Hou Yi, the soul of Freshippo (Hema). While working at JD.com (China's second-largest e-commerce platform), he pitched a bold plan to founder Richard Liu: build offline fresh grocery retail, reinventing traditional supermarkets with internet thinking. Liu's answer: "JD.com is an e-commerce company. We don't do offline."
After being rejected, Hou Yi took his plan to Alibaba. Daniel Zhang (then Alibaba's COO, later CEO) saw its potential and brought it to Jack Ma. Ma's reaction was the polar opposite:
"The era of pure e-commerce is over. In the next ten to twenty years, there will be no 'e-commerce' — only New Retail. Online, offline, and logistics must merge to create true New Retail."
— Jack Ma, Yunqi Conference, 2016
In January 2016, the first Freshippo (Hema) store opened in Shanghai's Jinqiao district. This was no ordinary supermarket — overhead conveyor belts for automated picking, live seafood you could buy and have cooked on the spot, checkout only via the Hema App (no cash accepted). Jack Ma personally visited the store, ate king crab, and promoted it heavily on social media.
Hou Yi later reflected: "Richard Liu rejecting me was one of his biggest mistakes." JD.com's later launch of "7Fresh" supermarkets was essentially a Freshippo copycat — but two years too late.
First principles: Hou Yi didn't ask "how do we put a supermarket online?" He asked "what are consumers' actual pain points when buying groceries?" The answer: lack of freshness, lack of convenience, lack of certainty. So Freshippo's solution was: store-warehouse integration ensures freshness, 30-minute delivery solves convenience, full App tracking eliminates uncertainty. Not "e-commerce + supermarket," but a fundamental reimagination of what "buying groceries" means.
From its 2016 opening to 2025, Freshippo endured a long cash-burning period. Alibaba never separately disclosed Freshippo's losses in its earnings reports, but industry estimates put cumulative losses at tens of billions of RMB.
First Full-Year Profit
FY2025 adjusted EBITA turned positive (confirmed in Alibaba's Feb 2025 earnings)
GMV Over 75B RMB
Gross Merchandise Value, with online orders exceeding 65% of total
Revenue Growth >40% YoY
Primarily driven by rapid expansion of Hema NB discount stores
420+ Stores
Freshippo flagship stores + Hema NB discount stores combined
~300 Stores
Hema NB Outlet discount stores, rapidly expanded in 2024-2025
IPO Failed
2023 IPO planned at $10B target valuation; investors offered only $4B; shelved
$4 Billion
2023 valuation (halved), against a $10B target
9 Years
From 2016 opening to first full-year profit in 2025
Alibaba first confirmed Freshippo's single-quarter profitability (adjusted EBITA) in its February 2024 earnings, progressing to full-year profitability in FY2025. The key inflection point came from: shutting down unprofitable store formats, aggressively expanding low-cost Hema NB discount stores, and ruthless headcount reductions.
The critical pivot: Freshippo's path to profitability wasn't about "selling more fresh groceries." It came from three things: (1) Killing all failed store formats (Mini, X Membership, Neighborhood), converging to just two formats; (2) Hema NB discount stores cracked single-store profitability through ultra-low rents + minimal buildout + high-margin private-label products; (3) Replacing the CEO — from a visionary founder to a financially disciplined operator. Profitability wasn't the result of expansion; it was the result of contraction.
Traditional supermarkets have stores in one place and warehouses elsewhere. E-commerce has warehouses but no stores. Freshippo did something unprecedented: every store simultaneously serves as a forward warehouse.
While customers browse the aisles, the overhead conveyor belt system simultaneously fulfills online orders — staff place items in bags, bags glide along rails to the back room, and delivery riders pick them up and dispatch via electric scooters.
This means:
- Online and offline sales share the same inventory, drastically reducing spoilage
- Each store doubles as both a sales venue and delivery origin, saving the cost of building separate warehouses
- What customers see in-store is exactly what's available online — what you see is what you get
Freshippo promises delivery within 30 minutes for customers within a 3 km radius of any store. This speed was an industry first in 2016.
Why 3 km? Because that's the maximum distance an electric scooter can cover in congested urban traffic, including picking time, within 30 minutes. Freshippo's site selection logic revolves entirely around this radius: opening a new store essentially means drawing a 3 km circle on the map, serving all households within it.
Freshippo stores feature live Boston lobsters, king crabs, and abalone. You can pick your seafood on the spot — an open kitchen right next to the tanks cooks it for you, and you eat it in the store's dining area.
"We didn't add a restaurant inside a supermarket. We use dining to prove how fresh our ingredients are. When you watch that crab pulled from the tank, cooked, and served to your table — are you really going to question our freshness?"
— Hou Yi, on Freshippo's dining strategy
The dining operation itself doesn't make money (razor-thin margins), but it solves the biggest trust problem in fresh grocery retail: consumers don't believe your products are actually fresh. On-site cooking is the ultimate proof of quality.
One of Freshippo's most controversial moves at launch: no cash accepted — payment only via Alipay through the Hema App. This sparked criticism in China about "discriminating against the elderly" (regulators later required cash acceptance).
But Hou Yi's logic was clear: without App binding, you can't capture the data. He needed to know what each customer bought, how often they visited, where they lived, what they preferred. This data is the foundation for Freshippo's product selection optimization, demand forecasting, and spoilage reduction.
Core formula: Freshippo's "New Retail" essentially transforms a single store into a three-in-one operation — "Experience Center + Shopping Center + Logistics Hub." The experience center (dining, live seafood) builds trust, the shopping center (grocery retail) generates revenue, and the logistics hub (forward warehouse + 30-minute delivery) creates convenience. Traditional supermarkets only have the shopping function; Freshippo layers all three into the same physical space.
China's fresh grocery retail is a market exceeding 5 trillion RMB (~$700B) annually, but with razor-thin margins (average 15-20%), extreme spoilage (20-30% for fruits and vegetables), and ferocious competition. Freshippo doesn't face one or two competitors — it faces an entire red ocean.
| Platform | Model | Store Count | Delivery Speed | Target Segment |
|---|---|---|---|---|
| Freshippo (Hema) | Store-warehouse integration + 30-min delivery | ~130 (flagship) | 30 minutes | Middle-class families, quality-oriented |
| Hema NB | Community discount store | ~300 | In-store pickup | Price-sensitive, community residents |
| Sam's Club | Membership warehouse + delivery | ~50 | 1 hour (express) | Upper-middle income families |
| Costco China | Membership warehouse | ~7 | No delivery | High income, bulk buyers |
| Dingdong Maicai | Pure forward warehouse + delivery | ~1,000 dark stores | 29 minutes | Busy office workers |
| Meituan Maicai (Xiaoxiang) | Forward warehouse + instant delivery | ~2,000 dark stores | 30 minutes | All segments, high-frequency users |
| JD 7Fresh | Store-warehouse integration (Freshippo copycat) | ~70 | 30 minutes | JD ecosystem users |
The brutal reality: Fresh grocery retail is one of China's most savage battlefields. MissFresh (formerly a leading player) went bankrupt, Dingdong Maicai drastically contracted, and Xingsheng Youxuan (community group buying) shrank after burning through billions. Freshippo survived not because it made no mistakes (it made more than anyone), but because Alibaba's deep pockets kept it alive. On this battlefield, survival itself is victory.
Freshippo may be the brand with the most "trial and error" iterations in Chinese retail history. Hou Yi's strategy: launch new formats fast, validate with data, kill what doesn't work. The problem: the cost of experimentation was staggering.
One-third the size of a standard Freshippo store, located in residential neighborhoods. The logic: "fill gaps where flagship stores can't reach." The problem: limited floor space meant limited product selection and no dining experience — losing Freshippo's core differentiation. Most have been closed.
Launched in 2021, directly competing with Sam's Club and Costco. Annual fee of 258 RMB (~$36), bulk packaging, premium quality. The problem: Sam's Club has operated in China for 20+ years with deeply entrenched supply chains and brand loyalty. Hema X opened fewer than 10 locations; all were closed by August 2025.
Launched in 2021, chasing the community group buying trend. Users order one day, pick up at community collection points the next. The problem: the sector had already been scorched into a red ocean by Meituan Select and Pinduoduo's Duo Duo Maicai through massive subsidies. Freshippo had no price advantage. Major contraction by 2022.
Located in upscale malls, focusing on imported foods and premium ingredients. The problem: average transaction values were too high, foot traffic too low, and revenue per square meter couldn't cover the rent. Closed after limited pilots.
- Hema Mini (community small stores)
- Hema X Membership Store (targeting Sam's Club)
- Hema Neighborhood (community group buying)
- Hema Premier (high-end store)
- Hema Market (wet market renovation)
- Hema F2 (convenience store)
Common cause of death: Strayed from the core competency of "store-warehouse integration," competing head-on with specialists in unfamiliar territory.
- Freshippo flagship stores (~130 locations)
- Hema NB Outlet discount stores (~300 locations)
Why they survived: Flagship stores leverage Freshippo's core competency (store-warehouse integration + experience); NB Outlets cracked "budget Freshippo" with ultra-low costs + private labels.
Core logic: Flagships build brand and experience; NB stores drive coverage and profit.
The cost of experimentation: Hou Yi once proudly called Freshippo "retail's laboratory," but laboratory costs are paid in real money. Each new format required tens of millions in investment; launching over a dozen formats meant burning through billions. The ultimate lesson: innovation doesn't mean trying everything — true innovation is finding one model and executing it to perfection. It took Freshippo 7 years to learn this.
| Force | Intensity | Analysis |
|---|---|---|
| Rivalry Among Existing Competitors | Extreme | Sam's Club (Walmart), Meituan Maicai (Meituan), Dingdong Maicai, JD 7Fresh, community group buying (Duo Duo Maicai). Every competitor is backed by a tech giant's capital, and differentiation in fresh grocery is extremely hard to establish. |
| Threat of New Entrants | Medium | High store investment barriers (a Freshippo flagship costs tens of millions RMB), supply chain development takes years — but pure-online forward warehouse models have relatively lower entry barriers. Douyin (TikTok China) e-commerce is testing fresh grocery. |
| Threat of Substitutes | High | Traditional wet markets (cheaper, more "local"), traditional supermarkets (Yonghui, Sun Art), community group buying (next-day but cheaper), food delivery platforms. Consumers have too many choices for buying groceries. |
| Bargaining Power of Suppliers | High | Fresh products are highly non-standardized with wide quality variation. Premium suppliers (e.g., Yangcheng Lake hairy crab, Dandong strawberries) are scarce. Freshippo uses private labels and direct sourcing to reduce supplier power, but with limited effect. |
| Bargaining Power of Buyers | High | Consumer switching costs are near zero (phones have Hema, Meituan, and Dingdong all installed) and price sensitivity is extreme. Freshippo's prices run 30-50% above traditional wet markets — any economic downturn triggers user attrition. |
Key finding from Five Forces: Fresh grocery retail is a "hell-level" arena where all five forces are high. Many competitors, many substitutes, non-standardized suppliers, and highly sensitive buyers. Freshippo's survival in this battlefield relies not on advantage in any single force, but on a combined moat of "experience + convenience + quality." But building this moat is extraordinarily expensive — which is why Freshippo burned cash for 9 years before turning profitable.
- Mature store-warehouse integration model; industry-leading fulfillment efficiency
- Alibaba Group technology and capital backing
- Growing private label share (margins 10-15% above third-party products)
- High brand awareness — "Hema" = New Retail
- Hema NB discount stores have proven single-store profitability
- Flagship model is capital-intensive; expansion speed is constrained
- Prices generally 30-50% above traditional channels
- Nine years of experimentation burned massive capital and management bandwidth
- Heavy dependence on Alibaba Group (capital, technology, traffic)
- Fresh produce spoilage rate still above industry average
- China's online fresh grocery penetration remains low (~15%); significant growth runway
- Hema NB can be rapidly replicated to Tier 2-3 cities
- Private label expansion into prepared meals and ready-to-eat foods
- Consumption upgrade trends driving sustained demand for quality fresh products
- Independent listing could attract external capital to accelerate expansion
- Sam's Club accelerating China expansion (targeting 100 stores)
- Meituan Maicai and Pinduoduo waging price wars to seize market share
- Economic downturn triggering consumption downgrade; middle-class customer base shrinking
- Alibaba Group restructuring may reduce resource allocation to Freshippo
- Food safety incident risk (prior controversies over relabeling expiration dates)
| Stage | Traditional Supermarket | Freshippo Model | Value Shift |
|---|---|---|---|
| Procurement | Distributors add layers of markup (farm → primary wholesaler → secondary wholesaler → store) | Direct sourcing + custom private labels | Eliminates 2-3 layers of middlemen, improving margins by 10-15% |
| Warehousing | Central warehouse → store, slow turnover | Store-warehouse integration; the store IS the warehouse | Eliminates one transfer step, reducing spoilage by 5-8% |
| Display | Shelf display, relying on promoters | Live seafood display + on-site cooking + App recommendations | From "passive display" to "active experience" |
| Sales | In-store checkout counter | App ordering + 30-min delivery (online >65%) | From "come to the store" to "delivered to your home," expanding coverage radius |
| Delivery | Consumers carry groceries home themselves | 3 km, 30-minute delivery | Shifts "last mile" cost from consumer to platform |
| Data | Only POS transaction data (only knows what sold) | Full-funnel data (browsing, cart, payment, repurchase, reviews) | From "transaction ends at checkout" to "continuous optimization" |
The key value chain shift: Freshippo's biggest disruption to traditional supermarkets isn't in "sales" — it's in "data." A traditional supermarket only knows "we sold 100 kg of pork today." Freshippo knows "which neighborhoods, which users, at what times, prefer which cuts of pork — and last Saturday's rainy day sales were 23% higher than a sunny day." This difference in data granularity determines the vast gap in procurement accuracy and spoilage rates.
In 2023, after Alibaba's "1+6+N" restructuring (splitting the conglomerate into six business groups), Freshippo was the most anticipated candidate for independent listing. Initial target valuation: $10 billion. But during the roadshow, investors were only willing to offer $4 billion — less than half the target.
Reasons: persistent losses, uncertain profitability model, weak Chinese consumer market, and investor enthusiasm for "New Retail" having cooled considerably. Freshippo's IPO was postponed indefinitely.
In 2023-2024, Freshippo shuttered dozens of stores, including all Hema X Membership stores, most Hema Mini locations, and several underperforming flagship stores. Multiple rounds of layoffs cut thousands of positions across headquarters and stores.
In March 2024, Freshippo founder Hou Yi was "retired," his title changing from CEO to "Honorary Advisor." His successor was an Alibaba veteran known for cost control and operational precision.
Industry interpretation: Hou Yi is a classic "0-to-1" founder — full of imagination, willing to experiment, but burning cash without blinking. When Freshippo needed to shift from "cash-burning expansion" to "disciplined profitability," Hou Yi's style was no longer the right fit.
"Hou Yi's contribution was inventing the New Retail prototype. His problem was that he wanted to use the same approach to invent ten prototypes."
— A former Freshippo executive (anonymous)
Between 2023 and 2025, Alibaba massively restructured its retail portfolio: selling off Sun Art (its hypermarket chain, at well below acquisition price) and divesting Intime Department Stores. While Freshippo wasn't sold, Alibaba's stance shifted from "full support" to "sink or swim on your own."
This was a double-edged sword for Freshippo: it lost the unlimited lifeline of capital, but gained greater operational autonomy. Being forced to break even actually accelerated Freshippo's path to profitability.
In August 2025, Freshippo announced the closure of all X Membership stores. This was a public admission: in the membership warehouse sector, Freshippo couldn't beat Sam's Club. Sam's Club's brand loyalty and supply chain depth in China, built over 20+ years, cannot be replicated in just a few years of imitation.
The essence of these controversies: All of Freshippo's controversies point to the same problem — "running at the right direction but the wrong speed." The New Retail direction was correct, but simultaneously launching over a dozen store formats and competing head-on with specialists in every sector scattered resources and amplified losses. The ultimate solution wasn't more innovation — it was more discipline.
Liu Bei (161-223 AD), the founding emperor of Shu Han during China's Three Kingdoms period, wandered for twenty years before finding his greatest strategist. He sought refuge under one warlord after another — Gongsun Zan, Tao Qian, Cao Cao, Yuan Shao, Liu Biao — always a guest with no territory of his own. Yet Liu Bei held firm to one conviction: a ruler of virtue would ultimately win the realm.
The famous "Three Visits to the Thatched Cottage" isn't just a story about humility before talent — it's about "having the right direction but needing to try repeatedly when the timing isn't right." Liu Bei's first two visits to the sage Zhuge Liang's cottage came up empty. Most people would have given up. But Liu Bei knew: finding the right person (Zhuge Liang, who would become his legendary military advisor) was worth any wait.
Freshippo's nine years follow the same logic. The "New Retail" direction was right — the fusion of online and offline, data-driven operations, experience upgrades — these trends are irreversible. But finding the right model (flagship store-warehouse integration + NB discount stores) required repeated experimentation, repeated failure, repeated convergence. Liu Bei wandered for twenty years before the strategic masterplan; Freshippo burned cash for nine years before full-year profitability. The right direction doesn't guarantee quick success, but it guarantees eventual success — if you survive long enough.
Shang Yang (c. 390-338 BC) was a statesman who transformed the ancient Chinese state of Qin from a backwater into a military powerhouse — laying the foundation for Qin's eventual unification of China. His reforms weren't implemented all at once. He first tested new policies in a single county (a pilot program), then rolled them out nationwide. The reforms themselves went through multiple iterations: the first wave focused on agriculture and military merit rewards; the second addressed administrative structure and land reform.
Freshippo's experimentation with over a dozen store formats follows essentially the same logic: a new system (New Retail model) can never be perfect from the start — only practice reveals what works and what doesn't. Hema Mini didn't work. X Membership didn't work. Neighborhood didn't work — but each failure helped Freshippo understand more clearly "what actually does work."
The price of Shang Yang's reforms was Shang Yang himself — he was executed by the old guard's backlash (literally torn apart by chariots). The price of Freshippo's experimentation was Hou Yi being "retired" (the backlash of burning too much money). But the reforms left behind the foundation of a powerful Qin; Hou Yi left behind the New Retail prototype. The reformer's fate is often this: the one who plants the tree never enjoys its shade, but the tree keeps growing.
Freshippo's "New Retail" direction was correct — by 2025, Chinese retail is indeed moving toward online-offline integration. But Hou Yi committed a classic founder's error: equating "right direction" with "everything is worth trying." He simultaneously launched over a dozen store formats, each requiring massive investment and a dedicated operations team, resulting in scattered resources and mediocre execution across the board.
Business takeaway: The right direction is a necessary condition, but not a sufficient one. You also need "the right speed" and "the right focus." If you have a good idea but limited resources, put 80% of your resources into the single most promising model — not 10% into eight different models. Had Freshippo stuck to just "flagship stores + discount stores" from the start, it might have achieved profitability in three years.
Launching new store formats is exciting — every new experiment is full of possibility. But shutting down unprofitable formats is painful — you must admit failure, lay off teams, and explain losses to investors. Freshippo spent seven years expanding (2016-2023) but only two years contracting (2023-2025) — and then turned profitable.
Business takeaway: In a company's lifecycle, "knowing what NOT to do" matters more than "knowing what to do." When Steve Jobs returned to Apple, the first thing he did was kill 70% of the product line. Freshippo killed over a dozen formats to keep just two. Contraction isn't retreat — it's focus. Focus is what creates penetrating power.
Hou Yi was Freshippo's soul — he invented the New Retail prototype. But when Freshippo needed to shift from "creative exploration" to "disciplined profitability," Hou Yi's freewheeling style became an obstacle. Alibaba made a brutal but correct decision: move the founder to an advisory role and bring in someone skilled at managing by the numbers.
Business takeaway: Different stages require different types of leaders. 0-to-1 needs a visionary founder (high risk tolerance, boundless imagination); 1-to-100 needs a disciplined operator (cost optimization, efficiency focus). The most successful companies know when to make this switch. Apple's transition from Jobs to Tim Cook is the textbook example.
Without Alibaba, Freshippo would have died long ago — nine years of losses and billions in investment are beyond most investors' tolerance. But Alibaba's backing also had side effects: Freshippo never needed to earn its own keep, fostering a "spend freely" culture. Only when Alibaba restructured and demanded self-sufficiency did Freshippo seriously confront the profitability question.
Business takeaway: Corporate ventures have a structural paradox: the parent company's resources ensure you won't die, but also eliminate the urgency of "make money or perish." The ideal state is "have a safety net but act as if you don't" — leverage the parent's resources, but manage with a startup's discipline.