Latest Thinking · FAQ

Using the line "Only Apple Can Do" that swept the market after the foldable iPhone Duo debuted at Apple's September 10, 2026 event, the article argues that Apple is not a typical product innovation company but more like an experience innovation company. Every parameterizable metric of a foldable — thickness, crease, battery life, chip — was pushed to its limit by Android makers years ago. Apple's difference is not a single breakthrough but a different logic of innovation: most companies add on the product dimension (more features, stronger specs), while Apple subtracts from the user's point of view (less operational cost, less cognitive load, fewer experience breaks). The conclusion: specs win a comparison, experience wins a long-term relationship.

Because users never pay for specs themselves: 120W fast charging is not the need — "never worrying about battery" is; a 200-megapixel camera is not the need — "a great shot any time" is. Every technical spec has to be translated into context and turned into a perceptible experience before its value is actually realized. Most companies' innovation stays at the level of technical language and never completes the translation into user value — which is exactly why specs keep getting stronger while users find the difference harder and harder to feel.

The article points out that AI has sharply cut the cost and cycle time of code development, feature design, and capability replication: a feature innovation used to hold a lead for about a year, but now it can be matched in months, even weeks. So the most dangerous rat race of the AI era is every company producing increasingly similar products at greater speed. As features and specs become easy to copy, what becomes hard to replicate is the complete experience co-created by hardware, software, chips, ecosystem, services, and organization working in concert — what makes "Only Apple Can Do" hard was never one technology, but getting the entire system to serve a single user experience.

The transmission chain the article gives: sustained experience creates user habits → habits settle into trust → trust converts into willingness to repurchase, tolerance for a price premium, and ecosystem lock-in → a long-term business moat. Specs depreciate quickly as technology iterates; the habits and trust built through long-term experience do not. So the unit of competition for innovation is upgrading: from "what can I do" (creating new features), to "how much better am I than others" (spec contests), to what will truly matter in the AI era — "what does the user actually feel" (differentiated experience and a sense of value). Product innovation asks what else can be added, experience innovation asks how much friction can be removed, and business innovation asks whether this experience can become a long-term relationship.

This article analyzes the deep structural reasons behind online platforms repeatedly engaging in monopolistic practices, using Ctrip’s ¥5.179 billion antitrust fine as a starting point. It argues that the vicious cycle of pure online information-asymmetry platforms squeezing both merchants and consumers stems from the platform’s business model itself, not just the morals of individual companies — and the only way out is rebuilding commercial structures founded on symbiotic value creation.

Pure online platforms do not own core supply (hotels, restaurants, vehicles) and bear no offline delivery costs. When information asymmetry fades and traffic growth stalls, they resort to rule manipulation and merchant squeezing to maintain profitability. Fine one platform, and another will inevitably fill the gap and repeat the same story.

People earn offline but spend online chasing rock-bottom prices. Offline businesses bear all fixed costs but are forced into online price wars, while platforms capture profit without delivery burden. This squeezes margins, depresses employment, and drives consumers toward even lower prices — a self-reinforcing negative spiral.

Platforms must build moats on real delivery capability rather than rule design. Merchants need to invest in their own brands and private domains. Consumers should acknowledge that genuine quality has a cost. Only when all parties return to their proper roles can the industry escape the deadlock.

"A kind word warms three winters; a harsh word chills even June." In January 2026, a pre-made-food controversy that began in September 2025 escalated again: the long-range feud between Luo Yonghao and Xibei founder Jia Guolong ended with both Weibo accounts muted simultaneously. Even muted, both kept posting on WeChat Moments: Luo confirmed a 15-day ban, chose to "read it kindly as protection", and said he would not respond further; Jia complained that "Xibei suffered a cliff-edge loss while Luo's livestream sales quintupled", saying the controversy piled misery on an already battered Xibei. A discussion that should have centered on consumers' right to know and the definition of standardized food degenerated, over four months of emotional confrontation, into a public-resource-draining shouting match.

True emotional stability is not "never getting angry" but "seeing the emotion, accepting it, then responding rationally". It differs fundamentally from swallowing resentment: the former is active mastery, the latter passive suppression — suppressed emotion eventually erupts more violently, while mastered emotion becomes an asset. Emotional stability is never about holding your tongue in humiliation; it is the clarity of "a kind word warms three winters; a harsh word chills even June". Only by guarding the bottom line of emotion can we preserve the warmth of relationships and the scope of a life. As Xunzi wrote in "On Honor and Disgrace": "Kind words to others are warmer than cloth;"

When we lose emotional control, we are in fact being hijacked by a kind of "subconscious"—in psychology it is called the ABC principle; put plainly: what happens (A), and how we think about it (B), determines what we do (C).

The Luo-vs-Jia feud escalated precisely because both fell into automatic emotional loops: Luo construed Xibei's failure to label pre-made dishes as "deceiving consumers" (B) and responded with public call-outs plus a bounty for evidence (C); Jia construed Luo's public questioning as "malicious slander" (B) and responded with hardline counterattacks and threats to sue (C). This loop of "stimulus → negative cognition → emotional confrontation" turned an industry issue fit for rational discussion into a personal war. Life inevitably brings setbacks and disagreements; the emotionally stable stay clear-headed through the swings.

"The law does not believe in tears, business has no mercy for pathos, and human nature cannot withstand testing!" Li Yinuo is an education entrepreneur and force for good whom I deeply respect. On the surface of the ETU dispute, the law is never a warm mediator but a cold arbiter of rules: it recognizes only objective facts and explicit contracts — not tears, not ideals, not dreams. And what happened to ETU is the abyss that every dream-driven innovator struggles to cross.

The collapse of Yitu Zhizhi School had its answer written long ago: when Huazhang, the legal representative, was ruled to bear joint and several liability for a 50 million yuan loan, none of it mattered—whether her decision-making power had been hollowed out, how pure Li Yinuo's educational dream was, or how tragic the team's move to mortgage their homes to fill the gap. The iron law of "rights and duties defined by law" could not be changed. The Zhizhi side misappropriated the loan and the school stalled; personal assets faced liquidation. The root of this tragedy is singular: replacing law with sentiment, and testing human nature with trust. The only code to lasting business has always been "law sets the boundaries, rules guard against human nature, settle terms first and you stay friends longer."

No business tragedy erupts from a single mistake; it is the compounding of two failures — "no legal boundaries" and "blindly testing human nature". ETU's collapse is the concentrated eruption of four such traps. 1. Trap one: legal mismatch of power and liability + betting on "the other side being ethical" leaves the "fall guy" no room to argue. To secure a school license and stable premises, Li Yinuo's team made a fatal compromise: Huazhang became the legal representative bearing unlimited joint liability, yet the corresponding decision rights were never written into legal documents — the ETU side held the board majority, and core matters like large fund transfers and loan collateral required no consent from the liable party. The legal principle that power and liability are defined by law does not lapse because someone was "sidelined"; an oral agreement of "liability without power" is worthless in court, and Huazhang ended up the prime target of debt recovery.

"NVIDIA is essentially a platform computing company!" On the evening of October 28, at the GTC conference, Jensen Huang spent two hours painting a detailed blueprint for an AI-driven new industrial revolution. Huang is a devout believer in first principles, having spoken on many occasions about using them to think through problems and run the company. In what follows, we use first principles to read NVIDIA's strategic blueprint.

There are two typical views of NVIDIA: some see a hardware company focused on high-performance GPUs, dominating the global market through superior compute; others stress its software identity, arguing that the CUDA-centered programming platform and ecosystem are its real moat. Jensen Huang has stated clearly that NVIDIA is neither purely hardware nor software but a "platform computing" company that exists to solve computing problems. This positioning has run through NVIDIA's history: from the start, every strategic move revolved around it. Its early GPU bet came from the insight that general-purpose CPUs faced severe performance bottlenecks in graphics rendering and complex scene simulation, and a dedicated computing architecture was needed to break the limits.

Huang's twin-engine direction of "computing paradigm + industrial form" is, in essence, a precise lock on the three core computing pain points of the next 5-10 years, drawing the problem space for NVIDIA's strategy: breaking the compute ceiling — with Moore's Law fading, general CPUs cannot support the exponential compute demands of AI training and quantum simulation (large-model training requires trillions of operations). Hence the shift to "GPU + AI accelerated computing", replacing general architectures with dedicated ones — hardware-form innovation to break through traditional compute limits;

Every product in the announced full-stack hardware-software matrix is a tailored answer to a computing pain point in the new direction — none strays from "solving computing problems" — together forming a complete toolchain covering compute production, compute adaptation, and compute application. Hardware: breaking the physical limits of computing — the Blackwell platform (GB200 NVL72) achieves 10x performance through extreme co-design while cutting token-generation cost, addressing the energy-efficiency shortfall of large-scale AI computing; NVQLink connects QPUs and GPUs at 4-microsecond latency, solving the delay problem of quantum-classical collaboration — both use hardware innovation to push past physical limits of performance and latency.

"The first principle of competition: rather than being better, be different." Breaking the delivery-war deadlock: JD Delivery is piloting daily professional disinfection of full-time riders' delivery boxes, so riders carry clean boxes to every customer. This signals the delivery war's turn from "price slaughter" to "service breakout".

Once upon a time, delivery-platform competition was like same-brand milk fighting with buy-one-get-one-free: whoever subsidized more grabbed traffic for a while. But since 2025 the unsustainability of this model has become stark, forcing the whole industry to seek a new way out. On the platform side, low-price strategy has hit the double red line of profit and policy: China Chain Store & Franchise Association delivery-industry data from October 2025 shows 75% of incremental delivery orders were paid at under 15 yuan, and sustained subsidies trapped leading platforms in "traffic up, profit down". On the merchant side, low-price orders crush margins: a chain fast-food executive revealed in an interview that platform commission rates run 25%-30%, and after ingredients, packaging, and delivery, low-price orders net less than 5% margin.

Service competition is not empty talk detached from the business — it is detail-level optimization of the whole "order-fulfillment-aftersales" chain around user needs. Making service perceivable and quantifiable is becoming the new moat. The service turn is not only changing the competitive landscape but reshaping the industry's underlying value: it moves users, merchants, and platforms from mutual attrition to shared gains — all effects grounded in real data, no fictional forecasts. For users, "fair price + quality service" becomes the norm: when users willingly pay a premium for box disinfection and after-sales guarantees, a positive cycle forms — users pay, merchants profit, service upgrades again. Users no longer have to trade off between "cheap" and "good".

JD Food Delivery leverages Dada Logistics and JD Supermarket resources to create a "instant retail + food delivery" synergy service. When users order food in the delivery channel, they can simultaneously buy supermarket daily necessities, achieving "pickup within 5 minutes within 3 km, delivery within 1 hour," while sharing fulfillment capabilities such as "meal-box disinfection" and "temperature-controlled delivery." Alibaba links Taobao Instant Shopping with Tmall Supermarket to offer "food delivery + online membership" services for brands like Watsons and Starbucks. Users earn brand points when ordering via delivery, and repurchase rates rise 50% over pure offline consumption. This "ecosystem synergy" turns the platform from a "single food-delivery tool" into an "all-scenario life entry point," strengthening user stickiness.

"The essence of a business problem is optimizing the path under a shared goal — not an either-or choice." When news broke of Pang Donglai's 20,000-yuan-per-person visit fee, the retail industry quickly split into two camps: supporters saw it as a tool to "filter for aligned learners and purify the learning ecosystem"; skeptics pointed at "a bar too high for small businesses and insufficient cost transparency".

(1) The previous model (free assistance) fit the consensus on "transmitting the essence" but strayed from the consensus on "efficiency". When Pang Donglai helped a Yonghui supermarket store in 2025, it used a one-on-one, in-store overhaul: three senior executives stationed for three months, landing "pay raises (junior staff monthly salary from 4,500 to 6,000 yuan) + service upgrades (a new urgent-restock service)". The store's sales rose 210% and labor efficiency 58% — successfully transmitting the "employees first" philosophy.

A close reading of the debate over the "20,000-yuan visit fee" shows the two sides, though seemingly opposed, are working toward the same goal and share three underlying consensuses. Consensus 1: reject "fragmented model-copying" — the "essential philosophy of healthy operations" must be transmitted.

Supporters believe official visits can crush third-party "high-fee, low-value" harvesting — some agencies charge 30,000 yuan per person yet offer only store tours, photo ops, and boilerplate handouts, transmitting no actionable logic. Skeptics worry that once officially monopolized, philosophy transmission degenerates into standardized lecturing — executives reading scripts, staff posing for photos, with visitors still learning no real operational detail. The core contradiction of the whole controversy: under the shared goals of transmitting the essential philosophy, ensuring learning efficiency, and purifying the ecosystem, how to solve the path-fit problems of covering small businesses, making costs transparent, and diversifying the ecosystem. Any concrete design must balance "filtering aligned learners" with "covering diverse needs", "cost balance" with "transparent trust", and "official leadership" with "ecosystem openness" — rather than satisfying a single dimension of value.

"In ability and seniority, Cao Hui is indeed no match for some executives who fought alongside me for years. But he is Fuyao's major shareholder, and the root that binds this family to the company. What I am handing over is not the chairman's seat — it is the product system and management institutions Fuyao has built over decades." On October 16, 2025, Fuyao Glass announced its chairmanship succession. Unlike most companies' secrecy around successors, and unlike a simple "son inherits father's business" story, Cao Dewang pushed his son Cao Hui to the front without hiding his weaknesses or mythologizing his strengths — sketching a long-planned strategic succession with three keywords: "ten years of grooming", "handover at the peak", and "institutional continuity".

The common predicament of succession in Chinese private enterprises often stems from the "passivity of timing"—handing over in haste because of the founder's health, or "dumping the burden" when performance declines, leaving the successor mired in the infighting of "cleaning up a mess." Cao Dewang's choice broke out of this frame: he stepped down voluntarily while the business was performing well, creating an operating environment for Cao Hui that had "no performance pressure and institutional support." The formation of this situation grew out of a complete decade-long groundwork, from "cultivating the person" to "validating the capability."

Cao Dewang chose to retire while Fuyao Glass was in a period of steady growth. In 2015, when Cao Hui became vice chairman, revenue was around the 10-billion-yuan level; by 2024 it had climbed to nearly 40 billion, with net profit rising in step. A decade of sustained growth meant the succession faced no "firefighting" pressure — Cao Hui inherited not just a title but a market-tested operating system. Compared with private firms that force successors on stage amid earnings turmoil, Fuyao's "handover at the peak" is in essence a transfer of confidence: employees see careers unshaken, partners see policy continuity, and the market sees an unchanged growth logic with no succession risk to fear.

Cao Hui's succession was not a "parachute coronation" but the result of staircase growth. Appointed vice chairman of Fuyao Glass in 2015, over the following decade he moved from observer to core participant, deeply involved in globalization and product-line upgrades. In globalization, he led the landing and operation of Fuyao's US plant: facing early challenges in cultural fit and supply-chain coordination, he stationed himself on site to push "local procurement + lean production" reforms, eventually driving the plant to profitability and making it a pillar of Fuyao's overseas business.

On October 14, JD teamed up with GAC and CATL to launch a "National Car" — is this cross-industry move a feast of traffic, or an experiment in reshaping the auto industry centered on user needs and powered by ecosystem integration? JD's car-making strategy extends its digital DNA from e-commerce while breaking the rigid division of labor in the traditional auto chain through an "asset-light + strong-synergy" model; in essence it is an ecosystem rebuild of the entire car-consumption lifecycle.

JD's entry this time did not replicate the single "e-commerce selling cars" model; instead, by integrating verified user data, supply-chain capability, and industry resources, it built a closed loop covering "demand insight – product definition – sales service – energy replenishment."

Relying on 600 million annual active users and 20 million car-owner data points, JD advances product design through a C2M reverse-customization model; users can participate in model voting and configuration selection on the JD app. As of October 2025, over 20,000 users had completed requirement forms, focusing on family-user needs such as "flat rear floor + smart cockpit." Based on this data, the new car is set to adopt CATL's Choco-Swap battery-swap technology, achieving 5-minute ultra-fast replenishment and solving users' range-anxiety pain point.

"If you are not willing to fail, your innovation is not radical enough." SpaceX's Starship completed its 11th test flight at 18:23 US Central Time on October 13, 2025 (07:23 Beijing time, October 14) from Starbase in Boca Chica, Texas, achieving for the first time a full closed loop from launch to controlled splashdown — a successful finale to the second-generation Starship's development phase.

Extreme heat-shield testing: the ship deliberately flew with some heat tiles removed, especially in vulnerable zones without backup ablative layers, to verify tolerance under extreme heat. Live footage showed surface temperatures exceeding 1,427°C on re-entry with no damage to key structures, proving the improved tile-sealing technology works. Engine configuration and redundancy: the booster used a new landing-burn strategy — igniting 13 Raptor 2 engines to decelerate, then switching to 5 for lateral maneuvering — providing redundancy against future engine failures. This design will go directly into the third-generation Starship (Block 3). Dynamic maneuvering and guidance optimization: in the final re-entry phase the ship performed a "dynamic banking maneuver", simulating the complex trajectory control of a future return to the launch site and testing the precision of subsonic guidance algorithms.

Starship's breakthrough was never the single technical victory of "a super-heavy rocket's maiden flight" — it is a rebuild of the underlying logic of how innovation happens. It broke the aerospace industry's habit of "must start with zero defects", proving with "$300 million per trial" and "48-hour rapid iteration" that what matters more than avoiding failure is building a mechanism to learn from failure efficiently. 1. The conceptual pillar: from "avoiding failure" to "deliberate trial and error" — the first 10 flights' "nine defeats, one victory" were not accidents but boundary tests (removing heat tiles on purpose to verify heat tolerance; shutting down engines deliberately to test redundancy). First principles in action: strip away industry habit and return to physics (abandoning carbon fiber for 301 stainless steel — 50% stronger at cryogenic temperatures at 1/50 the cost). Vision-anchored long-termism…

From the essential choice of abandoning carbon fiber for stainless steel to the courage of "exploding for data," the core lesson reaches far beyond aerospace: for all innovators chasing breakthroughs, the real barrier is never technical difficulty but the mental shackle of "not daring to trial-and-error." Starship's practice is the best annotation: exploring the unknown needs no "perfect plan" — move forward fast with the problems in hand, and let every iteration become a step closer to the goal. This is the most precious underpinning of humanity's breaking of boundaries.

"A great company's endurance depends on institutional systems, not personal authority." On September 12, 2025, Zong Fuli formally resigned as chairwoman of Wahaha Group after a tenure of only 378 days. The data decline during this brief period bluntly exposed the fatal flaw of "personal handover": terminal outlets plunged from 5 million to 80,000, AD Calcium Milk sales in East China fell 37%, and purified-water market share shrank from 18% to 12%. Wahaha's plight is essentially a clash between "accidental success" and "inevitable longevity": it rose on Zong Qinghou's personal judgment and the dividends of an era, but under family-style governance never built a complete institutional system — hence today's fragmentation.

Wahaha in the Zong Qinghou era was built around the founder's personal prestige, without a sustainable succession system. After Zong Fuli took over, the fragility of this person-dependence erupted into systemic crisis. (1) The institutional defect in the equity structure

In sharp contrast to Wahaha, New Hope Group achieved a smooth founder-to-successor transition by building institutions in advance, showing how institutional design underpins corporate succession. (1) A staircase design for the transfer of power

The succession practices of Wahaha and New Hope show that building a systematic institutional system is the key for private enterprises to break through the succession predicament. This requires a governance ecosystem in which the founder, the successor, and the enterprise jointly exert force.

The platform makes the star, traffic pays the bill, the locality benefits; he is not a "grassroots comeback" but an "ecosystem instrument" selected by the platform. Mention "Fried Chicken Bro" and you will likely be moved by this narrative: a 48-year-old street vendor rooted in Jingdezhen for over 20 years, running his snack stall for 9, went viral through the sincerity of "6-yuan fried chicken served with emotional warmth", was named a "cultural-tourism ambassador" by the local tourism bureau, and boosted National Day tourism. The public reads it as an inspiring tale of an ordinary man rising through honesty — a heartwarming case of "authenticity beating traffic tricks".

Do you really think this is all thanks to "sincerity"? Why did Fried Chicken Bro sell chicken cutlets in Jingdezhen for 9 years without going viral, only to blow up exactly 7 days before National Day? Why could the culture-tourism bureau's supporting services move faster than an influencer gains followers? Is it really just a "grassroots underdog story"? Fried Chicken Bro's fame is no accident at all — it is a carefully laid "open scheme" by the platform to push its second curve of "local life"!

To understand this open scheme, you must first grasp why the platform "switched tracks": the leap of its business logic from the "first curve" to the "second curve" is a proactive choice in step with the times.

In earlier years, platforms made money on the "first curve" of live commerce, whose core logic was "quick money from emotional stimulation": find IPs that generate high contrast and strong emotion, then sell goods and collect tips. Take Wen Huijun, who exploded in 2023: a positive, hardworking father in the morning, a versatile singing-and-dancing uncle at noon, and a gloomy "decadent literature master" at night — grabbing eyeballs through that split persona. In December 2023 his main account "Wen Shen" grew from 210,000 to 4.77 million followers in 10 days, two livestreams drew over 55.2 million cumulative views, and at his peak a single stream received 6,000 "Carnival" gifts. But this model has a fatal flaw: it leans too hard on short-term emotion — once the novelty fades, it goes cold.

"The pain of switching curves is the cost of growth, not a signal of failure." Recently, Huawei Cloud's reorganization drew wide attention — cutting the EI (Enterprise Intelligence) product line and consolidating around a "3+2+1" system (three basic capabilities: general computing, AI computing, storage; two core platforms: AI PaaS and databases; one safeguard: security). Unofficial reports put the changes at nearly a thousand staff. Many read it as "business contraction", but looking through the surface, this is not a decline signal — it is the inevitable pain of moving from the "cloud-computing second curve" to the "AI third curve". Such pain is not unique to tech giants; it is a strategic question every company crossing a technology cycle must face.

Before judging Huawei Cloud's move, we need the industry's underlying logic: technology waves succeed one another as growth curves replace each other — the internet era centered on "information-connection efficiency", the cloud era on "computing-delivery efficiency", and the AI era turns to "intelligence-generation efficiency". For companies, the core contradiction of curve switching is not "whether to transform" but "how to migrate old-curve resources (technology, teams, customers) efficiently to the new curve". This is the first principle of technology-driven companies: the essence of business growth is resource-reallocation efficiency, and the pain of switching curves is the inevitable cost of old resources mismatching new demand.

Huawei Cloud's adjustment is not "blind transformation" but is driven by three real constraints — profit targets, industry competition, and capability boundaries. These three constraints are also checkpoints every company must pass when transforming toward AI; it is only that Huawei Cloud, as an industry front-runner, faced and addressed them earlier.

In 2024 Huawei Cloud was still in the red, and "achieving profitability" was set as a clear core goal for 2025. This goal forced the business to shift from "scale coverage of a single curve" to "cross-curve value focus": previously Huawei Cloud's business spanned Cloud EI tools, edge computing, large AI models, and more, with some businesses (such as basic algorithm tools) facing dual pressure from "customer self-substitution" and "low-price competition," and continuously declining input-output ratios; this disbanding of the Cloud EI department and focus on the "3+2+1" system is, in essence, divesting "non-strategic assets" and concentrating manpower, capital, and other resources on high-value areas such as AI computing and AI PaaS — by narrowing the business boundary, it improves the precision of resource investment and shortens the profit cycle.

"This is a classic war of misaligned value perception: the company's efficiency logic and the user's experience expectations are simply not on the same channel." "Paying 830 yuan for a meal that tastes like day-old pre-made food — anyone would complain." In September 2025, a Weibo post by Luo Yonghao pushed Xibei into the storm: five diners ordered 13 dishes, the lamb chops smelled gamey and the stew arrived lukewarm. He called it "expensive and disgusting" and offered a 100,000-yuan reward for evidence that Xibei uses pre-made dishes.

Many reduce restaurant competition to "wok-fresh vs. pre-made," but miss a more fundamental logic: what users pay for has never been "wok-fresh" or "pre-made" itself, but a combination of "functional value + emotional value." This is the first principle of the food-service industry: "the money / time a user pays must match the dual expectations of 'practicality (function)' and 'psychological satisfaction (emotion).'"

More crucially, for different categories of dining, users' expected weighting of "dual-dimensional value" is completely different. For example, when you eat at McDonald's, the default is "convenience first" and you accept 90% pre-made semi-finished products; but when you spend 80 yuan at Xibei, the default is "fresh-made first," and even mere raw-material pre-processing feels "not worth it." This is the "expectation baseline" that category attributes set for users — once this baseline is breached, controversy is bound to erupt.

Xibei's problem has never been "the central kitchen doing raw-material pre-processing," but "charging high-end dine-in prices without meeting the emotional-value expectations of high-end dining" — specifically, it stepped into three pits:

"Amap's non-commercial breakthrough with its Street-Scan Ranking hits Meituan's fatal weak spot!" In September 2025, on Alibaba's 26th anniversary, Amap quietly launched the "Street-Scan Ranking" — billed as the world's first ranking of offline services based on real behavior — pushing Meituan-Dianping into a corner. Some call it "a new platform defeating an old one", but the truth is not that simple.

As many platforms grow in scale, they fall into a trap: believing that "many users, many merchants, big subsidies" is an unshakable moat. The moat of platform economics has never been these "surface data," but a "closed loop of underlying capabilities" built around the "ultimate value of platform economics (user value + merchant value + ecosystem sustainability)" that rivals cannot easily copy. It must satisfy three core characteristics, none of which can be missing:

A true value anchor is the long-term pain point of users / merchants — for example, users' need for "real recommendations" (fear of bad experiences, fear of fake reviews) and merchants' need for "low-cost customer acquisition" (not being held hostage by traffic). It is not the platform's short-term goal of "how much to earn this quarter." Like Amap's Street-Scan Ranking, which from the start anchored on users' need to "find genuinely good stores," rather than rushing to profit from paid rankings — that is the key to its rapid user appeal.

A true core capability is the underlying logic plus technology / data barriers that "rivals cannot learn or copy." Take Amap's "navigation behavior data": with 13 million daily life-service navigation records, behaviors such as users "navigating specifically to a store from 5 km away" or "repurchasing more than 3 times" simply cannot be faked through brush orders the way Meituan's "reviews" can. By contrast, Meituan's "scale advantage" can be siphoned off by JD with "zero commissions" or by Amap with "real recommendations" — the essence is that its core capability lacks "irreplaceability."

"The first principle of innovation is user needs, not feature stacking!" DJI and Insta360 have finally met head-on: Insta360 launched a panoramic drone, DJI launched a panoramic camera. Who do you favor? I own products from both, and I personally favor Insta360 — why?

On the surface, the competition centers on specs (stabilization, sensor size), features (underwater shooting, AI editing), and market reach (online channels, KOL partnerships). The deeper conflict: DJI believes "stable hardware + full-scenario coverage" is the core, leaning on its flight-control heritage (gimbal control, obstacle-avoidance algorithms) to build a moat via "sky-to-ground" technology migration.

DJI's moat comes from migrating its flight-control technology. The gimbal control (Zenmuse series), obstacle-avoidance algorithms (OcuSync transmission), and battery optimization accumulated in its drone business naturally fit action cameras' hard need for stable footage. The "RockSteady" stabilization of the Action series is essentially drone attitude-control algorithms miniaturized; omnidirectional obstacle sensing is industrial drone perception adapted downward. The advantage of DJI's route is certainty: users get stable footage out of the box (cycling follow-shots, diving) without learning complex settings. The limitation: hardware integration carries high marginal cost (custom hardware per scenario), and over-reliance on spec leadership can neglect users' need for flexible creation (multi-angle shooting, quick editing).

Insta360's underlying logic is "turn users into creators, not recorders". Its core technologies — 360° panoramic stitching, AI subject tracking — essentially move post-production forward to the moment of shooting. The One RS modular lens design lets users switch between wide-angle, 360°, and 1-inch sensor modes with one click; the FlowState stabilization algorithm corrects footage in real time with AI, eliminating re-cropping in post; the "invisible selfie stick" removes the stick from the frame algorithmically, outputting cinematic camera moves directly. The advantage of this route is freedom: users produce high-quality content (vlogs, travel shorts) without professional editing skills.

"The delivery war will eventually shift from a zero-sum price game to creating incremental value." Wanting coffee one morning, I flipped through three apps — Meituan, Ele.me, JD — and the pages looked copy-pasted: Meituan "30 minus 15", Ele.me "20 off your first order", JD's new-user coupon cutting a latte to half price. It is like three supermarkets selling the same milk, all fighting with "buy one get one free". After JD entered, why can't the delivery war escape the subsidy-for-traffic loop? And where is the way out?

The essence of food delivery is "getting the goods the user wants, at an acceptable price, within an acceptable time, to a designated place" — and the core is one thing: fulfillment efficiency. Deconstruct the platform's logic: it is a hub connecting merchants, riders, and users, with a value chain of just three links — "findable (merchant coverage), fast (rider efficiency), worth buying (user deals)".

The delivery war among Meituan, Alibaba, and JD is, in essence, homogeneous competition among platform companies. On the surface the fight looks lively: JD patches its instant-retail gap with supermarket supply chains, Meituan defends its "delivery + in-store" base, and Alibaba's Ele.me widens scenarios with supermarkets and pharmacies. But the core contradiction is not "whose service is better" — it is that platform players hold nearly identical weapons and cannot find a differentiated breakthrough.

Delivery "quality" is the merchant's supply capability — taste, packaging hygiene, ingredient freshness are all decided by the merchant. The platform can at most add a "safe to eat" label; it cannot directly intervene. The same Sichuan restaurant sells the same yu-xiang pork on Meituan and JD Daojia with no difference in taste or portion; many merchants operate on multiple platforms, so platforms simply cannot differentiate on "quality". Moreover, if you sign KFC, I can sign McDonald's; if you land a local supermarket chain, I will recruit the convenience store next door — no one can monopolize exclusive supply.

"Every viral phenomenon is a precise resonance between user needs and business efficiency." The Little Monster of Langlang Mountain broke 500 million yuan at the box office within 7 days of release; the Douyin topic "I want to leave Langlang Mountain" surpassed 8 billion views; Luckin's tie-in coffee sold 100,000 cups in 3 hours; character blind boxes traded second-hand at 2x premium. Behind these numbers lies a key question: why can a "pot-scrubbing little pig demon" move both children and office workers? And why can an animation IP extend from cinemas into coffee and stationery?

The core logic of the animation industry has never been "the more polished the visuals, the better", but two inseparable fundamentals: emotional resonance and content innovation. Emotional resonance means precisely matching users' emotional needs. Animation is essentially "characters and stories carrying unmet emotions": children need the immersion of "growing brave", office workers need the comfort of "breaking through hardship". The Lion King matched the public's struggle with responsibility through Simba's arc from escape to duty, ultimately earning $1.04 billion — the power of resonance.

The resonance of Langlang Mountain comes from turning workplace pain into the daily life of little monsters: the pig demon "studying three years to enter the Great King's Cave" mirrors civil-service exam takers; "scrubbing pots until his fur falls out, yet still denied" feels exactly like an employee's all-night proposal being rejected; the toad demon "clutching his work badge while fleeing" maps to the need for security of those who complain about work but dare not quit. The team nailed three workplace pains — effort-reward mismatch, lack of security, wanting to break out yet hesitating — letting viewers see themselves in the little monsters. That mirror feeling became the engine of organic sharing and paid viewing. Perhaps everyone needs to learn, through the trials of "Langlang Mountain", to distinguish "the value of the tool" from "the glory of the creator". The world beyond the mountain may still be mountains, but between the mountains lies the gulf of self-definition.

The contest between efficiency-for-everyone and quality-privilege is redefining how Chinese consumers shop. In August 2025, two landmark events hit Chinese retail: JD's discount supermarket opened its first store in Zhuozhou, Hebei, drawing over 60,000 visitors in a single day — an industry record; Sam's Club, after delisting 200+ exclusive items and bringing in mass-market brands, saw its renewal rate fall below the 90% red line for the first time. This is not simply "low price beats premium" but a head-on collision between the "efficiency" and "quality" routes, revealing a rebuild of retail's underlying logic. Every retailer sits on a two-dimensional value axis: the horizontal axis is efficiency (cost, speed, coverage); the vertical axis is quality (exclusivity, quality control, experience premium).

The essence of retail has never changed: deliver the best products to the people who need them most at the lowest cost. But "the best product" means different things to different people, and "the lowest cost" varies with the goal — the difference in JD's and Sam's Club's "people, product, place" comes down to anchoring differently on the "efficiency–quality" axis.

Big data shows that price sensitivity for staples among lower-tier-market consumers reaches 78%, but their attention to fresh-food quality is 23% higher than in first-tier cities — they want "low prices," yet fear "cheap but poor quality" even more. So JD's Suqian store adopts "farm-direct sourcing + community group buying": vegetables picked from Shandong bases in the morning reach the shelves by noon, with twice-daily restocking to ensure freshness, yet priced 15% below the wet market. This model precisely hits the pain point of users who "want low prices but fear getting burned," essentially using "efficiency" (direct sourcing to cut costs, high-frequency restocking) to meet "basic quality needs."

"Whatever type of user need you anchor to, you must back it with the matching capability." In July 2025, Starbucks' China business was reportedly up for sale at a $9 billion valuation, its market share shrunk from 34% to 14%; meanwhile Luckin's share surged to 35%, with quarterly revenue double that of Starbucks China. This battle is not a fluke of "cheap beating premium" but an inevitable choice on the coffee-retail "efficiency-experience" value axis: Starbucks anchors to "experience premium", Luckin occupies "efficiency for everyone" — and the outcome hides in whether capability matches the value anchor.

The first principle of coffee retail has always been: "use core capability to satisfy users' value priority." What users drink is never just coffee — it may be the hard need to refresh in 3 minutes, or a 2-hour social scenario. The "efficiency–experience" value axis is precisely the key tool to see this competition clearly: the horizontal axis is efficiency (price, speed, coverage density), the vertical axis is experience (space, emotional identity, brand premium). Starbucks stands in the "high-experience, mid-efficiency" range, Luckin in the "high-efficiency, mid-experience" range; the core of the difference is the judgment of "users' demand priority."

Starbucks' "experience-sensitive users": its core customers are urban white-collar workers willing to pay for the "third place"; of the 38-yuan average ticket, 40% is the cost of the space experience. Their priority is "social scene over price and speed" — but when young consumers prefer to "grab and go", the experience premium becomes a burden. Starbucks' below-20% penetration in lower-tier markets is exactly because users there care more about "good coffee for 10 yuan" than "sitting down to chat". Luckin's "efficiency-sensitive users": data shows the post-90s generation's price-performance sensitivity for coffee reaches 72%, and Luckin's 16.49-yuan ticket precisely targets the hard need of "staying alert".

Starbucks' "standardized experience": it insists on globally unified quality control, classic items exceed 60% of the menu, and local innovations must pass global headquarters review with R&D cycles up to 6 months. This "slow innovation" protects the promise that a latte tastes the same in New York and Shanghai, at the cost of lagging behind China's flavor iteration — when users want tea-coffee or liquor-coffee blends, it always reacts half a beat late. Luckin's "localized efficiency": it reads trends from 91.7 million monthly active app users; from Coconut Latte to Moutai-flavored latte, its R&D cycle is compressed to one month, with a single hit selling 44 million cups in its first month. This "fast innovation" rests on digital insight: every tap and review in the app feeds product iteration — efficiency capability responding rapidly to demand.

"Pop Mart's growth logic is the survival rule of the whole industry." In the first half of 2025, Pop Mart's revenue reached 13.88 billion yuan, the single IP LABUBU brought in 4.81 billion yuan, and the Americas market grew 11.4x. Behind these numbers lies the first principle of designer toys: the business is not selling toys, but selling "emotional companionship" — every commercial move is an answer to the question of how an IP can keep delivering emotion.

The value of a designer-toy IP lies in precisely matching a specific emotion: Molly stands for "self-expression", LABUBU for "healing companionship", Dimoo for "fantastical imagination". Pop Mart's key move is a tiered IP matrix that spreads emotional-value risk: in the first half of 2025, LABUBU accounted for 34.7% of revenue (4.81 billion yuan) and kept leading; Molly earned 1.36 billion yuan (9.8%); Skullpanda and CRYBABY contributed 1.22 billion yuan each (8.8%); Dimoo delivered 1.11 billion yuan (8%) — a structure where the head IP carries the flag and the mid-tier fills the gaps.

What Gen Z wants from designer toys is long-term emotional companionship, not a one-off purchase. The core of Pop Mart's user operations is turning IP into everyday emotional symbols: online, a collector community with 8 million monthly active users where 60% of users voluntarily share IP scenes; offline, themed stores with IP photo walls, plus extensions like HEYTEA co-branded cups and Meituan delivery boxes — so users "meet the IP" while sipping tea or picking up packages. In 2024 member repurchase hit 65%, far above the industry average of 40%.