Before you compare suppliers, answer these.比价之前，先回答这些问题。比價之前，先回答這些問題。

Start from the operation, not the equipment从业务问题出发，而不是从设备出发從業務問題出發，而不是從設備出發

No supplier can price a problem that has not been stated. These answers decide whether the project is measurable at all.没有人能为一个没被说清的问题报价。这一部分决定项目是否可以被衡量。沒有人能為一個沒被說清的問題報價。這部分決定項目是否可以被衡量。

• Which process is failing today — throughput, labour cost, accuracy or safety — and how is it measured?今天失效的是哪个环节——吞吐、人工成本、准确率还是安全？现在是怎么衡量的？今天失效的是哪個環節——吞吐、人工成本、準確率還是安全？現在是怎麼衡量的？
• What are current volumes, average and peak, and what growth must the solution survive over three to five years?当前业务量的均值与峰值是多少？未来三到五年，方案需要承受多大的增长？目前業務量的均值與峰值是多少？未來三到五年，方案需要承受多大的增長？
• Which constraint binds first: space, labour, systems or capital?最先触顶的约束是什么：场地、人力、系统还是预算？最先觸頂的約束是甚麼：場地、人力、系統還是預算？
• What has already been tried, and why did it fall short?之前尝试过什么？为什么没有解决问题？之前嘗試過甚麼？為甚麼沒有解決問題？

Site conditions decide what is buildable现场条件决定方案边界現場條件決定方案邊界

These constraints filter the shortlist long before product selection.这些约束在选型之前就已经筛掉了大部分方案。這些約束在選型之前就已經篩掉了大部分方案。

• Aisle widths, clear height, floor flatness and load rating — are these measured, not estimated?通道宽度、净高、地面平整度与承重：是实测数据，还是估算？通道寬度、淨高、地面平整度與承重：是實測數據，還是估算？
• Power capacity, charging locations — and is network coverage verified in every operating zone, including corners and mezzanines?供电容量与充电位置如何？各作业区的网络覆盖是否已实测，包括角落和夹层？供電容量與充電位置如何？各作業區的網絡覆蓋是否已實測，包括角落和夾層？
• Which fire, egress and safety rules limit where equipment can sit, move or charge?哪些消防、疏散与安全规定限制了设备可以停放、行驶或充电的位置？哪些消防、疏散與安全規定限制了設備可以停放、行駛或充電的位置？
• What are the shift patterns and crew levels, including nights and seasonal peaks?班次与人员配置如何？包括夜班和旺季峰值。班次與人員配置如何？包括夜班和旺季峰值。
• What can be modified physically — and what genuinely cannot?哪些现场条件可以改造？哪些确实不能动？哪些現場條件可以改造？哪些確實不能動？

Most delays start at the interface大多数拖延始于接口大多數拖延始於接口

In warehouse automation, the equipment is rarely the unknown. The interface between existing systems and new ones is.在仓储自动化里，设备本身通常不是未知项，未知项是现有系统与新系统之间的接口。在倉儲自動化裡，設備本身通常不是未知項，未知項是現有系統與新系統之間的接口。

• Which system owns task allocation — WMS, WCS, or the fleet manager?任务分配由谁负责——WMS、WCS，还是车队管理系统？任務分配由誰負責——WMS、WCS，還是車隊管理系統？
• What interfaces exist today — API, flat file, middleware — and who is able to expose them?现有接口是什么形式——API、文件还是中间件？由谁负责开放？現有接口是甚麼形式——API、文件還是中間件？由誰負責開放？
• What happens to in-flight tasks when a network handoff drops?网络切换中断时，执行中的任务如何处理？網絡切換中斷時，執行中的任務如何處理？
• Which record is the source of truth: order, inventory or location?哪一份记录是唯一事实来源：订单、库存还是库位？哪一份記錄是唯一事實來源：訂單、庫存還是庫位？
• Who may change an interface after go-live, and how is the change tested before release?上线后谁有权变更接口？变更在上线前如何测试？上線後誰有權變更接口？變更在上線前如何測試？

Someone must own the seams接缝处必须有人负责接縫處必須有人負責

When three vendors deliver three subsystems, failures live between them. Name the owner before signing, not after a breakdown.当三家供应商交付三个子系统时，故障往往发生在它们之间。责任要在签约前写明，而不是在故障之后再讨论。當三家供應商交付三個子系統時，故障往往發生在它們之間。責任要在簽約前寫明，而不是在故障之後再討論。

• When fleet, WMS and WCS come from different vendors, who owns integration?当车队、WMS 与 WCS 来自不同供应商时，集成由谁负责？當車隊、WMS 與 WCS 來自不同供應商時，集成由誰負責？
• Who is accountable for a failure that spans two vendors?跨两家供应商的故障，谁负责到底？跨兩家供應商的故障，誰負責到底？
• What is the escalation path at 03:00 on a Sunday?周日凌晨三点出问题时，升级路径是什么？週日凌晨三點出問題時，升級路徑是甚麼？
• Which party signs off each subsystem, and against what evidence?每个子系统由谁签字确认？依据什么证据？每個子系統由誰簽字確認？依據甚麼證據？
• What are the commercial terms when responsibility is shared?责任共担时的商务条款是什么？責任共擔時的商務條款是甚麼？

"Passed the pilot" is not a definition"试点通过了"不是定义「試點通過了」不是定義

An acceptance criterion is a measurement, a period and an owner. If it cannot be written that way, it cannot be enforced.验收标准是一个测量值、一个时间段和一个责任人。写不成这三样，就无法执行。驗收標準是一個測量值、一個時間段和一個責任人。寫不成這三樣，就無法執行。

• What does "passed" mean: peak throughput, availability at full crew, or accuracy?"通过"指什么：峰值吞吐、满员可用率，还是准确率？「通過」指甚麼：峰值吞吐、滿員可用率，還是準確率？
• Over what period, and under which conditions — normal weeks, peak, or stocktake?在什么周期、什么工况下考核——平时、旺季，还是盘点期？在甚麼週期、甚麼工況下考核——平時、旺季，還是盤點期？
• Measured by whom, with what data, and recorded where?由谁测量？用什么数据？记录在哪里？由誰測量？用甚麼數據？記錄在哪裡？
• If the target is missed: remedy, retest, or exit?未达标时怎么办：整改、复测，还是退出？未達標時怎麼辦：整改、複測，還是退出？
• Which criteria are contractual, and which are aspirational?哪些标准写入合同？哪些只是期望？哪些標準寫入合同？哪些只是期望？

The record of what actually happened真正发生了什么，要留下来真正發生了甚麼，要留下來

Deployment is where the assumptions get tested. Recording the outcome is what makes the next project cheaper to define.部署是检验假设的环节。把结果记录下来，下一个项目的需求定义才会更省力。部署是檢驗假設的環節。把結果記錄下來，下一個項目的需求定義才會更省力。

• Who records planned versus actual after go-live?上线后，计划值与实际值由谁记录？上線後，計劃值與實際值由誰記錄？
• Which assumptions in the brief proved wrong, and how quickly was that discovered?需求文件里哪些假设后来被证明是错的？多久之后才被发现？需求文件裡哪些假設後來被證明是錯的？多久之後才被發現？
• How does that record flow into the next site or the next phase?这份记录如何回流到下一个现场或下一期项目？這份記錄如何回流到下一個現場或下一期項目？
• If you scoped this again tomorrow, what would you change?如果明天重做一次需求定义，你会改什么？如果明天重做一次需求定義，你會改甚麼？

What this checklist is not这份清单不是什么這份清單不是甚麼

• Not a procurement document, and not a vendor list.不是采购文件，也不是供应商名单。不是採購文件，也不是供應商名單。
• It does not choose technology. It makes requirements comparable.它不替你做选型，只让需求变得可比较。它不替你做選型，只讓需求變得可比較。
• Answered honestly, it will disqualify some options early. That is the point — it is cheaper before a purchase order than after.认真回答后，一些选项会提前出局。这正是目的——在下单之前出局，比之后便宜得多。認真回答後，一些選項會提前出局。這正是目的——在下單之前出局，比之後便宜得多。