The situation
An international apparel and footwear brand sold heavily into a market that had just tightened its import rules. Customs officers there could now detain goods if any input, at any tier, was linked to forced labor, and the burden of proving otherwise fell on the importer. A competitor had already had containers held at port.
The brand knew its 212 tier-one garment factories well and had audited most of their fabric mills. Beneath that, its knowledge came from supplier questionnaires. The autumn outerwear range, worth $64M, was due to ship in nine weeks.
The question
Can any input in the outerwear line be traced to a facility with indicators of forced labor? The answer had to meet the standard a customs authority would apply, which meant every link in the chain needed a source that could be shown to an official.
What we did
A supply chain lead, two trade data analysts, a corporate records analyst, two regional language specialists and a geospatial analyst worked the project, with a red-team reviewer assigned at the start. The client gave us its tier-one supplier list and nothing else. We contacted no suppliers.
1. Rebuilt the tiers from shipment records
Shipment and customs records showed which mills delivered fabric to the 212 factories, which spinners delivered yarn to those mills, and which processors delivered raw fiber to the spinners. Tenders, certifications and the suppliers' own published material filled the gaps. The map grew to 1,840 entities across four tiers.
2. Graded every link
Each relationship was graded as confirmed, probable or possible. By the end, 71 percent of links were confirmed by two or more independent sources. The rest were labeled for what they were.
3. Found the concentration nobody had seen
Three tier-one suppliers bought fabric through different trading companies. Registry filings, read in the original language, showed that all of it came from the same tier-three mill. The brand believed it had three independent sources. It had one.
4. Followed the mill downward
The mill's own filings listed two fiber processing facilities among its suppliers. Official notices and regional press, again in the original language, described workers being moved to both facilities under a government labor transfer program. Recruitment advertisements for the sites described restrictions on leaving the premises and the retention of identity documents.
5. Confirmed the sites
Commercial satellite imagery confirmed both facilities at their registered locations, operating, with dormitory blocks inside a fenced perimeter built in the same period. Assessed against the eleven recognized indicators of forced labor, five were documented at each site.
6. Tried to break it
The red-team reviewer argued that the facilities might supply the mill's other product lines, not the fabric our client bought. Shipment records showed that the fiber type and product codes moving from the facilities to the mill matched the outerwear fabric in the relevant months. What public records cannot do is follow one bale into one garment, and the report said so.
| Link | Evidence | Grade |
|---|---|---|
| Facilities to tier-three mill | Mill filings, shipment records | Confirmed |
| Mill to two trading companies | Shipment records, shared officers in registry filings | Confirmed |
| Trading companies to three tier-one suppliers | Shipment records, supplier certifications | Confirmed |
| Forced labor indicators at both facilities | Official notices, regional press, recruitment advertisements, imagery | Confirmed, five of eleven indicators |
| Facility output in the client's outerwear fabric | Matching fiber type and product codes by month | Probable |
What we found
- High confidence: two fourth-tier facilities with documented forced labor indicators supply the tier-three mill.
- High confidence: that mill supplies three of the client's tier-one suppliers through two trading companies.
- Moderate confidence: material from the facilities entered the client's outerwear line during the period reviewed.
Entity relationships can be confirmed from records. The path of a specific lot of fiber cannot, without the mill's internal production data. We reported the product-level link as probable and told the client exactly which documents a supplier would need to produce to disprove it.
The outcome
High-severity findings were reported in week three, as soon as they were confirmed. The full report followed in week five.
- The client put direct questions to the three suppliers, using our list of what a credible answer would contain.
- Two suppliers moved their orders to mills with documented traceability within five weeks.
- The third could not demonstrate separation and was phased out of the line.
- The $64M autumn range shipped on schedule, with an evidence file ready for customs. No shipment was detained.
- The map is now refreshed every quarter, and the client is told when a mapped entity is newly listed or reported.
What made the difference
The brand's audits were sound. They simply stopped where its contracts stopped. The exposure sat two tiers below the last place anyone had visited, behind trading companies that made one mill look like three.
Because every link carried its source and its grade, the compliance team could separate what was proven from what was probable, act on the first and ask precise questions about the second. The method is described in our guide, Past tier one.