An H&M supplier list identifies the factories the group discloses. It can show where those sites are and what work they do. It cannot answer every question about a garment or the people who made it.

That does not make the list unhelpful. It makes careful reading essential. A public factory record is a useful starting point for a student, buyer, journalist or customer. Its value depends on the question being asked and on the other evidence used with it.

This guide explains how to read H&M supplier transparency in practice. The sources were reviewed in October 2026. Company statements are attributed to H&M Group. The sample records below are invented. So are the figures used in the calculations. Neither describes a named factory.

Start with the scope of the disclosure

H&M Group says independent suppliers and manufacturers make its products. The group does not own those factories. Its supply-chain page describes a group supplier list, first published in 2013. The list identifies tier 1 manufacturing and processing sites. It also includes available details on certain fabric suppliers. H&M Group supply-chain disclosure

The word “group” matters. The corporate group publishes the disclosure. It does not map every product sold by the H&M brand. Before using a record, write down what the source actually covers.

Three questions will prevent many errors:

  • Which company or group published the information?

  • Which stages and products are included?

  • Which date or reporting period applies?

Keep the answers beside the data. A downloaded file can become detached from its explanatory page. Without the notes, a reader may later mistake a partial scope for a complete supply chain.

Choose the question before choosing the data

Suppose you want to know whether a named factory appears in a published list. A site name, address and date may answer that narrow question.

Now suppose you want to know where the cotton in one shirt was grown. A list of garment factories will not usually establish that link by itself. You need evidence connecting the item, its material and the relevant production stages.

A third question might concern wages at a particular site. A factory's presence in a list does not supply payroll evidence or worker testimony.

These are different research tasks. Treating them as one task creates conclusions that the data cannot support.

Supplier, factory and production stage are different ideas

A supplier is a business relationship. A factory is a physical site. A production stage describes an activity. One supplier can be connected to several sites, and a site can perform more than one activity.

H&M's own explanation distinguishes manufacturing factories, processing factories and fabric mills. It also warns that supply chains vary by material and product. Use the tier definitions given in that system. Other organisations may label their stages in a different way. H&M Group's explanation of supply-chain stages

Keep the business name and site name in separate fields. Give the address and activity their own fields too. Combining them into one free-text field makes later comparison harder.

Term in a review

What it should identify

Question to ask

Supplier

The business relationship

Is this the contracting entity or another group company?

Factory or facility

A specific site

Is the address detailed enough to distinguish it?

Activity

The work performed

Does the record cover sewing, processing or another stage?

Product connection

The item or order link

Is that connection directly documented?

Disclosure date

The time boundary

When was this information current?

This is a proposed reading table, not a reproduction of the current H&M download.

A place name alone is a weak identifier. Two sites may have similar names. A company may change its name. An address may be translated or shortened. Resolve these issues before counting unique factories.

Transparency and due diligence answer different questions

Transparency makes information available. Due diligence concerns how a company identifies, addresses and follows up on risks and harm.

The OECD describes due diligence as a process. It helps businesses understand and address their impacts on workers, human rights and the environment. Its garment and footwear guidance provides a common framework for that work. Publishing a list can support the process, but the process extends beyond publication. OECD guidance for garment and footwear supply chains

H&M Group's due-diligence page describes risk assessment, action, monitoring, communication and remediation. It also discusses the use of local teams, worker input and grievance channels. These are the company's descriptions of its approach. To assess effectiveness in a specific case, seek evidence of what happened after a concern was raised. H&M Group due diligence

A useful review separates four levels of evidence:

A published policy states an expectation.

A process description explains how the company intends to act.

An activity record shows that a stated action took place.

An outcome record helps show what changed for the people affected.

A policy can be clear while an outcome remains uncertain. An activity count can rise without proving that the underlying problem has been resolved. This is why each level should be labelled.

Why a public supplier list still has value

A public record can make a relationship easier to investigate. It can help different parties identify the same site and direct questions to the relevant businesses.

Fashion Revolution explains one use of supplier disclosure. Workers and their representatives can identify buyer relationships. They can then seek responses to problems. Its historical list also gives a clear limit. Inclusion does not endorse a company's ethical conduct. That article's list dates to April 2020. It explains this principle but cannot serve as a current brand directory. Fashion Revolution on supplier transparency

The practical benefit is reduced uncertainty about where to start. First identify the site. A researcher can then check reports, ask focused questions or compare records over time.

Still, availability is only the first step. Information must also be clear enough to use. An incomplete address or an unexplained scope can leave a public record difficult to interpret.

A strong review therefore asks both whether information is published and whether it answers a defined question.

Read dates before comparing figures

A webpage, a supplier file and an annual report can have different time boundaries. A recent page view does not mean every number on the page describes that same day.

H&M's supply-chain page dates several figures to January 2026. It says the supplier list is updated quarterly. Trade-union data within that list is updated yearly. Those footnotes show why a single “last checked” date is not enough. H&M Group supply-chain notes

Record at least three dates where available:

The date the source says the data covers.

The date the source was published or updated.

The date you accessed or downloaded it.

If a date is missing, say so. Do not quietly replace it with the date of your review.

H&M's reporting page gives the period for its 2025 annual and sustainability report. It runs from December 1, 2024, through November 30, 2025. It also explains that reporting structures and boundaries can change. A year-on-year comparison should check those changes before treating two figures as directly comparable. H&M Group sustainability reporting

A simple example of a misleading comparison

Imagine a January list contains 900 disclosed sites. An October list contains 940. It is tempting to say that the company's production network grew by 40 sites.

That conclusion may be wrong. The later list could cover a newly disclosed stage. It could correct duplicates. It might add sites that were already active but absent from the earlier list.

The observed fact is that the published count changed. The business explanation requires further evidence.

This distinction is especially important when a report changes its definitions. A break in a data series should remain visible.

Cross-checking a factory record with Open Supply Hub

H&M points readers toward Open Supply Hub as another place to find supplier information. Open Supply Hub's FAQ explains who can contribute records. It also says historical affiliations can remain visible. Some public lists are uploaded by the platform's team. Others are managed by the named organisation. The FAQ explains this distinction. Open Supply Hub FAQ

Use the platform as a source of attributed records. Check who supplied the data and which date applies. Confirm that the site matches the record you started with.

A useful cross-check follows this sequence:

  • Compare the country and full address.

  • Compare the facility name and known name variants.

  • Check any stable facility identifier shown.

  • Read the contributor and list information.

  • Note the date and any stated relationship status.

  • Record unresolved differences instead of forcing a match.

A matching name is a clue. A well-supported site match is stronger. Neither by itself proves that one specific garment passed through that site.

If two records disagree, preserve both versions and state the conflict. A silent merge may create a neat spreadsheet while hiding the most important research issue.

What a disclosure percentage actually means

A percentage is meaningful only when its numerator and denominator are clear.

Suppose a hypothetical brand says it has disclosed 90% of its supply chain. That phrase leaves several questions unanswered. Does it mean 90% of suppliers, factories, purchase value, product units or a defined production stage?

The following example shows why the choice matters. None of the figures describe H&M.

A brand buys 1,000,000 garments from 100 factories. It discloses 90 factories. Those 90 factories make 800,000 of the garments.

Factory-count coverage

Factory-count coverage = (Disclosed factories ÷ Total factories in scope) × 100

Factory-count coverage = (90 ÷ 100) × 100 = 90%

Unit-based coverage

Unit-based coverage = (Units from disclosed factories ÷ Total units in scope) × 100

Unit-based coverage = (800,000 ÷ 1,000,000) × 100 = 80%

Both results are correct under the stated assumptions. They answer different questions.

Neither result tells us the share of raw materials traced to origin. Neither measures the quality of working conditions. The scope of the denominator determines what can be claimed.

What happens when the total is unknown?

Imagine you can count 90 public factory records but cannot verify the total number in scope. You can report 90 disclosed records. You cannot calculate a defensible coverage percentage.

Do not use the published records as both the numerator and denominator. Doing so would produce 100% by construction and reveal nothing about missing records.

Keep group and brand denominators separate

A group may publish data across several brands. A product-specific review may cover one category or market. Do not mix a group numerator with a narrower denominator. Each number may be accurate, yet the resulting percentage can mislead.

Before calculating, write the scope in words. If the two parts do not describe the same population and period, stop and resolve the mismatch.

H&M supplier transparency — hm supplier transparency factory disclosures

Working conditions need evidence beyond a site name

H&M Group's working-conditions page sets out its expectations for suppliers. It also describes local engagement and safety programmes. The company also has a human-rights page. It discusses its responsibility to address risks within its operations and value chain. These statements help a reader understand the company's stated approach. They do not establish the conditions experienced by every worker at every site. H&M Group working conditions H&M Group human rights

When investigating a specific concern, ask what evidence would answer it.

Question

Evidence that may help answer it

What a supplier-list entry cannot establish

Were workers paid correctly?

Relevant payroll records and worker evidence

That all payments were correct

Was a safety issue corrected?

Dated findings, corrective action and follow-up

That the issue is closed

Could workers raise concerns?

Accessible channels and evidence of their use

That a channel was effective

Did a buyer relationship end?

Dated relationship and exit information

Why the relationship changed

Was one order made at the site?

Order and production records

A product-specific link

The table is a research guide. Access to some records may be restricted, and personal worker data needs appropriate protection. A public article should report only what its evidence supports.

Factory evidence needs to be organised. ApparelDiary explains this in its guide to buyer audit preparation and traceability records. Its quality-audit discussion should not be mistaken for proof of labour conditions.

Read wage disclosures with particular care

A wage figure needs a label. Is it basic pay or total earnings? Does it describe a legal minimum, a negotiated rate or a living-wage benchmark? Those terms are not interchangeable.

H&M's wage page states what it expects from suppliers. It describes factory data collection, checks and corrective action. The group says it does not pay these garment workers directly. Independent suppliers employ them. It also recognises that factories can serve several buyers. H&M Group wage approach

A reader should ask what the reported wage includes and which workers it covers. Check the period, hours, currency and whether the figure is an average. An average can conceal variation between roles or groups.

Do not infer that all workers receive a living wage because a supplier signs a code. Do not infer the opposite from a missing wage field in a public list. Each claim needs its own evidence.

If the source provides only a policy, report it as a policy. If it reports a measured outcome, preserve the method and scope. If a benchmark is used, identify who produced it and the location and year to which it applies.

A practical desk review of one factory record

The following fictional case shows how to keep a conclusion within the evidence.

A researcher finds “Example Garments Ltd” in a group supplier file. The record gives a city, an industrial-area address and a garment-manufacturing activity. A second source has a similar name but a different street number.

Establish identity first

Do not assume that the records describe the same site. Compare the full address, legal name and any facility identifier. Check whether one source describes an office while the other describes a production site.

If the match remains uncertain, write “possible match; not confirmed.” This is more useful than an unsupported yes-or-no answer.

State the relationship narrowly

The list may identify a supplier facility at a stated date. Report that dated relationship as the finding. It does not establish the exact share of the site's business supplied to the group.

It also does not establish that a particular shirt was made there. For that, seek a product or order connection.

A production BOM can define the materials needed for an order. It cannot prove their full physical route by itself. The relevant production and transaction records must connect those materials to the actual item.

Separate the unresolved questions

The researcher might finish with three findings:

The site appears in the specified public list.

A possible second record needs identity confirmation.

No product-level or wage outcome conclusion was established.

This is a useful result. It gives the next researcher a clear starting point and avoids converting gaps into claims.

Build an evidence register that another person can follow

A good review should be reproducible from its notes. Keep the source URL, file name, data date, access date and relevant passage or field.

Add a column for the claim you intend to make. This forces you to check whether the source supports the words, not merely the general subject.

Then add a limitation column. Examples include:

Group-level data; brand split unavailable.

Factory count; production volume not stated.

Historical relationship; current status not confirmed.

Policy description; outcome evidence not reviewed.

Similar site name; identity unresolved.

Use plain language. A long list of technical labels will not help if another reader cannot understand the boundary.

Keep your own calculations separate from published figures. Mark hypothetical examples clearly. If you correct a data error, preserve the original value and record the reason for the change.

Common mistakes and how to correct them

Counting rows as unique factories

Look for repeated sites. One site might perform several activities or appear in more than one list. Decide the counting unit before reporting a total.

Treating a missing entry as proof of misconduct

A missing record can have several explanations. Identify the disclosure scope and date before drawing a conclusion.

Treating inclusion as a seal of approval

A disclosed relationship creates a path for scrutiny. Evaluate working conditions and outcomes with appropriate evidence.

Mixing reporting periods

Use comparable dates or explain why the comparison is limited. A quarterly list and an annual report may describe different populations.

Treating an average as every worker's experience

Ask how the average was calculated and what variation it hides. Preserve the difference between aggregate data and individual outcomes.

Repeating a target as if it were achieved

A target describes an intended result. Look for dated evidence of progress and completion before changing the wording.

Assuming more fields always mean better evidence

More information can help, but clarity, accuracy and relevance matter. A well-defined field is more useful than several unexplained columns.

Comparing two published lists without losing the changes

A net change can hide a much larger set of movements. Consider another fictional example. A first list contains 100 unique sites. A later list retains 90 of those sites, removes 10 and adds 15. The later count is 105.

Later disclosed count = Retained sites + Added sites = 90 + 15 = 105 sites

Net change = Added sites − Removed sites = 15 − 10 = 5 sites

The net increase is five sites, but 25 records changed membership. Reporting only the net figure would hide that difference.

Create three review groups: retained, added and removed. Keep a fourth group for uncertain matches. An uncertain match should not become an addition or removal until the identity issue is resolved.

Then investigate the reason for each material change. An added record may reflect a new relationship, wider disclosure or corrected data. A removed record may reflect a relationship change, a duplicate correction or a change in scope. The lists alone do not establish which explanation applies.

For retained sites, compare fields as well as membership. A changed activity or address can matter even when the site count stays the same. Preserve the earlier record so another reader can follow the comparison.

Begin the final statement with the published changes. Which records appeared or disappeared? Which fields changed between the two dated files? Explain business causes only where separate evidence supports them. This method keeps the review accurate without ignoring useful signals.

Questions worth asking after the first review

Once the basic scope is clear, deeper questions become more precise.

How is a newly added factory approved? What happens when a concern is found? How are workers involved in checking a remedy? How are relationship changes reflected in public records? Which stages remain difficult to trace?

These questions should follow the evidence already collected. Avoid sending a broad list of demands when one unresolved fact would clarify the next step.

For an academic assignment, choose one question and explain the evidence gap. For a sourcing review, connect the question to the intended decision. For a public article, make the distinction between verified facts and open questions visible to readers.

Frequently asked questions

Does H&M own its garment factories?

H&M Group says independent suppliers and manufacturers make its products. It does not own those factories. The exact relationship for a site should be checked in the relevant source.

Does a factory list trace every garment to raw material origin?

A factory list and a product-level chain are different records. Establish the item and material connections before making an origin claim.

Can two transparency percentages both be correct?

Yes. One may measure factory count while another measures units, value or a different stage. Read the denominator and period.

Does Open Supply Hub prove a current buyer relationship?

Read the contributor, dates and status. Its records can include historical information. A record should be interpreted within that context.

Is public disclosure enough to judge a brand?

It is a useful part of the evidence. A broader judgment also needs information on risks, actions and outcomes, with clear limits on what has been verified.

Conclusion

H&M supplier transparency is most useful when read as a set of scoped, dated disclosures. Start with the entity, production stage and time period. Resolve site identities before counting. Keep product tracing, worker outcomes and public relationship records distinct.

A useful final note might read: the two dated lists show the same named site, but its address field has changed. The reviewer has saved both records and asked whether this is a move or a data correction. That statement is narrow, useful, and open to checking. It neither clears the factory of every concern nor invents a concern from the changed address. This is the discipline that makes public supplier data worth using.

Disclosures are easier to read alongside buyer-side practice, so see buyer audit preparation, buyer requirements, buying houses and buying offices and the buyer's journey in Bangladesh.

References