that So-Young would reach RMB100B within 5 years now feels “so embarrassing,” yet today “I actually believe it even more.”

Deep dive

1. Six years as a listed company rode 2 waves; growth had already begun slowing before the IPO

• 金星 attributes So-Young’s rise without taking credit: “First was the broad mobile-internet wave; second, the past 10 years really were a decade of rapid medical-aesthetics development in China.” So-Young was founded in 2013 and listed on Nasdaq in 2019.
• The growth curve was the tell: after commercialization, revenue grew 600% YoY in 2016, 300% in 2017, and 200% in 2018, before dropping to 87% in 2019—“you could clearly feel growth declining bit by bit.” The business truly went off target 2 years after the IPO, when the major platforms moved in.
• Discussing the documentary Ignition and the baggage around his public comments, he stressed that he has always been transparent: “If someone knowing your strategy causes a major setback, that means the company has no competitive moat and no differentiation.”

2. The introverted small-town boy and the life formula of “creation minus consumption”

• He grew up in Qitaihe, a small coal-mining city, and was “an especially withdrawn kid even by Northeastern standards”; his record was spending an entire summer without stepping outside. He ranked second in science across his school and wanted from an early age to do something significant that combined technology and management.
• In middle school, he came up with his own formula: “A person’s value equals what he creates minus what he consumes; the larger the difference, the greater the value of his life.”

3. Not a proof-seeker like 安传东, but a problem-solver

• On 安传东’s obsession in Ignition with using entrepreneurship to prove himself, he said he did not strongly relate: “I don’t care that much about what other people think… After I took this company public, I cared even less, because I had already proved myself once.”
• He considers himself a “chosen entrepreneur” for 2 reasons: exceptional resilience and a powerful creative impulse—“like someone who paints and desperately wants to create a painting that a lot of people will see.”
• The host put it bluntly and he accepted the characterization: “You’re not an ambitious person?” “I’m a problem solver.” That mindset also foreshadowed his later failure to build a community.

4. The IPO was too conspicuous: advertising amplified visibility and exposed the company to regulators and Big Tech competition

• He admitted he regretted the launch strategy: “We were too high-profile at the IPO. We showed everyone how profitable this category was… and a lot of Big Tech companies rushed in.” A Focus Media campaign ran for more than a month and cost nearly RMB70M, with the slogan “So-Young Medical Aesthetics—beautiful, beautiful, beautiful, beautiful”; “you can’t control who sees it.” In a heavily regulated industry, the government may naturally focus on the most conspicuous platform, and enforcement became stricter every year from 2020.
• But So-Young’s internal review classified the high profile as only an external factor: “This business model simply cannot solve the consumer’s fundamental problem.” The core pain point remained medical delivery, which an internet platform could not reach; that became the seed of his fourth venture, this time offline.

5. How Big Tech entered: So-Young shared data with Meituan, while Douyin’s livestreaming business was reported

• Meituan initially operated only a beauty channel. 金星 was close to the executive in charge and “shared a lot of our data with them… I’m just someone who shares too much.” Douyin entered local services to compete with Meituan, and its livestreaming volume ramped quickly, potentially overtaking Meituan medical aesthetics within a year. After someone reported the business, the government prohibited medical-aesthetics livestreaming, pushing Douyin toward a more cautious, conservative approach.
• Xiaohongshu has moved into medical aesthetics in recent years and also benefited from the early platform economics, but fake content remains an unsolved problem: many cases are fabricated, “even when the cases in the videos are fake,” and specialized companies mass-produce fake diaries to sell to institutions. “We still haven’t seen what new solution Xiaohongshu can actually find.”
• Big Tech’s structural advantage is that medical aesthetics is embedded inside platforms with abundant traffic, allowing them to subsidize consumers with thicker margins. They also do not need to buy external advertising as frequently as So-Young, and their regulatory exposure looks different. The result is a structural problem So-Young cannot easily solve: the same store and procedure may be cheaper on Meituan.

6. The biggest mistake was copying Meituan: product-level granularity was So-Young’s own path

• Platform breadth determines scale: “Meituan can probably serve as an information intermediary across 100 industries… A super app is like a giant island—the bigger it is, the more stable it is.” So-Young’s category breadth was too narrow; given its position, going deeper was the rational choice.
• So-Young’s internal review identified its biggest mistake as “learning from Meituan.” Without Meituan’s traffic and margins, “you simply didn’t have the foundation for that kind of intensive operation.”
• Its original differentiation was e-commerce, not LBS. Meituan used the store as its smallest unit; So-Young used the product—searching for double-eyelid surgery returned specific procedures at individual institutions. A user in a fourth-tier city might travel across the region to find “the best nose job in all of Beijing.” The model should have been “Taobao, not Meituan.” He agreed with the host’s summary: “You forgot who you were.”

7. The near miss with Xiaohongshu: the trust trap of a low-frequency platform

• He had known 瞿芳 since 2013 and urged her in Shanghai to build a community, explaining the logic behind his own women’s community. He did not build one himself—“because I think in terms of solving problems… I may have missed other things, but that isn’t necessarily bad.”
• When the host argued that Xiaohongshu’s medical-aesthetics content is better today, 金星 asked whether that was being measured by data before explaining the frequency logic: medical aesthetics is inherently low frequency, and “when you interact with a platform very frequently, your trust in it keeps increasing.”
• So-Young’s postoperative support can build deep trust, “but it’s one-time trust. Once the recovery period is over… the user may simply uninstall the app.” The platform must also compete with institutions for users through top-ups, membership cards and repeat purchases, making retention its most painful challenge.

8. So-Young tested offline in 2015, then used a “rural encirclement” strategy in 2021

• In 2015, it partnered with a Taiwanese team on its first offline experiment, “Cloud Clinic.” It shut the project down after a year without success. At the time, the internet business was growing too quickly, with large numbers of new users arriving every year: “Whether they downloaded 3 apps or 5 apps, they would definitely look at So-Young.” The platform model still worked.
• In 2021, the company formally decided to pivot to industrial internet, comparing the strategy to the Communist Party retreating to the countryside after losing the cities to the KMT. It strategically abandoned the platform business and shifted attention, resources and everything else toward offline.

9. The first failure: an emptied office and a Beijing wiped clean

• The most indelible scene came when he could not make payroll and told employees to take the computers. He hid in the office, then emerged to find “the computers and the people were both gone.” “No one even said, ‘Let’s hold on a little longer’… It wasn’t just the business that failed; I failed as a person.”
• His girlfriend then broke up with him. After drinking heavily, he lost his wallet and ID at the airport, while his mother helped dispose of the Beijing apartment he had bought in 2003. “Everything in Beijing was cleared out.” He took a suitcase to Shenzhen, feeling “as if heaven deliberately wanted to take everything away from me.”
• His personality split at that point: from a freewheeling optimist who might spend 5 nights a week in nightclubs to someone who took 2-3 years to recover and became low-key, steady and rigorous—“almost like 2 different people.” His conclusion: “Truly successful entrepreneurs have hearts of steel, but a heart of steel is forged through thousands of blows. The shell becomes hard; the inside stays warm.”

10. Tencent’s healing and the theory of value differentials

• He joined Tencent “partly to heal, partly to understand why this company had become so successful.” No one paid particular attention to his experience at Renren or Mop; he “honestly integrated as an ordinary employee.”
• The framework he took away was that every company creates 2 types of value: user value and commercial value. The quality of its reputation depends on the gap between the 2. Tencent delivered a 10 on user experience and a 3 on monetization, so users were tolerant; “suppose Baidu delivered a 3 on experience and a 5 commercially—users would feel over-harvested and leave.”

11. The second startup went nowhere: morale follows the business

• He joined a company run by his former boss at Qianxiang, the IM2.0 venture he mentioned, and rebuilt the Zhimei business from his first startup. The company was later acquired, but his own mandate failed. “It wasn’t especially painful; it was just depressing.” At one point he shouted at subordinates in the office: “Why is there no entrepreneurial spirit, no drive, no passion to fight?”
• His reflection came back to himself: “First, my leadership wasn’t good enough.” Team morale is “directly tied to the business.” So-Young had the same low morale during its exploratory phase; only after the data began growing and the company raised money from Matrix could it attract stronger people. Managing stronger people required “a particularly hands-off management style.”

12. The post-IPO collapse: “the stock price is a man’s self-respect”

• The host noted that the market regulator’s August 2021 Guidelines on Enforcing Medical-Aesthetics Advertising Laws banned the creation of appearance anxiety, after which the stock fell from $10 to $5.33 in 3 months. 金星 added the effects of COVID on O2O, the China-concept-stock crisis and Big Tech competition. The stock bottomed at $0.50 at the end of 2022, down 97% from its post-IPO high of $22, nearly triggering a Nasdaq delisting warning. “It lay flat on the ground for 4 full years.”
• The pain was not about cashing out: he has not sold a single share since the IPO and increased his stake from 16% to 25%. “The stock price is a matter of self-respect—a man’s self-respect.” He also felt guilty toward pre-IPO investors. The host noted that the team’s options had collapsed in value and some employees had left; 金星 said the stock price was also a major blow to the team because options were an important part of compensation.
• The 2 pains were different. The first failure was emotionally painful but rationally acceptable—“who succeeds on their first startup?” This time, “everyone thought we had already made it ashore—how did we end up falling back into the river?” He knew the pivot was the right path, but “every day you still expected a miracle… and the fantasy simply never appeared.”

13. The true darkest hour: the app was taken down without warning for 5 months

• The host said perhaps only 20%-30% of listed companies going through the same period truly fought their way out. 金星 admitted, “Did I ever think about lying flat? For a moment, I had thoughts like that,” but his principle was “to leave as a winner… How could I explain to my children what kind of experience their father had gone through?”
• Darker than the stock price was the uncontrollable event: the app was taken down for 5 full months. “No one notified us when it was taken down. Not once did anyone give me a written explanation of why.” The company even prepared to build a new app. “That was the true darkest hour.”
• He began the recovery with his body. At his worst, he was taking a handful of pills a day and had high blood sugar, high cholesterol and high blood pressure. He began exercising, losing weight, taking semaglutide and following a 16:8 fasting schedule. “The founder’s condition is the company’s condition. If I want this company to recover, I have to get myself into a better condition first.”

14. The RMB100B boast was “so embarrassing,” but stores may be evolution’s winner

• When footage resurfaced of him saying before the IPO, “Give So-Young 5 years and we’ll reach RMB100B,” his first reaction was, “Oh my God, that’s so embarrassing.” But today, “I actually believe it even more.”
• His confidence comes from arithmetic that can be seen and touched offline. “The internet really is a traffic game… somewhat rough-and-ready management. If the company loses money, it loses money; it will make money in the future, but nobody knows when.”
• His evolutionary counterargument is that the profitable store model has existed for more than 1,000 years—“could it actually be the winner of evolution? Like cockroaches, which have survived all this time?” Models built on “the wool coming from the pig” may be discarded by history. “That’s why Bezos says retail is the sexiest business model.”

15. The industrial platform defined: Didi-style decomposition of operating elements

• “We still believe we are building a platform.” Didi does not operate taxi companies; it decomposes the driver and the car, then sets prices and standards. A medical-aesthetics service consists of 3 elements: the doctor, the premises, and the equipment/products, coordinated and combined through the internet.
• The blueprint has layers. So-Young already has more than 100 doctors and aims to become “China’s largest medical-aesthetics physician group.” Clinics provide the premises, while the supply chain handles products and equipment. The platform is not merely information; “it also includes people, premises and equipment.”
• The old model’s original sin was making money from businesses while having to win consumer trust. “Once merchants paid for advertising, they wanted more exposure, regardless of whether consumers wanted to see it.” That is “the shortcoming no platform has ever been able to solve.” The new model aims for multi-party gains, consistent with his view that business models range from sophisticated to crude depending on whether the parties sacrifice for one another or move in parallel.

16. A major supply-side clearing: CR10 doubled, but the industry did not shrink

• The host cited external data showing more than 30,000 medical-aesthetics institutions closing nationwide in a year, with Beijing above 30%. 金星 pointed out that China has only a little more than 20,000 licensed institutions nationwide; the host clarified that the figures also included unlicensed operations. 金星 nevertheless confirmed the reality: “Chinese medical-aesthetics institutions are suffering widespread losses.”
• The framework is an industry life cycle: supply grows faster than demand, eventually creating excess capacity and competition. The top 10 medical-aesthetics groups now account for 25% of the market, up from less than 10% 5 years ago. Of the more than 20,000 institutions, 95% are small clinics occupying a few hundred square meters and employing only 1 or 2 doctors; many are now closing.
• The host cited China’s medical-aesthetics penetration rate rising from 4% to 6%, versus 22% in South Korea. 金星 said upstream companies had grown from 100-200 5 years ago to at least 1,000 today, and new entrants with R&D capabilities would drive prices lower. The host observed that light medical aesthetics is becoming routine anti-aging maintenance. 金星 said home devices are only a supplement: “The energy is orders of magnitude different from what medical-aesthetics institutions use.” Ultimately, it remains a medical procedure.

17. Countering “low prices are just customer acquisition”: a new species targeting Korea’s 25% net margin

• Responding to the host’s claim that high-end medical aesthetics serves the middle class and above, while low prices are merely a customer-acquisition tactic before upselling premium services, 金星 laid out the cost math. Well-run Chinese institutions earn about a 5% net margin, and Lancy has disclosed the same; Korean light-medical-aesthetics chains charge an average of KRW100,000, or about RMB550, yet achieve a 25% net margin.
• The gap comes from 3 factors: customer-acquisition costs account for 30%-50% in China, products are more expensive, and operating efficiency is lower. A Korean chain in the same footprint may see 10x China’s monthly traffic, spreading fixed costs over one-tenth the base. So-Young’s sales productivity per square meter is 3x that of ordinary Chinese institutions, and “mature stores have net profit margins above 10%”—better than most institutions claiming to operate in the high-end segment.
• He rejects the positioning: So-Young is not Mixue Bingcheng and should “absolutely not become a discount supermarket.” What China most lacks is a Sam’s Club or Fat Dong Lai-style model of “high quality at low prices.” Users feel smart—“I spent only RMB1,000, while you have to spend RMB5,000 for the same procedure”—and spread the word themselves. He says the store environment, service, products and doctors are all higher quality; while they cannot be compared directly with top specialists, they can reach at least 80%-90% of that standard.

18. Density and offline traffic: malls are giant traffic-concentration points

• Responding to real examples of people flying to South Korea for cheaper procedures, he admitted that some services still cannot match Korean prices. But light medical aesthetics is performed once every 2-3 months, making density decisive. So-Young already has 7 light-medical-aesthetics clinics in Beijing and plans to reach 10 by year-end. “We hope that in the future users can find one within a 3-kilometer radius, or reach one in a 15-minute taxi ride.”
• South Korea points to the end state: as light-medical-aesthetics chains became highly developed, “everyday beauty services in Korea have basically shrunk to almost nothing.” In China, offline traffic is heavily concentrated in malls. Some malls span 1 million square meters and “become giant concentrations of offline traffic,” potentially making offline customer acquisition cheaper than online.

19. The 奇致激光 acquisition and the “Xiaomi of upstream” thesis

• So-Young paid RMB791M for an 84.9% stake in 奇致激光. Its founder is regarded as China’s “first person in photofacials”; the company has operated for 25 years, contributes tens of millions of RMB in annual profit and holds a substantial portfolio of Class III device approvals. “Good light-based equipment companies are scarce—extremely scarce.” The deal closed in 1-2 months. As for why it was not the market leader: “The best one definitely won’t sell to you.”
• His multi-year observation is that upstream companies command high public-market valuations—IMEIK and Jinbo Bio are examples—but still rely on highly traditional, relationship-driven salesforces. M22 sold 600 units in China in a year, already a historic peak. High prices limit volume, and low volume keeps prices high: “It’s a vicious cycle.”
• The opportunity is to build “the Xiaomi of upstream.” The high price of large equipment is not driven by the hardware: “The number of components may be smaller than in an iPhone—an iPhone may contain 10,000 components.” Consumer marketing could drive volume; equipment could even be provided free, with fees based on usage or revenue share. That would put it within reach of more than 10,000 small and midsize institutions and create the possibility of shipping several thousand units a year.

20. 新氧臻品 and the bet on 4 injectables

• 新氧臻品 combines customization with centralized procurement: So-Young customizes and promotes the products itself, securing near-cost pricing in exchange for annual purchases of at least 100,000 units, and sometimes 200,000-300,000. Most are domestic products; regenerative materials are mainly produced by Chinese companies, and every product must have the required approval.
• The industry’s rule is to sell commercialization rights when a product enters clinical trials. A company may pay tens of millions of RMB for exclusive distribution rights; if the clinical trial fails or the registration certificate is not granted, the money is refunded. In 2023, So-Young signed rights agreements for 4 major injectables, including next-generation 童颜针, extracellular-matrix gel and silk-fibroin protein products. Their registration certificates are not expected until 2026-2027, by which time the store network should be at the right scale.
• The risks include the partner failing, choosing the wrong products and sinking time costs. 金星 acknowledged that the first test is “our judgment.” He looks at products already validated overseas, but demand does not always align between China and overseas markets, so the company still has to make its own call. It also runs small-scale tests before broad clinical trials to reduce risk. Upstream investment has exceeded RMB1B over 4 years; cash peaked at RMB2.8B and stands at RMB1.1B today. “We haven’t spent a penny of the money raised after Series C, because we’ve been profitable ever since.”

21. The 1,000-store engineering plan: build the middle platform first, then move cautiously from 0 to 1 and quickly from 1 to 100

• So-Young addressed the “how do you build a middle platform without stores?” paradox in 2022 through the virtual chain 新氧优享, operating shop-in-shop locations with nearly 200 institutions. So-Young set the product lineup and prices, defined the SOP and handled customer acquisition; the institutions handled only medical delivery. “I didn’t have a single store of my own, but I already had 200 virtual offline stores.” When asked whether partners might steal its customers, he replied: “It was a transitional phase. I could accept it; I established my own middle platform.”
• The pace was deliberately conservative. It opened its first store in 2023, then waited 10 months before opening the second to fully validate the model. It went from store 2 to store 20 in another 10 months, opened 20 stores last year and plans 30 this year. “To be honest, I have the financial capacity, but I would feel that it was unsafe.”
• Franchising will use a fully managed model and will not accelerate openings, because the management difficulty is essentially the same as direct operation. The company plans to pilot 3 stores this year before deciding the mix. “We are an extremely cautious and conservative company, but once we prove that something works, we become very bold.” An internet-hospital license also enables compliant online consultations.