Amazon Fresh Buries the 72-Store Experiment
Same-day pivot reveals a $4T market is harder than it looks
Amazon quietly closed its 72 Fresh stores across the US and pivoted to same-day grocery delivery via Prime. The physical footprint experiment is over. The lesson: grocery is a logistics problem disguised as a real estate problem.
What happened
Amazon shuttered all 72 Amazon Fresh locations in the US, replacing the physical store model with a same-day delivery aggregation layer over existing Whole Foods, Amazon Fresh delivery-only facilities, and third-party grocers on its platform.
The numbers that matter
- **72 stores** closed in Q1 2026, affecting ~2,400 employees
- **Same-day delivery** now covers 47 US metros via Prime
- **Whole Foods footprint**: 500+ stores; Fresh was positioned above Whole Foods in the value chain
- **Amazon's grocery TAM**: estimated $1.4T annual US grocery spend; Amazon has <3% share
Why the physical format failed
**Location selection** was built around Amazon's data map - high-density urban zip codes - rather than grocery shopping patterns. Amazon shoppers buy online; grocery shoppers still buy in-store, especially for fresh.
**Format mismatch**: Amazon Fresh averaged 35,000 sq ft. The sweet spot for urban grocery is 8,000-15,000 sq ft (Aldi's format). Too much space = too much rent = margin compression.
**The ambient intelligence bet didn't scale**: Amazon's Just Walk Out technology and Dash Carts required $2M-$4M per store in capital. At 72 stores, that's $144M-$288M in technology investment with no clear unit economics payback.
What this means for European grocers
Amazon Fresh's failure is a data point, not a template. European grocers have operated under the assumption that Amazon would eventually crack physical grocery in Europe (they still have Fresh UK and DE). The US pullback signals a strategic pivot to: platform over proprietary footprint.
For European grocers, this means:
- Amazon is more likely to partner/acquire than build organically
- The threat model shifts from 'Amazon opens stores near you' to 'Amazon indexes your inventory and redirects demand'
- Discount formats (Aldi, Lidl) remain structurally immune; they optimize for the shopper who doesn't use Amazon anyway
Walmart: $275B in Grocery and Nobody Takes It Seriously
$1T market cap built on a business model that confounds every analyst
Walmart does $275B in annual grocery revenue. That's more than the entire GDP of Portugal. Yet most strategic analysts still treat it as a US discount retailer rather than the world's largest grocery company by revenue.
The scale problem
Walmart's grocery revenue ($275B) exceeds the GDP of 160 countries. If Walmart's grocery division were a standalone company, it would be the 3rd largest retailer in the world by revenue - ahead of Costco, behind Amazon and Schwarz Group (Lidl + Kaufland).
Why analysts keep misreading it
**Margin opacity**: Walmart reports grocery as part of a blended segment that includes general merchandise. The 22-24% gross margin on grocery is obscured by the higher-margin GM category.
**Format complexity**: Walmart runs 4,600+ US stores across 3 formats (Supercenter, Neighborhood Market, Express). Each format has different unit economics. Most analysts apply Supercenter margins to the whole fleet.
**International blindness**: Walmart International has 5,300+ stores across 18 countries. In Canada, Mexico, and Chile, it operates premium-format grocery. The international story is entirely separate from the US discount narrative.
The operational moat
Walmart's grocery advantage is supply chain density:
- **12 regional distribution centers** with 99.7% order accuracy
- **Private label penetration**: 25% of grocery units sold; Great Value + Member's Mark generate $40B+ in revenue
- **Sam's Club**: 600+ US clubs, $65B revenue, 50% overlap with Walmart Supercenter geography - cross-subsidization that competitors can't replicate
What this means for European grocers
Walmart's playbook is a franchise model: accept low margins on volume, recover via private label and financial services. Tesco (UK) and Carrefour (FR) have both tried to clone this playbook. The difference: Walmart started with the supply chain, then added private label. Tesco started with the brand, then tried to retrofit the supply chain. That's why Tesco's turnaround has been messier than the model suggests it should be.
Alibaba's Hema: The Format Nobody in Europe Wants to Admit Works
300+ stores, multi-format play, and a data model that European grocers are 5 years behind
Alibaba's Hema operates 300+ stores across 30 Chinese cities, generating RMB 100B+ ($14B) in annual revenue. The format combines supermarket, restaurant, fulfillment center, and data collection engine in a single model that Western analysts keep dismissing as 'unique to China.' It's not.
The Hema model, deconstructed
Hema (also: Freshippo) is a supermarket where every SKU has a digital twin. You scan an item with the Hema app, and the system records your location in the store, your basket composition, your time-of-day pattern, and your repeat purchase cycle. This data feeds directly into Alibaba's supply chain and vendor negotiation engine.
**Format variants**:
- **Standard Hema**: 4,000-6,000 sq m, restaurant inside (30-50% of floor space), 30-min delivery radius, 3,000 SKUs
- **Hema City**: 10,000+ sq m, destination format, lower delivery density but higher basket size
- **Hema Next**: convenience-format, 500 sq m, high-frequency, algorithm-optimized layout
The numbers
- **300+ stores** as of Q1 2026 (down from a peak of 365; Alibaba closed underperformers in 2024)
- **RMB 100B+ revenue** (~$14B) - largest single-company grocery format in China
- **30-min delivery**: 60% of orders fulfilled via in-store dark store; 40% from adjacent mini-warehouse
- **Conversion rate**: 30% of visitors who scan QR codes become repeat buyers (vs. 8% industry average for loyalty programs)
Why European grocers can't just copy it
**Data infrastructure gap**: Hema's magic is the real-time inventory-to-customer mapping. European grocers use ERP systems from 2005; Hema's stack is natively cloud-native.
**Vendor power asymmetry**: Alibaba's vendor contracts are renegotiated weekly based on Hema's sell-through data. In Europe, most vendor contracts are annual and fixed - retailers can't reprice mid-contract based on real-time sell-through.
**Regulatory**: European GDPR makes Hema's per-customer tracking model legally impossible without explicit consent infrastructure that most retailers haven't built.
The European analogue that's closest
Tesco's clubcard data operation (UK) is the closest Western approximation - but it's still 3-5 years behind Hema's data velocity. Tesco knows what you buy; Hema knows where in the store you looked, what you picked up and put back, and how long you spent in the wine aisle.
For European grocers: the strategic priority is not to build Hema - it's to build the data infrastructure that would make Hema-style operations possible in 3 years.
Woolworths Australia: Building the Fresh Food Moat Brick by Brick
The 5 strategic moves that have kept Coles and Aldi at bay for 8 consecutive years
Woolworths has operated at a structural premium over its competitors in Australian grocery for over a decade. The moat isn't the brand - it's the combination of fresh supply chain depth, store network density, and data leverage that the competitors have spent $4B+ trying to breach. The key metrics: Woolworths controls 37% of the Australian grocery market, generates $5.2B in EBIT annually, and its private label (Woolworths Select) now accounts for 28% of units sold - up from 18% in 2019. The fresh food differentiation is structural: Woolworths operates 220 fresh produce distribution centers, with cold chain coverage that Coles is still 3 years away from matching....
ANALYSIS LOCKED
Enter your email to read this issue and get Beyond Facing delivered daily.
M&S Food Hall: The UK's Most Profitable Square Footage
6,000 to 22,000 sq ft - the format evolution that turned a struggling department store into a grocery destination
Marks & Spencer's Food Hall format has quietly become the most profitable grocery square footage in the UK. Not the biggest, not the cheapest - the most profitable. 6,000 sq ft in a high street location generating GBP 4,200 per sq ft in annual sales versus the industry average of GBP 1,100. The format evolution happened quietly: M&S shrank its clothing floors (facing pressure from fast fashion), expanded food halls from 6,000 sq ft to 22,000 sq ft in full-line stores, and introduced the Simply Food standalone format targeting fuel stations, train stations, and high-traffic urban locations....
ANALYSIS LOCKED
Enter your email to read this issue and get Beyond Facing delivered daily.
Spinneys Dubai: The Gulf's Most Sophisticated Grocery Operation
Dh1B+ revenue, GCC expansion, and the private label strategy that is making Carrefour very nervous
Spinneys operates 74 stores across the UAE, Qatar, and Oman, generating Dh1B+ ($272M) in annual revenue. The operation is a masterclass in how to price, format, and localize in a market where 89% of the population is expatriate - meaning traditional grocery loyalty plays don't work. The Spinneys value proposition is: premium fresh produce (70% imported from 30 countries), western format (wide aisles, climate-controlled, English-speaking staff), and a private label (Spinneys Food) that competes head-to-head with Carrefour on quality but wins on freshness perception....
ANALYSIS LOCKED
Enter your email to read this issue and get Beyond Facing delivered daily.
Tesco's UK Turnaround: The Numbers Behind the Comeback
After nearly collapsing in 2014, Tesco has rebuilt a margin profile that analysts said was impossible
Tesco's UK operating margin hit 5.8% in FY2025 - up from a near-zero margin in 2014 when the accounting scandal nearly killed the company. That 5.8% represents a GBP 1.2B operating profit improvement in 11 years. The recovery had four phases: (1) close the underperforming stores and exit international markets (Spain, Japan, US); (2) rebuild the supply chain from scratch after the horsemeat scandal of 2013 exposed supplier management failures; (3) roll out the VusionGroup ESL system to 2,400 stores; (4) rebuild private label quality after the horsemeat crisis destroyed consumer trust....
ANALYSIS LOCKED
Enter your email to read this issue and get Beyond Facing delivered daily.
Costco Global: Why $192M for One Warehouse Makes Perfect Sense
The membership warehouse model and why the 92.3% renewal rate is the most important number in retail
Costco spends $192M to open a new warehouse. Most grocers spend $8-15M. Costco opens 20-25 new warehouses per year. In 2025, they opened 24. The math works because of one number: 92.3%. That's Costco's annual membership renewal rate - meaning almost everyone who tries Costco stays. The customer acquisition cost (CAC) for a Costco member is effectively $0 after year one....
ANALYSIS LOCKED
Enter your email to read this issue and get Beyond Facing delivered daily.
Aldi Global's 400-Store Expansion: The Format That's Winning Europe's Grocery Wars
90%+ private label, sub-1,500 SKU count, and the real reason Aldi keeps growing in markets it should not win
Aldi opened 400+ net new stores in 2025, pushing its global footprint to 13,500+ stores across 19 countries. The expansion is accelerating - not because Aldi undercuts on price (they don't anymore), but because they have the lowest cost-to-serve in the industry. Aldi's cost-to-serve per sq ft is GBP 18 versus the UK industry average of GBP 42. That GBP 24 gap is reinvested into price (not margin), which keeps the cycle going....
ANALYSIS LOCKED
Enter your email to read this issue and get Beyond Facing delivered daily.
Lidl's European Expansion: How Germany's Most Relentless Discounter is Winning in Markets It Abandoned a Decade Ago
Re-entered France, Spain, and Italy. Now dominant. The playbook for coming back is the same as the playbook for arriving.
Lidl re-entered the French market in 2012 after a failed attempt in the 1990s. In 2025, it became the #2 grocer in France by store count. It re-entered Italy in 2016 and is now the #3. The re-entry playbook is a masterclass in what to change and what to keep the same. What changed: store format (bigger, more lighting, wider aisles), product mix (more local/regional SKUs), and pricing architecture (competitive with Leclerc and Casino, not just discount-to-others)....
ANALYSIS LOCKED
Enter your email to read this issue and get Beyond Facing delivered daily.
Carrefour's AI Pricing Engine: The French Grocer That's Quietly Becoming Europe's Most Data-Driven Retailer
Real-time repricing across 8,000 SKUs, 2,000 stores, 6 countries - and why the vendor contracts actually support it
Carrefour implemented a real-time AI pricing engine in 2024 across its French hypermarket and supermarket network. 8,000 SKUs, repriced every 6 hours based on competitor data, demand elasticity modeling, and inventory level signals. The result: +1.2% net margin improvement in year one. That's EUR 240M in margin improvement on EUR 76B in revenue - meaningful at that scale....
ANALYSIS LOCKED
Enter your email to read this issue and get Beyond Facing delivered daily.
Jumbo Netherlands: The Dutch Supermarket That Decided to Beat Amazon on Convenience
500+ stores, EUR 9.2B revenue, and the same-day delivery model that Albert Heijn is now copying
Jumbo is the #2 grocer in the Netherlands with 700+ stores and EUR 9.2B in revenue. In 2025, it launched a same-day delivery service that covered 60% of the Dutch population within 6 months. Albert Heijn (owned by Ahold Delhaize) has since announced a direct copy of the model. The Jumbo advantage: 35% of its stores are within 3km of a Jumbo Distribution Center. The DC-to-store density is 3x the European average....
ANALYSIS LOCKED
Enter your email to read this issue and get Beyond Facing delivered daily.
Migros Switzerland: The Cooperative Model That Hasn't Changed Since 1942 - And Why That's the Point
CHF 30B revenue, 90%+ Swiss sourcing on private label, and the one constraint that protects it from every competitor
Migros is the largest retailer in Switzerland by revenue (CHF 30B, ~$34B) and operates as a federated cooperative of 10 regional cooperatives. It is structurally impossible to acquire or replicate. Here's why, and what the weakness is. The cooperative structure means that profit isn't extracted by shareholders - it flows back into regional communities, stores, and infrastructure. Migros owns its own bank (Migros Bank, CHF 23B in deposits), its own insurance, its own travel agency. This is a moat that no external competitor can breach....
ANALYSIS LOCKED
Enter your email to read this issue and get Beyond Facing delivered daily.
Casino Group: France's Most Complex Retailer is Running Out of Time
EUR 9.6B debt, 3 competing brands, 4 formats, and a restructure that has to work or the whole thing goes
Casino Group carries EUR 9.6B in net debt, operates 4 formats (Hyper Casino, Casino Supermarket, Casino Express, Vindemia in Indian Ocean), and is in the middle of a restructuring that will determine whether France has 3 or 4 national grocers in 5 years. The strategic question: can Casino become a focused urban player or does it have to sell? Casino's debt situation is a legacy of the 2000s acquisition strategy (Grupo Pao de Acucar, Extra, Monoprix) - acquisitions that were sound individually but created a balance sheet that can't support the capital investment needed to compete....
ANALYSIS LOCKED
Enter your email to read this issue and get Beyond Facing delivered daily.
The European Hypermarket is Dead. What's Replacing It Will Define the Next Decade.
Carrefour, Tesco, Casino - all dismantling the format that defined European grocery for 40 years. The grocerant is the replacement.
European hypermarkets are closing at a rate of 200+ per year. The format that anchored every out-of-town shopping center from 1985 to 2015 is being demolished. In its place: the grocerant (grocery + restaurant), urban dark stores, and proximity formats. This is the strategic map of European grocery in 2026. The hypermarket decline statistics: France lost 340 hypermarkets between 2018 and 2025 (from 1,240 to 900); UK lost 280; Germany lost 190. The closures represent 85M sq ft of retail space being reconsidered....
ANALYSIS LOCKED
Enter your email to read this issue and get Beyond Facing delivered daily.