The $795 Billion Mirage: Explosive Growth, Persistent Losses
By 2025, the global real-time instant delivery market is projected to process 795 billion orders annually, growing at a compound annual growth rate (CAGR) of 30.5% from 2020 to 2025 (iResearch). This staggering figure—equivalent to over 2.1 billion deliveries per day—has ignited investor enthusiasm, catalyzed IPOs, and fueled multi-hundred-million-dollar funding rounds. Yet beneath this blue-ocean narrative lies a stark financial reality: profitability remains structurally elusive.顺丰同城 (SF Now), the fourth publicly listed entity under Wang Wei’s empire, reported net losses of RMB 330 million, RMB 470 million, and RMB 758 million in 2018–2020. Its U.S.-listed peer, Dada Nexus, fared worse: USD 272 million, USD 242 million, and USD 246 million in net losses over the same period. These are not startup-phase deficits—they are systemic losses persisting across scale inflection points.
What makes this paradox so acute? Unlike traditional logistics segments—full-network express (3–5 days), warehouse-to-consumer (same-day/next-day)—instant delivery operates under a fundamentally different economic logic. Its core value proposition—30–60 minute standard delivery, with sub-15-minute ‘flash’ service increasingly expected—demands hyper-local, human-first, point-to-point execution. This eliminates the network efficiencies that define scalable logistics: no centralized sorting hubs, no batched routing algorithms optimized for volume, and no fixed asset amortization across thousands of SKUs. Instead, every order triggers a discrete, time-sensitive, labor-intensive event. As Southwest Securities analysis confirms, while China’s leading express carrier reduces per-parcel cost by 22% as daily volume rises from 20M to 50M parcels, instant delivery platforms like Meituan see no meaningful decline—and often a slight increase—in rider-per-order cost beyond 25M daily orders. Scale doesn’t bend the curve; it stretches the operational wire thinner.
The Human Cost Ceiling: Why Labor Dominates and Defies Optimization
Instant delivery’s cost structure is brutally transparent—and brutally inflexible. For SF Now, personnel-related expenses—including rider outsourcing fees and employee welfare—accounted for 97.8%, 97.3%, and 97.8% of total operating costs in 2018–2020. With 2.8 million registered riders and an average of 10.7 billion orders fulfilled in the first five months of 2021 alone, the platform’s financial engine runs on human capital, not automation or infrastructure leverage. This isn’t a transitional phase; it’s a structural imperative. Autonomous last-mile solutions remain confined to controlled campuses and low-speed zones, while AI-powered dynamic dispatch systems—though sophisticated—cannot eliminate the physical constraints of urban traffic, staircases, building access protocols, or real-time customer interaction.
Contrast this with the economics of full-network express. A single ZTO Express sorting hub processes over 10 million parcels daily, achieving marginal cost reductions through robotics, predictive analytics, and fixed-line transportation networks. In instant delivery, the ‘hub’ is the rider’s smartphone, the ‘line haul’ is a 3-kilometer scooter ride, and the ‘sortation’ happens in real time—by a person interpreting ambiguous addresses, managing multiple concurrent orders, and absorbing service-level penalties for lateness. Crucially, rider earnings are tightly coupled to local minimum wage regulations, social insurance mandates, and platform commission ceilings—none of which scale downward with volume. When SF Now’s average order value plummeted from RMB 12.44 in 2018 to RMB 2.85 in early 2021, its unit fulfillment cost only dropped from RMB 15.00 to RMB 5.90. The gap narrowed—but remained deeply negative. That RMB 3.05 loss per order isn’t inefficiency; it’s physics meeting policy meeting pricing power.
- Meituan Delivery: ~27.8 million daily orders (2020), 4.3% operating margin on food delivery—achieved only via massive cross-subsidization from high-margin ads, fintech, and enterprise SaaS.
- Dada Nexus: ~2.9 million daily orders (2020), negative 15.2% adjusted EBITDA margin—despite being embedded within JD.com’s ecosystem.
- SF Now: ~2.7 million daily orders (2020), negative 28.4% gross margin—with no captive marketplace to monetize data or drive ancillary revenue.
The Third-Party Illusion: Independence vs. Dependency in Platform Economics
SF Now positions itself as China’s largest third-party instant delivery platform—a distinction meant to signal neutrality, flexibility, and strategic independence from walled-garden ecosystems. But the numbers tell a different story. In 2020, 33.6% of SF Now’s total revenue came directly from its parent company, SF Holding; in the first five months of 2021, that dependency spiked to 38.6%. More revealingly, RMB 16.22 billion of its RMB 48.43 billion in 2020 revenue was derived exclusively from ‘last-mile delivery services for SF Holding’. That means over one-third of its business is functionally a captive internal logistics arm—not a diversified, demand-driven platform.
This dependency undermines its core strategic narrative. While Meituan and Ele.me control their own demand funnels (via apps, search, payments, and user data), and Dada leverages JD’s e-commerce traffic and merchant relationships, SF Now lacks any native commerce layer. It cannot set pricing, influence consumer behavior, or capture lifetime value beyond a single delivery fee. Its ‘third-party’ status is operational—not economic. And crucially, its independence comes at a steep cost: without integrated demand, it must compete for merchant contracts in a fragmented, price-sensitive B2B market where margins are razor-thin and churn is high. Its top clients include major chains like McDonald’s, Starbucks, and Hema, but none grant SF Now exclusivity—or guaranteed volume. Every contract is renegotiated quarterly, pressured by rivals offering deeper discounts or bundled tech services. True third-party viability requires either platform-level data moats (e.g., real-time demand forecasting APIs sold to retailers) or embedded logistics-as-a-service infrastructure (e.g., white-labeled delivery OS for grocers). SF Now has neither.
Proximity Commerce: The Promise and Peril of the ‘Near-Field’ Economy
‘Near-field commerce’—defined as fulfilling high-frequency, low-consideration needs within a 3–5 km radius—is widely heralded as instant delivery’s salvation. From community group buying (Meituan优选, Pinduoduo’s DuoDuo Maicai) to flash grocery (Hema, Dingdong Maicai), the trend promises higher order density, shorter routes, and repeatable baskets. Analysts at Shenwan Hongyuan estimate instant delivery could eventually displace 15% of full-network express volume—a multi-billion-dollar opportunity. Yet proximity commerce also intensifies the very pressures that cripple profitability.
First, near-field models thrive on ultra-low unit economics: a 20-minute grocery delivery must cost less than RMB 5 to be defensible against in-store pickup or 1-hour courier services. Second, they require deep vertical integration: dark stores, micro-fulfillment centers, proprietary inventory systems—all capital-intensive assets SF Now does not own or operate. Third, they reward first-party control: Meituan can route all Hema orders through its own fleet because it owns both the demand and supply sides. SF Now must bid for those same orders on open marketplaces—often losing to lower-cost, vertically aligned competitors. As a result, near-field growth hasn’t lifted SF Now’s unit economics; it has compressed them further. Its average order value collapsed by 77% between 2018 and 2021, reflecting a strategic pivot toward lower-margin, higher-volume micro-deliveries—exactly the segment where scale amplifies losses rather than efficiencies.
Pathways Forward: Beyond the IPO Mirage
The SF Now IPO isn’t an exit—it’s a lifeline. Like Dada before it, the listing provides runway to fund what the market refuses to finance organically: technology investment, rider incentives, and strategic M&A. But long-term viability demands moving beyond the ‘delivery-only’ trap. Three credible pathways exist:
- Logistics Intelligence Layer: Monetize dispatch AI, real-time ETAs, and predictive congestion modeling as SaaS for retailers—transforming from cost center to data vendor.
- Hybrid Fulfillment Networks: Partner with regional grocers to co-invest in shared micro-fulfillment hubs, capturing upstream inventory and downstream delivery margins.
- Embedded Financial Services: Leverage rider and merchant transaction data to offer working capital loans, insurance, or payroll solutions—creating recurring, high-margin revenue streams uncorrelated with parcel volume.
None are easy. All require abandoning the fantasy of pure-play neutrality. The $795 billion market won’t be won by the fastest rider or the most riders—it will be won by the platform that best arbitrages data, capital, and physical infrastructure across the entire near-field value chain. Until then, instant delivery remains less a logistics revolution—and more a high-stakes endurance race where every kilometer costs more than the last.
Source: 36Kr, “SF Now: There Is No Optimal Solution in Instant Delivery,” December 10, 2021
