Most supply chain risk is found where nobody has a contract. The buyer knows tier one. Tier one knows tier two. By tier three, knowledge has become assumption, and assumption is what regulators have stopped accepting.
Why questionnaires stop working below tier one
A questionnaire travels down a chain only as far as each supplier is willing to pass it, and comes back with whatever each one chooses to say. A supplier that relies on a problematic source has every reason to answer vaguely. A supplier that does not know its own sub-suppliers will answer confidently anyway.
Audits are better and narrower. An auditor sees one site on one day and has no view of where that site's inputs came from. Neither tool was designed to answer the question now being asked of importers: can you show, with evidence, what lies beneath your suppliers?
Public records can go a long way toward answering it, because goods crossing borders, companies registering and governments regulating all leave records behind.
Six families of record
| Record family | What it shows | Its main weakness |
|---|---|---|
| Shipment and customs records | Who shipped what to whom, when, by which route | Coverage varies by country. Domestic movements are invisible. |
| Corporate registries | Ownership, officers, addresses, subsidiaries | Nominees and secrecy jurisdictions hide control. |
| Tenders and procurement awards | Named sub-suppliers and contract values | Public sector buyers only. |
| Certifications and product registrations | Named facilities, processes and source materials | Can be out of date or cover a single product. |
| Suppliers' own publications | Customers, partners, factory locations, capacity | Written to impress. Claims need independent support. |
| Local-language press and official notices | Labor programs, penalties, disputes, ownership changes | Requires native reading and knowledge of the place. |
No family is sufficient alone. A shipment record proves that goods moved between two names. A registry filing proves who stands behind those names. A local notice explains what happens at the site. The map comes from reading them against each other.
The method, in five steps
1. Fix the starting set
Begin with the tier-one supplier list for one product line, and resolve each supplier to a legal entity with a registration number. Trading names are not enough. A surprising share of later errors start with a tier-one supplier matched to the wrong company of a similar name.
2. Work downward one tier at a time
For each entity, collect inbound shipments of the relevant materials over a fixed period, then repeat for each sender. Resist the temptation to jump ahead. Each tier should be reviewed before the next is built on it, because an error at tier two multiplies at tier four.
3. Resolve entities before counting them
The same mill may appear under a transliterated name, an abbreviated name, a trading company and a parent. Until those are resolved into one entity, the map will show diversity where there is concentration. This is usually where the most important finding appears: several suppliers that believe they are independent and share a single source.
4. Tie risks to sites, not names
A company may operate a dozen facilities, one of which carries the risk. Registrations, permits and satellite imagery establish which facility a record refers to, whether it exists at the stated location and whether it is operating.
5. Research the high-risk nodes in depth
Only now is it worth spending analyst time on ownership, labor sourcing and local reporting, and only for the entities the map shows actually matter to the product line.
Grade every link
A supply chain map drawn with uniform lines is making a claim it cannot support. Some links are documented by several independent records. Others are inferred from a single shipment. A reader deciding whether to drop a supplier needs to see the difference.
| Grade | Standard | Example |
|---|---|---|
| Confirmed | Two or more independent records from different families | Repeated shipments, plus the supplier named in the buyer's certification |
| Probable | One strong record, or several of the same family, with nothing contradicting | Regular shipments of the right material in the right period |
| Possible | An indication that fits but could have other explanations | A shared address, or a single shipment years ago |
In a well-documented product line we expect around two thirds of links to reach confirmed. The remainder should be labeled, not removed and not promoted.
Five errors that produce false findings
Name collisions
Common company names recur across provinces and countries. A match on name alone, without a registration number, an address or an officer in common, is a possible link at best.
Mistaking a trading company for a manufacturer
Shipment records name the shipper, which is often an intermediary. The goods were made somewhere else. Treating the trading company as the origin ends the trace one step too early.
Stale relationships
A relationship that ended three years ago is history, not exposure. Every link needs a date range, and findings should rest on the period that matters to the decision.
The shared address
Industrial parks and registered agent offices house hundreds of unrelated companies. Co-location is a reason to look further and never a finding in itself.
Translation by machine
Official labor program notices, court records and local reporting use terms that general translation tools render blandly or wrongly. The difference between a job fair and an organized labor transfer can rest on one phrase.
Assessing labor conditions from a distance
Public sources cannot replace a worker's own account. They can establish whether recognized indicators of forced labor are documented at a site: recruitment through state transfer programs, fees that create debt, retention of identity documents, restrictions on leaving the premises, isolation in guarded housing.
We report indicators, with the source for each, and we do not convert them into a verdict. Whether a given set of indicators meets a legal threshold is a question for the client's counsel and, ultimately, for an authority.
What public records cannot show
Be clear about the limit before a client relies on the map. Records establish relationships between entities. They do not follow a specific lot of material through a mill and into a specific product. That requires the supplier's own production data.
When a product-level link is probable but not confirmed, the useful output is a list of the documents a supplier would need to produce to prove separation: purchase records, production batch records and transport documents for the period in question. A supplier that can produce them has answered the question. One that cannot has answered it too.
Keeping the map alive
Suppliers change their own suppliers every season. A map is accurate on the day it is delivered and decays from then on. At a minimum, refresh shipment data quarterly, re-screen every mapped entity against sanctions and enforcement lists as they change, and treat any new intermediary on an established route as a reason to look again.