The FACT-M Framework
FACT-M replaces subjective impressions with five concrete, evidence-based criteria. Use it to score every candidate factory on the same scale — then compare apples to apples.
audit
trail
match
clarity
evidence
F — Factory Audit
What it measures: Does the physical operation match the description?
Verification question: Can the supplier provide a recent third-party audit report (not a summary photo) you can share with your compliance team?
A — Audit Trail
What it measures: Can the supplier trace materials and production steps for a specific order?
Verification question: If you request a lot-trace for a sample batch, how far back can they document raw-material origin?
C — Capability Match
What it measures: Can this factory produce your product, not just a product?
Verification question: Can they show a production sample of a construction similar to yours (e.g., structured 6-panel cap with a specific closure type)?
T — Terms Clarity
What it measures: Are commercial terms documented before you commit?
Verification question: Is the MOQ, sample-fee policy, payment schedule, and defect-replacement clause provided in a written quotation — not a chat message?
M — Manufacturing Evidence
What it measures: Is there proof that sample quality reflects bulk quality?
Verification question: Can the supplier share inline inspection records from a past production run, with photos, not just a final-pass rate?
Decision Table: Scoring a Hat Manufacturer
Score each dimension on evidence level. A supplier with "Documented" across all five dimensions is a stronger candidate than one with "Claimed" in three and "Missing" in two — regardless of how polished the website looks.
| FACT-M dimension | Documented (3 pts) | Claimed (1 pt) | Missing (0 pts) | Weight |
|---|---|---|---|---|
| F — Factory audit | Third-party audit report available for review | "We are audited" with no report | No audit reference | 2× |
| A — 审计跟踪 | 特定订单的批次追溯记录已记录在案 | “我们追踪材料”，但没有举例子。 | 未提及可追溯性 | 1.5倍 |
| C — 能力匹配 | 可提供类似结构的生产样品 | “我们生产帽子”，但没有可比样品。 | 没有相关样本 | 2× |
| T — 术语清晰度 | 提供包含所有条款的书面报价。 | 聊天或电子邮件中的部分条款 | 无正式条款 | 1.5倍 |
| M — 制造证据 | 附照片的在线检测记录 | “我们检查每一件产品” | 没有检查证据 | 2× |
总分：30 分。使用此分数对入围工厂进行排名——不要将其作为通过/不通过的评判标准，因为各个工厂的权重应反映您自身的风险优先级。
签字前需要询问的审批问题
- 工厂审核： “您是否可以授权第三方进行审核或分享最近的审核报告？我可以核实审核员的资质吗？”
- 审计追踪： “您能带我了解一下一个已完成订单的批次追溯流程吗？从原材料到发货？”
- 能力匹配： “您能否制作一个[特定结构]的样品，并记录该样品的制作步骤？”
- 条款说明： “能否提供一份形式发票，其中应明确最低订购量、样品费、付款节点、缺陷更换条款和取消条件？”
- 生产证据： “您能否分享一下与我的产品类似的生产批次的在线检验记录（包括照片）？”
如果对上述任何问题的回答是“我们不这样做”或“这不是标准做法”，请记录下这种差距。这种差距或许可以接受，也或许会成为交易的致命缺陷。关键在于，你现在知道这种差距确实存在。
营销文案隐藏的危险信号
- 没有第三方审计报告。拒绝提供审计报告的工厂未必就是不诚实的，但这确实是一个你必须考虑的风险。
- 条款仅存在于聊天记录中。如果最低起订量、定价和缺陷处理政策仅存在于微信聊天记录中，则没有可供争议解决的参考文件。
- 没有可比产品的生产样品。拥有500种款式的产品目录并不能证明他们能够按照您的规格生产您想要的款式。
- 在线检测记录属于机密信息。如果供应商无法共享已编辑的检测记录，则必须提供相关信息。
