concept
Getting Data Out of Deep-Tier Suppliers
Almost every blocked passport traces to a supplier who was never asked properly. How to structure requests that get answered, and what to do when they are not.
Supplier data collection succeeds when requests are grouped by attribute set rather than by product, sent to the person who holds the data rather than the account manager, and give an explicit route to answer that a value is not measured. Most failures are structural rather than about willingness.
What this gives you
The three mechanisms that get data out of deep-tier suppliers, what each costs in relationship terms, and the sequence that reaches a delegated act deadline.
Key takeaways
- The attributes a passport lacks usually sit at tier 2 or 3, where you have no direct contract and little influence.
- Grouping requests by attribute set rather than by product turns eleven identical chases into one.
- Without a route to say "we do not measure that", you receive estimates presented as measurements.
- Recording who asserted each value and when is what converts a returned number into evidence.
Passport programmes do not usually stall on software. They stall on a fibre origin figure held by a spinner in another country who has no contract with you, no compliance function, and no obvious reason to prioritise your spreadsheet.
That makes supplier data collection the critical path, and it is consistently the part least well designed. The failures are structural rather than about willingness, which is encouraging, because structural problems can be fixed.
Why visibility stops at tier 2
Most manufacturers know their tier-1 supplier well and their tier-2 supplier reasonably. Below that, the relationship is mediated: you know that your fabric comes from a mill, and the mill knows where the yarn came from, and nobody has ever needed to write it down.
The attributes regulation now asks for live precisely there. Process-stage country of origin, chemical inputs at dyeing, smelter energy mix, land geolocation — all of these sit below the tier where commercial relationships provide any real influence.
| Attribute | Typically held at | Your influence |
|---|---|---|
| Finished product weight and dimensions | Tier 1 | Direct contract |
| Component composition | Tier 1 to 2 | Direct or one step removed |
| Process-stage origin | Tier 2 to 3 | Indirect, through your tier 1 |
| Chemical inputs and treatments | Tier 2 to 3 | Indirect, often commercially sensitive |
| Raw material origin and geolocation | Tier 3 to 4 | Usually none directly |
| Recycled content certification | Tier 3 plus certifier | Through the certification scheme |
The four structural fixes
Four changes to how requests are constructed account for most of the improvement in response rate, and none require additional commercial influence.
- Group by attribute set, not by product. One mill supplying eleven of your styles should receive one request covering the fabric, not eleven covering each garment. Eleven identical requests produce one late reply and an irritated supplier.
- Send it to the person who holds the data. That is usually a production or technical manager, not the account manager whose job is to keep the relationship pleasant and who will forward it and forget.
- Name the regulation. "Regulation (EU) 2024/1781 requires this for our product group" gets a faster response than "for our sustainability programme", because the first is a fact about the world and the second is a request for a favour.
- Give a worked example per field. Ambiguity about units, boundaries or scope is what produces confident wrong answers, and it is entirely preventable.
What to do about the tier you cannot reach
Sometimes the holder of an attribute is genuinely beyond your reach. Three approaches work, in descending order of reliability.
The first is to route the request through your tier 1 as a contractual flow-down, making it their obligation to obtain rather than yours. This works where the tier 1 has the relationship and the commercial motivation, and fails where they are themselves a trader with no visibility.
The second is to use certification schemes as a proxy. A chain-of-custody certificate carries verified information about upstream stages without requiring you to contact them, which is what those schemes exist to provide.
The third is to change what you buy. Where an attribute is genuinely unobtainable and legally required, the material or supplier becomes non-viable for that market. That is an uncomfortable conclusion and occasionally the correct one, and reaching it early is much cheaper than reaching it after a delegated act applies.
Provenance is the deliverable
The output of collection is not a value. It is a value plus who asserted it, when, and by what method. Programmes that capture only the number find themselves unable to defend it eighteen months later when an auditor asks where it came from.
The test standard or calculation rule behind the figure.
Because screenings and certificates go stale.
A named organisation, not a department or a system.
Necessary, and on its own indistinguishable from an estimate.
Capturing those layers costs almost nothing at the moment the data arrives and is close to impossible to reconstruct afterwards, which makes it the highest-return field in the collection form after the not-measured option.
Running it as a campaign
Treating collection as a campaign rather than a per-product chase changes the operating model. A campaign has one attribute set, a defined supplier list, a due date and a response rate you can report on.
It also makes escalation proportionate. A supplier who has not responded to one clear, well-targeted request with a stated legal basis is in a different position from one who has ignored eleven confusing spreadsheets, and only the first case justifies raising it commercially.
Frequently asked questions
Why do suppliers not respond to data requests?
Usually because the request is unclear, sent to the wrong person, or duplicated across products. Refusal is rarer than assumed. A named technical contact, one request per attribute set, and a stated legal basis typically move response rates more than any commercial escalation.
What if a supplier genuinely does not measure something?
You need to know that, which is why the request must offer an explicit route to say so. A recorded gap can be planned around; an estimate presented as a measurement cannot, and it will fail verification later at a considerably worse moment.
How do we reach tier 3 suppliers we have no contract with?
Three routes, in descending order of reliability: flow the requirement down contractually through your tier 1 supplier, use chain-of-custody certification as a verified proxy for upstream stages, or accept that an unobtainable but legally required attribute makes that supply route non-viable for the market in question.
Should we use a supplier portal?
It helps at scale and hurts at the start. A portal adds an account and a login between a supplier and a five-field answer, which suppresses response rates for smaller suppliers. Structured email or a simple form usually outperforms a portal until volume genuinely justifies one.
How long does supplier data collection take?
One to two quarters for a first campaign, driven by how many tiers the request has to travel and whether upstream suppliers have to ask their own. It is the longest item in most passport timelines, which is why it should start before platform selection concludes.
What do we record alongside the value?
The asserting organisation, the date, the method or standard used, and where relevant the chain-of-custody model. Those four turn a number into evidence, cost nothing to capture at the point of receipt, and are effectively impossible to reconstruct a year later.
How often does collected data need refreshing?
On a change trigger rather than a calendar. A bill of materials change, a new supplier, or a candidate list update all invalidate specific attributes. Annual refreshes of everything waste effort while still missing the substance screening that went stale in March.
Sources
- Regulation (EU) 2024/1781 establishing a framework for ecodesign requirements — EUR-Lex, European Union, 2024-06
- Regulation (EU) 2023/1115 on deforestation-free products — EUR-Lex, European Union, 2023-06
Continue reading
- Getting fibre composition data rightA worked example of an attribute that must be assembled from several tiers.
- What data must a Digital Product Passport contain?The attribute families collection campaigns are built around.
- Verifiable Credentials explainedHow a returned value becomes a claim somebody else can check.
- Programme templatesA supplier data request structured the way this article describes.