Why spreadsheets fail garment OEM at 50+ styles為何試算表在50款以上時讓成衣OEM失靈为何表格在50款以上时让成衣OEM失灵

Every garment OEM company we meet started the same way: one merchandiser, a handful of customers, and a folder of Excel files that worked beautifully. The tech pack lived in one sheet, the BOM in another, the cost sheet in a third, and the person who built them held the whole picture in their head. At twenty styles a season, this is not a problem. It is, honestly, a perfectly good system.

我們接觸過的每一家成衣OEM公司，起點幾乎都一樣：一位跟單、幾個客戶、一個運作良好的Excel資料夾。TP在一個檔案，BOM在另一個，成本表在第三個，而建立這些檔案的人把全貌記在腦子裡。一季二十個款式，這不是問題——老實說，這是一套很好的系統。

我们接触过的每一家成衣OEM公司，起点几乎都一样：一位跟单、几个客户、一个运作良好的Excel文件夹。TP在一个文件，BOM在另一个，成本表在第三个，而建立这些文件的人把全貌记在脑子里。一季二十个款式，这不是问题——老实说，这是一套很好的系统。

Then the business grows. More customers, more seasons overlapping, a second merchandiser, then a third. Somewhere past fifty active styles, the folder of spreadsheets crosses an invisible line: it stops being a tool your team uses and becomes a liability your team services. Almost no one notices the moment it happens, because each individual file still looks fine. What breaks is everything between the files.

然後生意成長了。客戶變多、季節重疊、第二位跟單加入，然後第三位。在超過五十個進行中款式的某個時間點，那個試算表資料夾越過了一條看不見的線：它不再是團隊使用的工具，而變成團隊必須伺候的負擔。幾乎沒有人注意到那一刻，因為每個單獨的檔案看起來都還好。壞掉的是檔案與檔案之間的一切。

然后生意成长了。客户变多、季节重叠、第二位跟单加入，然后第三位。在超过五十个进行中款式的某个时间点，那个表格文件夹越过了一条看不见的线：它不再是团队使用的工具，而变成团队必须伺候的负担。几乎没有人注意到那一刻，因为每个单独的文件看起来都还好。坏掉的是文件与文件之间的一切。

The failure is structural失靈是結構性的失灵是结构性的

Garment OEM data has a shape that spreadsheets are bad at. A style is not one record — it is a family of related, versioned documents. The tech pack goes through revisions as fit comments come back. The BOM changes when a trim is substituted. The cost sheet must follow the BOM, and the quotation must follow the cost. Each document has a history that matters commercially: which version the customer approved, which version the factory quoted against, which version production actually used.

成衣OEM的資料有一種試算表天生不擅長的形狀。一個款式不是一筆記錄，而是一組互相關聯、各有版本的文件。TP隨著試身意見來回而改版；換一個輔料，BOM就變了；成本表必須跟著BOM走，報價又必須跟著成本走。每份文件的歷史在商業上都有意義：客戶核准的是哪一版、工廠是照哪一版報的價、生產實際用的又是哪一版。

成衣OEM的数据有一种表格天生不擅长的形状。一个款式不是一笔记录，而是一组互相关联、各有版本的文件。TP随着试身意见来回而改版；换一个辅料，BOM就变了；成本表必须跟着BOM走，报价又必须跟着成本走。每份文件的历史在商业上都有意义：客户核准的是哪一版、工厂是照哪一版报的价、生产实际用的又是哪一版。

In a spreadsheet operation, those relationships live in people's memories and in file names: FINAL_v3_updated_USE-THIS-ONE.xlsx. Version control is a naming convention. Cross-references are copy-paste. And copy-paste is where the money leaks out — a cost sheet quoting against last month's BOM, a purchase order keyed from a superseded trim list, a carry-over style duplicated from last season with one forgotten column still holding the old season's fabric.

在試算表營運裡，這些關聯存在人的記憶和檔名裡：FINAL_v3_更新_用這個.xlsx。版本控制是命名習慣，交叉引用靠複製貼上。而複製貼上就是錢漏出去的地方——成本表對著上個月的BOM報價、採購單照著已作廢的輔料清單打、延續款從上一季複製後有一欄忘了改、還留著舊季的面料。

在表格运营里，这些关联存在人的记忆和文件名里：FINAL_v3_更新_用这个.xlsx。版本控制是命名习惯，交叉引用靠复制粘贴。而复制粘贴就是钱漏出去的地方——成本表对着上个月的BOM报价、采购单照着已作废的辅料清单打、延续款从上一季复制后有一栏忘了改、还留着旧季的面料。

The five symptoms of the fifty-style wall五十款之牆的五個症狀五十款之墙的五个症状

1. The 'which file is current' tax. Before anyone can act, they first verify. Is this the BOM the customer saw? Did the factory get the revised spec? Multiply a few minutes of verification by every decision, every day, across the team, and you have hired an invisible employee whose only job is doubt.

一、「哪個檔案是最新的」稅。任何人行動之前都得先查證：這是客戶看過的那版BOM嗎？工廠拿到修訂後的規格了嗎？把每個決定前的幾分鐘查證，乘上每一天、每一位成員，等於雇了一位隱形員工，他唯一的工作就是懷疑。

一、「哪个文件是最新的」税。任何人行动之前都得先查证：这是客户看过的那版BOM吗？工厂拿到修订后的规格了吗？把每个决定前的几分钟查证，乘上每一天、每一位成员，等于雇了一位隐形员工，他唯一的工作就是怀疑。

2. Sample history in email. Seven rounds from proto to bulk, and the record of what was approved — and what the customer said in round three — is scattered across inboxes. When a fit dispute surfaces months later, the evidence is an archaeology project.

二、樣品歷史躺在郵件裡。從Proto到大貨七個輪次，核准了什麼、客戶在第三輪說了什麼，全散落在各人的收件匣。幾個月後版型出爭議，找證據像在做考古。

二、样品历史躺在邮件里。从Proto到大货七个轮次，核准了什么、客户在第三轮说了什么，全散落在各人的收件箱。几个月后版型出争议，找证据像在做考古。

3. Colorways as columns. Every added colorway widens the sheet; every material row must be checked against every colorway by eye. A missed cell — one material with no color assigned for one colorway — is invisible in a spreadsheet and very visible when the wrong goods arrive.

三、配色靠加欄位。每加一個配色，表就變寬一次；每一列物料都得用肉眼對每個配色檢查一遍。漏掉一格——某個物料在某個配色下沒指定顏色——在試算表裡看不見，等錯的貨到了就非常看得見。

三、配色靠加列。每加一个配色，表就变宽一次；每一行物料都得用肉眼对每个配色检查一遍。漏掉一格——某个物料在某个配色下没指定颜色——在表格里看不见，等错的货到了就非常看得见。

4. Retyping as a workflow. The customer sends a BOM as PDF. Someone retypes it. The factory sends a cost sheet as Excel. Someone retypes it. Every retype is unpaid work plus a fresh chance to introduce the error that the original file did not have.

四、重打資料變成工作流程。客戶寄來PDF的BOM，有人重打一遍；工廠寄來Excel成本表，有人再重打一遍。每一次重打都是無償工時，外加一次把「原檔沒有的錯誤」打進系統的機會。

四、重打资料变成工作流程。客户寄来PDF的BOM，有人重打一遍；工厂寄来Excel成本表，有人再重打一遍。每一次重打都是无偿工时，外加一次把「原档没有的错误」打进系统的机会。

5. The keyperson single point of failure. The operation depends on the two or three people who remember how the files relate. When they are on leave — or when they leave — the map of the business walks out the door with them.

五、關鍵人物的單點故障。整個營運依賴那兩三位記得檔案之間如何關聯的人。他們休假時——或離職時——整張營運地圖跟著走出大門。

五、关键人物的单点故障。整个运营依赖那两三位记得文件之间如何关联的人。他们休假时——或离职时——整张运营地图跟着走出大门。

What the alternative actually is替代方案究竟是什麼替代方案究竟是什么

The fix is not 'better spreadsheets' and it is not a generic ERP that treats a garment like a machined part. Across the industry, the systems that work share one property: the style is a single versioned record, and every other document derives from it. The tech pack, BOM, cost sheet, QC plan, and packing spec are tabs of one master, each with its own version history. The cost sheet is built from the BOM — change the BOM and the cost knows. The customer's file is imported, reviewed row by row, and committed as a version — not retyped. A carry-over style is a copy operation that brings the whole family along as a fresh draft.

解方不是「更好的試算表」，也不是把成衣當機械零件處理的通用ERP。綜觀業界，行得通的系統都有一個共同性質：款式是一份帶版本的單一主檔，其他所有文件都從它衍生。TP、BOM、成本表、QC、包裝規格是同一份主檔的分頁，各自帶版本歷史。成本表由BOM建立——BOM一改，成本就知道。客戶的檔案用匯入的：逐行檢查後存為新版本，而不是重打。延續款是一次複製動作，整組文件以新草稿一併帶入。

解方不是「更好的表格」，也不是把成衣当机械零件处理的通用ERP。纵观业界，行得通的系统都有一个共同性质：款式是一份带版本的单一主档，其他所有文件都从它衍生。TP、BOM、成本表、QC、包装规格是同一份主档的分页，各自带版本历史。成本表由BOM建立——BOM一改，成本就知道。客户的文件用导入的：逐行检查后存为新版本，而不是重打。延续款是一次复制动作，整组文件以新草稿一并带入。

This is also what makes the newer wave of AI capability real rather than decorative. Industry surveys of apparel digitalization keep landing on the same conclusion: AI is only as useful as the data structure underneath it. On a folder of disconnected files, AI can summarize a document. On a connected data model, it can review the operation — flag the order whose fabric has not arrived, draft the quotation in the customer's language, write the morning briefing from the system's own records.

這也是為什麼新一波AI能力能夠是真材實料而非裝飾。服裝業數位化的產業調查一再得出同一個結論：AI的用處取決於底下的資料結構。在一堆互不相連的檔案上，AI只能摘要一份文件；在一個互相連結的資料模型上，它能審閱整個營運——標出面料未到的訂單、以客戶的語言起草報價、根據系統自己的記錄寫出每天早上的簡報。

这也是为什么新一波AI能力能够是真材实料而非装饰。服装业数字化的行业调查一再得出同一个结论：AI的用处取决于底下的数据结构。在一堆互不相连的文件上，AI只能摘要一份文件；在一个互相连结的数据模型上，它能审阅整个运营——标出面料未到的订单、以客户的语言起草报价、根据系统自己的记录写出每天早上的简报。

If your team recognizes the five symptoms above, the spreadsheet folder has already crossed the line — the only question is how much longer you pay the tax before replacing the structure underneath. The companies that move early do not do it because Excel failed dramatically one day. They do it because they added up what doubt, retyping, and version archaeology were quietly costing every week.

如果您的團隊對上面五個症狀感到眼熟，那個試算表資料夾其實已經越線了——唯一的問題是，在更換底層結構之前，您還要繳多久的稅。及早行動的公司並不是因為某天Excel戲劇性地崩潰了，而是因為他們算了一筆帳：懷疑、重打與版本考古，每週都在安靜地吃掉多少成本。

如果您的团队对上面五个症状感到眼熟，那个表格文件夹其实已经越线了——唯一的问题是，在更换底层结构之前，您还要缴多久的税。及早行动的公司并不是因为某天Excel戏剧性地崩溃了，而是因为他们算了一笔账：怀疑、重打与版本考古，每周都在安静地吃掉多少成本。