The Hidden Subcontracting Report

Hidden subcontracting sits in the gap between the supply chain a buyer can see and the production network that actually completes an order. The risk emerges when the commercial relationship and the physical production pathway no longer match.

The strongest direct evidence in the dataset comes from a multi-country study covering more than 30,000 orders. In that study, 36% of orders were subcontracted without authorization. The factory population was not uniform: 56.6% of factories never subcontracted, 31.9% sometimes subcontracted and 11.5% always subcontracted. Those categories matter because sometimes-subcontracting factories accounted for 57.9% of observed orders, making intermittent behavior more important than a simple list of permanently high-risk suppliers would suggest.

The wider evidence base shows why hidden production matters beyond purchasing control. Global Better Work coverage exceeds 2,200 factories and 3.7 million workers across 11 countries, while the global textiles and garment workforce is around 90 million people. A well-audited approved factory does not automatically provide visibility into an undisclosed workshop producing part of the same order.

This report follows hidden subcontracting from order behavior and factory capacity through monitoring coverage, labour conditions, country manufacturing structures, gender, global-supply-chain exposure, traceability and due diligence. The statistical objective is to separate visible supplier relationships from verified production.

Executive Hidden Subcontracting Benchmarks

The numbers defining the visibility gap

The core order dataset establishes the scale of the issue. More than 30,000 orders were observed, and 36% were reported as subcontracted without authorization. Across 226 factories represented by the reported behavior counts, 98 factories had used unauthorized subcontracting at least once, equivalent to 43.4%. Factories that had ever subcontracted accounted for 64.3% of study orders.

Factory behavior separates the risk into three operational patterns. The 128 factories that never subcontracted represented 56.6% of factories and 35.6% of orders. The 72 sometimes-subcontracting factories represented 31.9% of factories but 57.9% of orders, with an unauthorized-subcontracting rate of 51.7%. The 26 always-subcontracting factories represented 11.5% of factories and 6.4% of orders, with a 100% subcontracting rate. A monitoring strategy focused only on suppliers with a permanent history would therefore miss a large volume of intermittent risk.

The same dataset contains strong evidence that hidden subcontracting leaves measurable signals. Out-of-sample order prediction accuracy exceeded 80%, while supplier-level prediction accuracy was about 70%. These figures do not turn a model score into proof of wrongdoing. They show that order characteristics, supplier behavior and production conditions can help prioritize verification. Data can narrow the search area; facility-level evidence must still establish what happened.

The operating environment is highly variable. Mean monthly factory capacity was 389,889 units, but the maximum reached 10 million units, a maximum-to-mean ratio of 25.6x. Mean order size was 2,508 units, while the maximum was 240,491 units, a 95.9x ratio. Mean lead time was 110 days, with a range from 1 to 503 days. Risk controls need to interpret orders relative to each supplier's normal scale, capacity and production pattern.

 

 

Why Hidden Subcontracting Requires a System-Based Benchmark

A supplier list is a commercial map, not necessarily a production map. When secondary facilities are hidden, the buyer loses the ability to connect purchasing decisions to actual capacity, working conditions and remediation responsibilities.

A system-based benchmark therefore needs several layers: production visibility, capacity credibility, order behavior, monitoring coverage and recurrence. Monitoring asks whether the workers and facilities involved sit inside an assessment and grievance system; recurrence asks whether a discrepancy disappears after remediation or returns with the next production peak.

This distinction is particularly important for intermittent subcontracting, making order-linked verification more informative than a static annual factory list.

 

The Scale of Unauthorized Subcontracting

When approved orders leave the approved production system

The 36% unauthorized-subcontracting rate in the core study is the report's clearest order-level benchmark. The denominator is not factories or audits; it is individual orders, which is closer to the commercial decisions buyers make every day.

Factory-level counts reveal a different picture. A majority of factories, 56.6%, never subcontracted in the observed data. Another 31.9% sometimes did so, while 11.5% always did. An intermittent supplier can move in and out of compliance as workload and circumstances change, making periodic inspection a weak standalone detector.

The order shares sharpen that conclusion. Never-subcontracting factories produced 35.6% of orders. Sometimes-subcontracting factories produced 57.9%, and always-subcontracting factories produced only 6.4%. In other words, the largest block of purchasing activity sat with factories whose behavior changed. The critical question is whether a current order creates conditions under which production is likely to move.

The dataset also shows that 98 of 226 factories represented by the behavior counts had ever used unauthorized subcontracting. That is 43.4% of factories. Yet those factories represented 64.3% of orders. Purchasing exposure was therefore more concentrated in factories with at least some subcontracting history than the factory count alone suggests.

Figure 1. Factory behavior is divided between never, sometimes and always subcontracting, with intermittent behavior forming a substantial middle group.

Figure 2. Sometimes-subcontracting factories account for the largest share of orders, making changing behavior central to detection.

 

Factory Behavior and the Intermittent-Risk Problem

The difference between persistent and intermittent behavior changes how a control system should work.

The sometimes group had a 51.7% unauthorized-subcontracting rate. That rate is neither low enough to dismiss nor stable enough to treat every order as subcontracted. It creates a strong case for order-level screening. Capacity utilization, order size, lead time, product complexity, historical behavior and unusual production spikes can be used to identify orders that deserve additional verification.

Factory scale adds another layer. The average factory in the study delivered 143.4 orders, but the standard deviation was 270.6 and the range ran from 1 to 2,772 orders. Factories handled an average of 2.8 product categories, with a range from 1 to 10, and served an average of 1.4 buyers, with a maximum of 11. Production networks differ materially in commercial breadth and order intensity.

The practical implication is that risk should be normalized to the supplier. A sudden increase from 20 to 60 orders may be more important for a small factory than a similar absolute increase for a very large manufacturer. Hidden subcontracting detection becomes stronger when it asks whether an order is unusual for that supplier rather than merely unusual in the entire industry.

 

 

Detecting Hidden Subcontracting Through Order Data

The ability to predict unauthorized subcontracting from operational data changes the role of compliance analytics. The reported out-of-sample accuracy of more than 80% at order level indicates that hidden production is not completely random. Supplier and order characteristics contain enough structure to distinguish higher-risk cases from lower-risk cases with useful accuracy. Supplier-level prediction of about 70% is weaker but still meaningful for prioritizing deeper review.

The distinction between order and supplier accuracy also reinforces the need for granular data. That is particularly important when 31.9% of factories sit in the sometimes-subcontracting group and those factories handle 57.9% of orders. A supplier can be low risk in one month and high risk in another because the commercial conditions have changed.

Figure 3. Reported predictive performance is stronger at the order level than at the supplier level, supporting granular screening.

 

Capacity Pressure, Order Volume and Production Movement

The commercial conditions behind subcontracting risk

Capacity is one of the most useful bridges between purchasing data and physical production. Mean monthly factory capacity in the study was 389,889 units, but the standard deviation was 864,023 units and the maximum was 10 million. The maximum was 25.6 times the mean.

Order size varied even more. The mean was 2,508 units, the standard deviation 5,828, the minimum 1 and the maximum 240,491. The maximum-to-mean ratio was 95.9x. Lead time averaged 110 days but ranged from 1 to 503 days, with a maximum-to-mean ratio of 4.6x. These figures describe an environment in which purchasing demands can differ by orders of magnitude.

Utilization also requires careful interpretation. Mean utilization in a non-idle month was 42.8%, but the reported range extended from 0.21% to 322%. Values above 100% indicate that nominal capacity can diverge sharply from observed production measures or that factories can stretch output through overtime, additional lines, outsourcing or other adjustments. The statistic is therefore less useful as a literal ceiling than as a signal that declared capacity must be reconciled with actual operating conditions.

A capacity credibility test should compare four elements at once: purchased volume, available lead time, declared capacity and observed output. Any discrepancy should be resolved with evidence.

Figure 4. The dataset contains extreme variation in capacity, order size and lead time, reinforcing the need for supplier-specific thresholds.

 

Supply-Chain Visibility: How Much Production Can Buyers Actually See?

Tier 1 visibility is an important starting point, but it captures only part of the production system. One global visibility indicator in the dataset places Tier 1 at roughly 20% of the production processes in a finished garment.

Hidden subcontracting turns this multi-tier complexity into an order-control problem. A buyer may have a strong relationship with the final assembler while knowing much less about temporary overflow units or specialized secondary facilities. Supplier disclosure therefore needs to move beyond company names toward facility identifiers and order allocation. A list that is correct once a year can still be wrong for a particular order next week.

Visibility improves in stages. Supplier name establishes the commercial counterparty. Factory address identifies an approved site. Capacity information tests whether that site can plausibly perform the work. Subcontractor declaration identifies secondary production. Order-to-facility mapping connects the purchase order to those facilities. Worker and facility verification then tests whether the map reflects reality. Each stage reduces uncertainty, but none should be confused with the next.

The strongest control is dynamic rather than static. Production sites should be linked to order IDs, process steps and dates so that changes can be reviewed before work moves. This approach also makes legitimate subcontracting easier to manage because the buyer can distinguish approved specialist processes from undisclosed overflow production.

 

 

Monitoring Coverage and the Audit Perimeter

Known factories versus the wider production network

Monitoring programs demonstrate the scale at which formal oversight can operate. Better Work reports coverage of more than 2,200 factories, more than 3.7 million workers, 240 brands and retailers and 11 countries, with a program history exceeding 20 years. These figures show that factory-level assessment can reach large production populations when facilities are known and participating.

Country coverage is substantial in several sourcing markets. Bangladesh records 501 Better Work factories and 1,323,334 workers. Cambodia records 797 factories and 761,915 workers. Indonesia records 206 factories and 469,014 workers, while Pakistan records 135 factories and 300,816 workers. Egypt, Ethiopia and Haiti add smaller but still meaningful monitored populations. The monitoring footprint is therefore large enough to shape industry practice.

The hidden-subcontracting challenge is that monitoring quality and monitoring perimeter are separate variables. A rigorous assessment at an approved factory says little about a workshop that never appears on the factory list. If production moves outside the perimeter, the buyer may lose access to assessment records, corrective-action plans, worker-management committees, grievance channels and documented employment information.

Figure 5. Better Work worker coverage is substantial across several major sourcing markets, but coverage depends on facilities being visible to the program.

 

Labour-Risk Context Behind Hidden Production

The labour significance of hidden subcontracting comes from the loss of verified conditions, not from an assumption that every subcontractor violates standards. Asia-Pacific garment, textile and footwear exports were valued at about $601 billion in 2019, representing 60% of global industry value in the cited dataset. The region employed about 65 million garment-sector workers, equivalent to 75% of the worldwide workforce in that measure. When production visibility fails in such a large employment system, even a small unmonitored share can involve many workers.

Working-time pressure is one channel. In Viet Nam, a high-hours group was associated with about 2,300 annual working hours, and many workers needed more than 50 overtime hours per month in the cited labour context. Indonesia's garment evidence indicated that nearly 60% of workers had excessive hours. These statistics are not measures of hidden subcontracting, but they show why lead-time and capacity pressure should be assessed alongside labour outcomes.

Gender and worker vulnerability add another dimension. Global garment manufacturing employment is reported as roughly 80% women, while the Asia-Pacific evidence shows women holding only 22.5% of management roles in 2018, up from 17.7% in 1991.

 

 

Workforce Scale and the Human Exposure Behind Supply Chains

The global textiles and garment workforce is measured at about 91 million workers in one 2019 estimate, including roughly 50 million women. A 2025 figure places textiles and clothing jobs at more than 90 million people.

Monitoring coverage should be interpreted against that denominator. Better Work's more than 3.7 million workers represent a substantial monitored population, yet it remains a fraction of the global sector. The gap does not imply that all other workers are unprotected; it simply shows that no single monitoring program covers the entire industry. Buyers therefore need their own facility mapping and due-diligence controls even when they participate in established programs.

Worker scale also changes the meaning of factory-level discrepancies. A production shift that appears small in commercial terms can move hundreds or thousands of workers outside the buyer's known network. Conversely, a large integrated facility can employ thousands of workers while maintaining strong internal controls. The number of workers is therefore an exposure measure, not a quality score.

 

Gender and Hidden Subcontracting Exposure

Why workforce composition belongs in the risk model

Garment production is unusually important to women's employment. The global dataset places women at about 80% of garment manufacturing employment. Cambodia reports women at roughly 75% to 80% of its garment, footwear and travel-goods workforce, Bangladesh reports about 55% to 60% in its RMG workforce, and Indonesia reports 58% of garment workers as women. Pakistan's Better Work factory population is markedly different, with women representing 12.5% of workers in the cited 2024 data.

Gender data therefore belong beside production data rather than in a separate social appendix. When a brand maps an order to a facility, it can also map the workforce reached by monitoring, grievance and remediation systems. That creates a clearer connection between sourcing decisions and social outcomes.

Figure 6. Women form a large share of garment workforces in several major production markets, although the composition varies substantially by country and program population.

 

Country-Level Unauthorized Subcontracting Signals

Large differences inside one study population

The core study reports substantial country variation in unauthorized-subcontracting rates. Viet Nam recorded 54.5%, Cambodia 52%, China 48%, Indonesia 16%, the 'Other' group 4.6% and Bangladesh 0.15%. The gap between Viet Nam and Bangladesh was 54.35 percentage points; Cambodia exceeded Indonesia by 36 points, and China exceeded Indonesia by 32 points. These are large differences within the study dataset.

The country values should not be treated as timeless national rankings. They reflect the suppliers, orders and period represented in the research. Country manufacturing systems evolve, buyers use different supplier portfolios, and enforcement or monitoring arrangements change.

Study composition also matters. China represented 59.7% of factories and 54.8% of orders. Bangladesh represented 10.6% of factories and 12.7% of orders. Indonesia represented 6.6% of factories but 13.1% of orders, a positive order-share minus factory-share gap of 6.5 percentage points. These differences affect how much each country contributes to the aggregate result.

The lesson is methodological: compare rates with denominators, sample shares and operational context. A country rate without study composition can be visually dramatic but analytically incomplete.

Figure 7. Unauthorized-subcontracting rates differ sharply across countries in the study; these values describe the study population, not permanent national rankings.

 

Bangladesh: Scale, Compliance and Subcontracting Exposure

Bangladesh illustrates why low observed unauthorized subcontracting in one dataset should not reduce the need for production visibility. The country's RMG sector employed about 4.2 million workers in 2017, with women representing roughly 60%. An earlier 2014 estimate placed employment at about 4 million and the female share between 55% and 60%. The sector was also described as indirectly supporting 40 million people, equivalent to about 25% of the population in that period.

Commercial scale is equally important. RMG exports reached about $24.5 billion in 2013-14 and generated more than 80% of export earnings in the cited historical data. A sector this important creates dense networks of factories, suppliers and specialized processes. Even when direct unauthorized subcontracting is low in a particular study sample, the economic incentive for production flexibility remains significant.

The human consequences of weak factory visibility are part of Bangladesh's industrial history. The dataset records 1,136 deaths at Rana Plaza and 112 deaths in the Tazreen fire. The broader lesson is that buyers need to know which facilities are producing their goods before safety and labour controls can function.

Current monitoring scale is substantial: Better Work reports 501 factories and 1,323,334 workers in Bangladesh. That creates a strong formal oversight base. The remaining control question is order linkage—whether purchased production can be reconciled to the monitored facilities rather than merely to a supplier name.

 

Cambodia: Factory Monitoring and Production Visibility

Cambodia combines a large export-oriented garment system with unusually extensive monitoring data. In 2024 the garment, footwear and travel-goods sector included 1,555 firms, up 30% from 2020, and employed more than 918,000 workers. By 2025 the reported firm count reached 1,810 and employment exceeded 1.11 million. Women represented around three quarters of the workforce, with a 2024 range of 75% to 80%.

The sector generated about $13.6 billion in GFT exports in 2024, with ready-made garments accounting for more than 70%. In 2025 GFT products represented 51.8% of total exports. The United States accounted for more than 38% of GFT export revenue and the European Union nearly 28%. This concentration connects factory visibility directly to major foreign-demand markets.

Monitoring coverage is broad. Better Work reports 797 factories and 761,915 workers in Cambodia, while 2024 data include 381 garment factory assessments and 75 travel-goods/bags assessments. Yet the hidden-production evidence shows why monitoring scope must remain dynamic: worker reports in 2015 identified 11 factories whose workers said their factory subcontracted out and 25 factories whose workers reported doing subcontracted work.

The country therefore illustrates both sides of the control problem. Strong monitoring infrastructure can generate detailed information about registered facilities, but order-level visibility is still necessary to identify production relationships between factories. A facility can be monitored and still participate in a production chain that is not fully visible to a buyer.

 

Indonesia: Production Scale and Supply-Chain Complexity

Indonesia's garment sector demonstrates how production scale and labour conditions intersect with visibility. The cited 2016 data place garment exports at $11.6 billion, equivalent to 6.6% of merchandise exports, while the sector contributed 1.4% of GDP. Employment exceeded 4 million workers, and women represented 58% of garment workers. More than 2,000 medium and large garment manufacturers were operating in the measured period.

The labour context adds pressure indicators that are relevant to capacity analysis. Monthly garment wages were about $154 in the cited 2016 dataset, nominal annual wage growth since 2012 was 8.8%, and the raw gender pay gap was 6.8%. Nearly 60% of workers were reported as working excessive hours.

Indonesia represented 6.6% of factories but 13.1% of orders in the core subcontracting study, giving it a 6.5 percentage-point positive gap between order share and factory share. Its unauthorized-subcontracting rate in that study was 16%. The combination shows why both prevalence and purchasing exposure matter.

Better Work coverage of 206 factories and 469,014 workers provides a formal monitoring base. The strongest buyer control is to connect that monitoring information to actual order allocation and to investigate production volumes that exceed the approved facility's plausible operating range.

 

Pakistan: Emerging Monitoring and Supplier Visibility

Pakistan's Better Work data provide a useful picture of an expanding formal monitoring system. In 2024 the program worked with 19 brands and retailers, 98 active factories and 219,327 workers. Women represented 12.5% of workers in that factory population. By 2025 the broader monitoring figures in the dataset list 135 factories and 300,816 workers.

Pakistan is not represented with a directly comparable unauthorized-subcontracting rate in the core country breakdown. That absence should be preserved rather than filled with an inferred national estimate. The available evidence supports a monitoring and workforce profile, not a prevalence claim.

For brands sourcing from the market, the practical focus should therefore remain facility-level: which site is approved, what volume can it produce, whether secondary facilities are declared, and whether shipment records reconcile with the production plan.

 

Wage Pressure and the Economics of Secondary Production

Cambodia's monthly minimum wage was $204 in 2024 and moved to $208 for 2025 in the cited data. Indonesia's historical garment wage benchmark was $154 per month in 2016. These values come from different years and institutional contexts and should not be treated as a direct cost ranking. Their role is to show that labour cost is one component of a much wider production equation.

Purchasing teams can make the equation more transparent by recording target price, order size, confirmed capacity, lead time, overtime assumptions and authorized external processes. When a supplier's commercial offer depends on capacity that is not visible in the approved factory, the issue can be addressed before production begins.

This approach also improves remediation. If hidden subcontracting is caused partly by unrealistic order changes or compressed lead times, simply warning the supplier may reproduce the same problem. Buyer purchasing practices and supplier production controls need to be examined together.

 

Global-Supply-Chain Exposure and Foreign Demand

Global supply chains connect production visibility to employment far beyond individual factories. One 2026 global statistic in the dataset places jobs linked to foreign demand at about 465 million. In South-East Asia, more than 75 million jobs were described as global-supply-chain related in 2023, with Viet Nam accounting for more than one quarter of those jobs.

Viet Nam's exposure is especially instructive. More than 35% of total employment was linked to global supply chains, manufacturing accounted for 49% of GSC employment, and textiles represented nearly one third of manufacturing GSC employment. More than 76% of Viet Nam's GSC jobs relied on six major foreign-demand markets. These figures demonstrate how buyer decisions in destination markets can influence production systems and employment at scale.

Hidden subcontracting can therefore be understood as a governance gap inside a much larger economic relationship. Orders move across borders through sophisticated commercial systems, but the last production allocation inside a sourcing country may still be difficult to observe. Digital purchasing records are highly detailed at the commercial level; production records need equivalent granularity at facility level.

 

Governance, Due Diligence and Buyer Responsibility

Moving from supplier codes to production verification

Governance frameworks establish the expectation that companies identify and address risks across garment and footwear supply chains. The dataset notes that 50 governments supported the OECD Garment Guidance and that the guidance itself runs to 192 pages. The scale of that framework reflects the reality that responsible sourcing involves more than a supplier code or a single social audit.

For hidden subcontracting, governance should translate into specific operational controls. Contracts can require authorization before production moves. Supplier onboarding can identify all owned and regularly used facilities. Capacity checks can compare order volume with production resources. Purchase orders can carry facility identifiers. Changes can trigger approval workflows. Verification can then test whether the declared production path matches records and physical activity.

Remediation should also distinguish causes. Deliberate concealment, emergency overflow, buyer-driven lead-time compression and inaccurate capacity data require different responses. A mature system corrects the production map and the commercial conditions that created the discrepancy, not just the paperwork.

 

 

Traceability and Circularity as Visibility Infrastructure

Traceability systems are often built around materials, environmental claims and circularity, but the same data architecture can strengthen labour visibility. The dataset notes that textile-to-textile recycling represents about 1% of textile production, highlighting how much work remains in material circularity. As brands invest in product passports and material traceability, facility and process identifiers can be designed into the same information flow.

A useful production record links product ID, purchase-order ID, supplier ID, facility ID, process step and shipment. If a process moves to another facility, the record changes before production proceeds. This creates an auditable chain that can support both environmental and social due diligence without requiring separate data silos.

Traceability is not automatically verification. A digital record can reproduce inaccurate supplier declarations. The control becomes stronger when facility IDs are connected to capacity records, assessment status, worker coverage and transaction evidence. Data integrity and physical verification must therefore develop together.

The benefit is operational as well as ethical. A buyer that knows where each process occurs can manage quality, lead time, recalls and compliance more efficiently. Hidden subcontracting becomes harder because an unexplained production step creates a visible break in the order record.

 

Building the Hidden Subcontracting Risk Index

The Hidden Subcontracting Risk Index converts the report into eight weighted control pillars. Unauthorized-subcontracting history receives 18%, the largest weight, because observed behavior is the most direct indicator of recurrence. Order and capacity mismatch receives 16%, reflecting the importance of whether the approved facility can plausibly complete the work. Facility and subcontractor visibility receives 15%, ensuring that supplier disclosure is translated into a current production map.

Monitoring coverage receives 13%, because known facilities need credible assessment and worker-protection systems. Labour-risk context receives 12% to capture working time, wages, safety and representation without assuming that visibility alone proves a labour violation. Purchasing and lead-time pressure receives 10%, recognizing that buyer practices can create production stress. Traceability and data integrity receive 9%, while governance and remediation receive 7%.

Scores should remain disaggregated. A supplier with excellent audit results but weak capacity reconciliation should not receive an unqualified strong score. Likewise, a supplier with a past subcontracting incident can improve if it demonstrates transparent authorization, reliable facility mapping and sustained remediation. The index is intended to expose where evidence is weak rather than hide weaknesses inside one average.

A practical interpretation uses five bands: 0–39 indicates limited visibility, 40–59 basic controls, 60–74 developing verification, 75–89 strong production visibility and 90–100 advanced order-to-facility traceability. These bands are management categories for the framework, not external industry standards.

Figure 8. The proposed index gives the greatest weight to observed subcontracting history, capacity credibility and facility visibility.

 

Hidden Subcontracting Market Challenges

The first challenge is incomplete factory information. Supplier databases often identify legal entities or primary sites but do not capture every production unit used during peak periods. The second challenge is timing. Intermittent subcontracting can begin after an audit and end before the next one. The third is commercial complexity: orders change, delivery dates move and suppliers rebalance production across lines and facilities.

The fourth challenge is audit perimeter. Monitoring programs can be rigorous inside known factories while having no visibility into an undisclosed workshop. The fifth is data fragmentation. Purchasing systems, compliance platforms, shipment records and factory capacity data may sit in separate systems, making it difficult to identify contradictions quickly.

The sixth challenge is incentives. A supplier that believes disclosure will automatically cause order cancellation may hide a legitimate capacity problem until it becomes a compliance issue. Controlled authorization and early escalation can create better incentives than a policy that exists only as a prohibition.

 

90-Day Hidden Subcontracting Verification Plan

Days 1 to 30 should establish the production baseline. Link those records to current purchase orders rather than storing them as a separate compliance list.

Days 31 to 60 should test capacity credibility. Compare purchase-order volume and lead time with production schedules, worker attendance, line allocation, input consumption, output and shipment timing. Flag unusual changes relative to the supplier's own history. A large order is not automatically risky; an order that cannot be reconciled to available production resources is the stronger signal.

Days 61 to 90 should verify the production pathway. Use targeted site checks, document reconciliation, worker interviews where appropriate, subcontractor validation and shipment review. When a discrepancy is found, determine whether it reflects authorized secondary production, inaccurate records or undisclosed subcontracting. Record remediation and check whether the same pattern recurs.

The plan should end with a verified production map and a set of supplier-specific risk thresholds. Those outputs are more useful than a one-time audit score because they can be applied to new orders as commercial conditions change.

 

 

Metrics Brands and Retailers Should Track

Production metrics should include order count, units, lead time, declared capacity, utilization, line allocation, worker hours, output and shipment volume. These measures establish whether the approved facility can plausibly complete the work. They become most useful when stored historically so that unusual changes can be identified relative to the supplier's normal pattern.

Visibility metrics should include approved facilities, declared subcontractors, facility IDs, authorized processes, production-site changes and the share of orders mapped to verified sites. Monitoring metrics should include assessment coverage, worker coverage, corrective actions, repeat findings and risk-triggered verification. The combination separates knowing a facility from knowing what it produced.

Outcome metrics should include unauthorized-subcontracting cases, recurrence, remediation time, worker complaints, unexplained capacity mismatches and traceability completeness. A low case count is positive only when detection coverage is credible. If few orders are reconciled to production evidence, a low incident rate may simply reflect low visibility.

Management dashboards should therefore show both performance and evidence coverage. For example, the share of orders with verified facility mapping can sit beside the share of those orders with discrepancies. This prevents the organization from interpreting missing data as good performance.

 

How Hidden Subcontracting Risk Changes by Business Model

Direct sourcing gives buyers a shorter commercial chain and can make capacity verification easier, but it also places more responsibility on the buyer to understand the factory's real production system. Agents and intermediaries add a coordination layer that can be useful for supplier management while increasing the distance between the buyer's purchase order and the facility allocation decision.

Large integrated manufacturers may have stronger internal systems, yet multiple campuses, buildings and legal entities can still make order allocation complex. Small workshops may offer flexibility and specialist skills but often have weaker documentation. Marketplace and multi-brand sourcing models can involve large supplier populations, making standardized facility identifiers and automated exception detection particularly valuable.

The same control framework can operate across these models if responsibilities are explicit. Someone must own capacity confirmation, subcontractor authorization, facility data, order allocation and remediation. When those responsibilities are split across purchasing, agents and compliance teams without a common record, hidden production becomes easier.

Business model therefore changes the route to visibility rather than the underlying objective. Every order should still be traceable to the facilities and processes that completed it.

 

The Hidden Subcontracting Report FAQ

What is hidden subcontracting?

Hidden subcontracting is production transferred to another facility without the disclosure or authorization required by the buyer or sourcing arrangement. It is different from authorized subcontracting, where secondary production is known, approved and subject to relevant controls.

How common was unauthorized subcontracting in the core study?

The study dataset covered more than 30,000 orders and reported that 36% were subcontracted without authorization. This is a study-specific benchmark rather than a universal global prevalence rate.

Why are sometimes-subcontracting factories important?

They represented 31.9% of factories but 57.9% of orders and had a 51.7% unauthorized-subcontracting rate. Their changing behavior makes static supplier labels less useful than order-level screening.

Can analytics detect hidden subcontracting?

The cited study reported more than 80% out-of-sample accuracy at order level and about 70% at supplier level. Analytics can prioritize investigation, but a prediction is not proof that production moved.

Is all subcontracting a labour violation?

No. Subcontracting can be a legitimate production arrangement. The control issue is whether it is authorized and visible; labour conditions at any facility should then be assessed separately.

Can audits solve the problem?

Audits are valuable inside known facilities. They cannot directly assess a site that is absent from the production map, so facility discovery and order-to-site reconciliation are necessary complements.

What should buyers compare with capacity?

Order volume, lead time, workforce, line availability, machinery, utilization, output and shipment records should be reconciled. Unexplained gaps are more informative than a single capacity number.

Which country has the highest risk?

The study reports different country rates, including 54.5% for Viet Nam, 52% for Cambodia and 48% for China, but these figures describe the study sample and period. They should not be converted into permanent national rankings.

What is the strongest practical control?

The strongest control is an order-to-facility record supported by capacity and production evidence. It connects the commercial purchase order to the physical sites that actually performed the work.

What should remediation achieve?

Remediation should correct the production map, address the reason production moved, verify any secondary facility and adjust purchasing or supplier controls when commercial pressure contributed to the problem.

Final Takeaway

Hidden subcontracting is best understood as a mismatch between commercial visibility and physical production. The core evidence is substantial: more than 30,000 orders, 36% subcontracted without authorization, and 43.4% of represented factories using unauthorized subcontracting at least once. Factories with any such history accounted for 64.3% of orders, showing that purchasing exposure can be greater than the supplier count suggests.

The most important pattern is intermittent behavior. Sometimes-subcontracting factories represented 31.9% of factories but 57.9% of orders and had a 51.7% unauthorized-subcontracting rate. Always-subcontracting factories had a 100% rate but represented only 6.4% of orders. A control system that looks only for permanently problematic suppliers will therefore miss much of the operational risk.

Data can help close the gap. More than 80% order-level prediction accuracy and about 70% supplier-level accuracy show that hidden production can leave measurable signals. Capacity and order statistics provide additional context: mean monthly capacity of 389,889 units sits inside a range extending to 10 million, while order size ranges from 1 to 240,491 units and lead time from 1 to 503 days. Supplier-specific baselines are essential.

The wider industry context makes visibility consequential. More than 90 million people work in textiles and clothing, women form a large share of garment employment, and major monitoring programs cover millions of workers. Those systems protect only what they can see.