高层住宅密集的香港，让送货上门在结构上就成本高昂——客户不在家，快递员重复上门，单件成本层层叠加。In high-rise Hong Kong, door-to-door delivery was structurally expensive — customers weren't home, couriers re-attempted, and cost per parcel compounded.
电梯、楼层、步行距离让纯人力配送不可能规模化盈利——问题不是跑得不够快，而是模式本身。Elevators, floors and walking distance made human-only delivery uneconomic at scale — the model, not the effort, was the problem.
没有统一基础设施把三方连接起来；订单、运输、柜机、仓储各自孤立，没有系统做端到端编排。No shared infrastructure connected the three; orders, transport, lockers and warehousing ran in isolation with nothing orchestrating end to end.
而是一套末端配送的操作系统——软件、硬件与合作伙伴集成，作为一个整体来设计。It was an operating system for the last mile — software, hardware and partner integration designed as one.
| 我们选择了We chose | 而不是Not | 为什么Why |
|---|---|---|
| 自持柜机 + 微仓基础设施Owning lockers + micro-warehouse infrastructure | 轻资产撮合平台An asset-light marketplace | 不掌握物理网络就无法保证服务水平——基础设施才是护城河。Without the physical network you can't guarantee service levels — infrastructure is the moat. |
| 自营车队 + 3PL 混合模式Hybrid in-house + 3PL fleet | 全自营或全外包All in-house or all outsourced | 用成本模型逐单权衡质量与成本——两个极端都算不过账。A per-order cost model balanced quality against cost — neither extreme penciled out. |
| 把 800 家便利店变成合作节点Making 800 convenience stores partner nodes | 与便利店自提竞争Competing with store pickup | 合作换覆盖：门店提供 24/7 人工溢出能力，柜机承接高频自动化流量。Partnership bought coverage: stores added 24/7 staffed overflow, lockers took the high-volume automated flow. |
| 开放 API 平台An open API platform | 封闭生态A closed ecosystem | 封闭会锁死网络增长——开放让 S.F. Express、UPS 与电商平台主动接入。Closed would cap network growth — open brought S.F. Express, UPS and e-commerce platforms onboard. |
部署首批核心柜机，用自营配送团队跑通单位经济模型——先证明每一单能算得过账，再谈规模。Deployed the first core lockers and ran an in-house delivery team to validate unit economics — prove each order pencils out before scaling.
上自动化分拣（全自动 + 半自动）支撑 1,000+ 柜机的流量；接入外部物流合作伙伴，网络开始双边生长。Added automated sorting (full + semi) to feed 1,000+ lockers; onboarded logistics partners — the network began growing from both sides.
从基础设施升级为品牌差异点：与 YSL Beauté 共创 AI 人脸识别快递柜、礼品自提活动、加急配送——高端品牌开始把网络当作体验渠道。Elevated utility into brand: YSL Beauté AI facial-recognition lockers, gift-redemption programs, express delivery — premium brands began treating the network as an experience channel.
以 800 家便利店合作把覆盖推到 2,000+ 点位，年处理 1,100 万件——用合作而非资本完成最后一段扩张。Pushed coverage to 2,000+ points via 800 convenience-store partnerships, 11M parcels a year — the final expansion bought with partnership, not capital.
| 岗位要求JD requirement | 生态中的实证Proof in this ecosystem |
|---|---|
| 覆盖仓储、履约与运输的端到端解决方案End-to-end solutions across warehousing, fulfillment & transportation | 三者集于一张网络——WMS（微仓）、OMS + Locker OS（履约）、TMS（运输）All three in one network — WMS (micro-warehouse), OMS + Locker OS (fulfillment), TMS (transportation) |
| 系统集成——WMS/TMS/OMS/ERP + API/EDI；测试与上线Systems integration — WMS/TMS/OMS/ERP + API/EDI; testing & go-live | 设计了 OMS 与开放 API；每次承运商接入均走规格 → 沙箱 → UAT → 切换 → SLA全流程Designed the OMS + open API; every carrier onboarding ran spec → sandbox → UAT → cutover → SLA |
| 主导实施：从方案设计到跨职能上线Lead implementation design → launch across functions | 5 年建设，0 → 2,000+ 网点，协同承运商、商家、运营与工程团队5-year build, 0 → 2,000+ points, coordinating carriers, merchants, ops and engineering |
| 数据/KPI 驱动的流程、自动化与成本优化Data/KPI-driven process, automation & cost optimization | Tableau 产能预测指导选址；混合车队成本模型逐单优化；自动化分拣按依赖节奏引入Tableau capacity forecasting for placement; hybrid-fleet cost model optimizing per order; automation phased in on the dependency curve |
| 客户关系管理与主动建议Customer relationship & proactive recommendations | 中小长尾客户与企业级承运商均以规格文档、SLA 和业务复盘完成接入；YSL 从客户变共创伙伴SMB long-tail and enterprise carriers onboarded with specs, SLAs and business reviews; YSL grew from customer to co-creation partner |
| 物流领域的英语 + 普通话双语能力English + Mandarin in the logistics domain | 双母语——且该网络正对应京东物流自身的一体化末端配送 / 自提点业务形态Native both — and this network mirrors JD Logistics' own integrated last-mile / pickup-point business |
| 京慧能力域京慧 domain | 我在快递柜生态中的实践What I ran on the locker ecosystem |
|---|---|
| 设计 / Design | 网络与节点设计——快递柜选址、微仓拓扑、覆盖策略Network & node design — locker placement, micro-warehouse topology, coverage strategy |
| 计划 / Planning | 产能预测、混合车队成本模型、跨节点库存布局Capacity forecasting, hybrid-fleet cost model, inventory positioning across nodes |
| 执行 / Execution | OMS + TMS + Locker OS 实时运转，支撑每年 1,100 万件包裹OMS + TMS + Locker OS running 11M parcels a year, in real time |
上门派送失败是第一大成本驱动；承运商、商家与消费者各行其道，缺乏统一系统。Failed door delivery was the #1 cost driver; carriers, merchants and consumers ran on separate rails with no unifying system.
从零设计并上线完整的末端配送生态——订单、运输、快递柜与微仓融为一体。Design and launch a full last-mile ecosystem from zero — orders, transport, lockers and micro-warehouses as one.
OMS + TMS + Locker OS + 微仓 WMS + IoT 硬件 + 开放承运商 API；先验证经济模型、再自动化、再体验、最后借 800 家门店合作完成扩张。OMS + TMS + Locker OS + micro-warehouse WMS + IoT hardware + open carrier APIs; economics first, then automation, then experience, then expansion through 800 store partnerships.
2,000+ 网点、SLA +30%、运营成本 -25%，拓展 2 个新市场带来 +40% 营收。2,000+ points, SLA +30%, operating cost -25%, and +40% revenue expanding into two new markets.