Complete Integration: Why Outsourcing Box-Build Is Safer Than Managing Multiple Subcontractors

The decision about a product’s integration model is usually made earlier than the R&D team assumes — and it affects the entire project life cycle: the ramp-up schedule, the total cost, and who bears responsibility for non-conformities detected after subsystems are assembled. Two models stand opposed: the distributed model (several specialized subcontractors coordinated by the OEM) and the consolidated model (complete integration at a single EMS partner, i.e. box-build).

The following study compares both models from an engineering perspective — hidden cost, change management, traceability, test strategy and IP protection — and indicates how to assess a partner’s real competences rather than accept its declarations.

1. System integration (Box-Build) versus the PCBA boundary

PCBA assembly covers the manufacture and test of a populated printed circuit board. Box-build (system integration — assembly and integration of the finished device) is the higher level: assembling the board and the remaining subsystems into a complete product — mechanical assembly, cable integration, configuration and software loading, marking, serialization, packaging, and in some cases logistics.

A condition for successful integration is the unambiguous demarcation of responsibility and the specification of the links and dependencies (interfaces) between subsystems — mechanical, electrical and software. The absence of a precise description of these links is the most common cause of non-conformities that surface only after integration.

It is also important that some of the decisions determining integration success are made at the R&D stage, not on the assembly line. This concerns testability in particular: it is in the design phase that it is decided whether the product architecture will allow an automatic functional test (FCT) after integration and access to the nodes required for the in-circuit test (ICT). Combining design for manufacturing (DfM) and design for test (DfT) early makes it possible to correct the design before changes begin to affect the ramp-up schedule. The consequences can be severe: if the product is coated with a conformal coating, access to the test points is blocked once the coating is applied — performing a functional test at a later stage becomes considerably harder or impossible. The “test–coating” sequence and the placement and accessibility of test points must therefore be resolved as early as the DfT stage.

In this area the scope of the EMS partner’s competences is decisive. ASSEL, for example, combines production preparation with the development of functional-tester software (test hardware and software) and in-circuit tests (ICT). An important caveat: this concerns tester software, not the product’s own software — the product’s firmware and applications remain with its creator (the OEM). Dedicated test and production tooling (and not product components) is produced using in-house CNC machining and 3D printing. This set of competences allows testability and manufacturability questions to be resolved in parallel with product development, rather than after it is closed. More on the scope of engineering support: engineering and DfM.

2. Risk analysis: distributed model versus consolidation

The distributed model is sometimes perceived as cheaper and more flexible. Risk analysis, however, points to hidden costs that surface only during integration and operation.

The junction of technological and competency interfaces. Each handover between subcontractors creates a boundary at which one party’s responsibility ends and another’s begins. Non-conformities arising at this boundary — for example defects at mechanical–electrical interface junctions — have no unambiguous owner. This results in a blurring of responsibility for system non-conformities: diagnosing such cases is longer and costlier than removing the fault itself. A particular case is mechanical damage where the stage of occurrence cannot be unambiguously established — whether it occurred after PCBA assembly, in transport between parties, or during mechanical integration. In the distributed model, assigning such a non-conformity to a specific link can be impossible.

Engineering change management. Introducing a design change in the distributed model requires it to be agreed and implemented in parallel at several parties, in different quality systems and schedules. The risk of documentation-version divergence and of delays due to desynchronization increases.

Coordination cost. Overseeing many suppliers — schedules, material flow, reconciling quality systems — burdens the OEM’s resources and is rarely captured in a comparative calculation. Added to this are the costs of the flow of semi-finished goods between parties: packaging, transport and warehousing (e.g. packing and shipping PCBAs to a mechanical-assembly supplier), and with them the cost of the packaging itself — significant in view of the new PPWR directive on packaging and packaging waste — as well as environmental costs and the disposal of waste arising in sub-processes.

Propagation of schedule risk. The probability of a slip grows with the number of independent links in the chain.

The box-build model consolidates these costs within a single point of responsibility: one party is responsible for the complete, working product, in one quality system. The table below organizes the differences:

3. Managing complexity and the supply chain

Managing complexity is a critical, not a peripheral, factor in box-build projects — which is why it warrants separate treatment. A complete product combines tens or hundreds of material items with different availability cycles, sourced from many suppliers. Without an appropriately structured process, this complexity translates directly into schedule and quality risk.

A well-designed process comprises several layers. Component-availability analysis conducted as early as the estimation and production-preparation stage makes it possible to detect shortage and EOL risk before it affects the ramp-up. Selection and oversight of suppliers of dedicated elements — in box-build the key suppliers become those of dedicated elements (mechanics, enclosures, special components), and their selection usually occurs already at the design or pre-production stage. The division of roles varies here: in many projects the client designates the suppliers and the contract manufacturer executes purchasing; ASSEL, however, also offers supplier selection for mechanical projects. For suppliers selected on its own side, ASSEL applies audits, assessment based on delivery history, and AQL levels set on the basis of quality history. In the consolidated model, one party runs and reconciles this rigour; in the distributed model coordination rests with the OEM.

The essence of the advantage is therefore not the mere number of services “under one roof”, but the continuity and coherence of the process spanning the supply chain, production preparation, integration and test. It is this continuity — not a single operation — that determines the product’s repeatability and predictability. Supply-chain discipline is described more fully in the study on supply-chain management.

4. Continuity of traceability — the role of the MES

In the distributed model, traceability is fragmentary: the origin data of components, sub-assemblies and integration arise in separate systems, and merging them in the event of a non-conformity can be time-consuming or unfeasible.

Integration at a single partner enables one continuous traceability chain — from component, through board and subsystems, to the complete, serialized product — maintained in one MES. Reconstructing the history of a specific unit then reduces to a query to the system.

The MES here performs a supervisory function, not merely a recording one. In ASSEL’s case, if a product has not passed a required operation, inspection or test, the system blocks its further flow; a unit that failed to meet a criterion is not directed to the next stage or to shipping. This ensures the detection of non-conformities within the process and a reconstructable history of every product. The scope of serialization and configuration is set per project, because requirements differ between medical, rail and industrial products. Structured, coherent traceability is also a condition for meeting the forthcoming requirements of the Digital Product Passport.

5. Test strategy: from component to system-level test (FCT / Run-in)

The test strategy in box-build comprises several levels, and its effectiveness depends on decisions taken at the R&D stage (DfT).

Subsystem-level tests — process control (SPI, 3D AOI), X-ray inspection, in-circuit test (ICT) — verify the correctness of the board and its connections. They do not, however, detect non-conformities that surface only at subsystem junctions, after integration. This category includes, among others:

• electromagnetic (EMI) phenomena appearing in the target enclosure and with the target cabling;
• thermal phenomena — limited heat dissipation in an installation that differs from the subsystem’s test-bench conditions;
• mechanical phenomena — resonances and stresses arising after mounting in the structure, and defects at mechanical–electrical interface junctions.

Detecting these classes of defect requires a functional test at complete-product level (FCT) and, in some applications, a burn-in / continuous-operation test (Run-in). This is why the system test is qualitatively different from the sum of subsystem tests.

The scope of tests should be verified with the specific partner. In ASSEL’s case, functional tests at product level are performed, and their tooling is developed internally (using CNC machining and 3D printing). Environmental stress screening (ESS), by contrast, is not currently performed, nor are qualification environmental tests (climatic chamber, vibration, salt fog), which remain with the OEM or accredited laboratories. This demarcation is presented explicitly in the competency matrix (section 7), because for the R&D team the transparency of the test scope is a qualifying criterion for a partner.

6. Managing intellectual property in regulated projects

Consolidating integration reduces the risk surface for intellectual property and information security — and this applies not only to dual-use projects, but to all regulated and sensitive projects (including medical devices per ISO 13485, and industrial and infrastructure applications).

In the distributed model, each subcontractor is a separate point of access to the documentation, hence a separate risk vector. The consolidated model limits access to one organization, one confidentiality agreement and one access-control regime.

In practice, ASSEL manages this area through access control to IT systems and to production premises, NDAs and data-protection procedures, run in line with ISO 27001 information-security-management standards. It should be noted that production is conducted essentially in a single hall — segmentation is based on access control (IT and physical), not on physically separated lines; for projects with the highest separation requirements this is a parameter to be settled at the specification stage. In projects intended for military or dual-use purposes, national special-trade rules (concessions, permits for trade in strategic goods) and EU export control (Regulation (EU) 2021/821) additionally apply — regardless of whether a given manufacturer currently runs such projects; the contract manufacturer then acts as an executor within the client’s regulated supply chain. The shared-responsibility model is described more fully in the study on data security in contract manufacturing (OEM–EMS).

7. Competency matrix: internal versus external

For the R&D team, an EMS partner’s transparency about its real competences is an elimination criterion — more important than the declarative breadth of the offer. The matrix below presents the demarcation of competences using ASSEL as an example: those performed internally and those remaining outside the plant (outsourced, limited or not offered). Disclosing the boundaries of scope is not a weakness — it is information necessary to assess a partner’s fit to the program’s requirements. This listing is a selection — the full list of technologies is on the PCB assembly / technology page.

The matrix in this form allows the R&D team to quickly assess which elements of the program will be covered internally and which require planning on the OEM’s side or by external parties — and it is the basis for preparing the interface specification and the validation plan.

8. Model selection criteria: optimizing TCO

Consolidating integration is not a universal solution. It optimizes total cost of ownership (TCO) under specific conditions:

• the product has high complexity and subsystem integration is non-trivial;
• traceability and system-level testing are required (including medical, rail, industrial, dual-use);
• the project is IP-sensitive, which justifies limiting the number of parties;
• the priority is a single point of responsibility and shortening the engineering-change chain;
• the OEM seeks to reduce its own coordination cost.

The distributed model remains rational for products of low complexity, very high and uniform volumes, and where the individual elements are already optimized at proven, specialized suppliers.

Assessing a box-build partner should rest on verifiable criteria: the real scope of internal integration (the competency matrix), the ability to perform functional and ICT tests at product level, continuity of traceability in one MES, supplier qualification, and an unambiguous IP boundary. A transparent demarcation of scope — including indicating the competences that remain outside the plant — distinguishes genuine system integration from coordination conducted under another name.

9. Frequently Asked Questions (FAQ)

Sources

1. IPC/WHMA — IPC/WHMA-A-620, Requirements and Acceptance for Cable and Wire Harness Assemblies. https://www.ipc.org/ipc-whma-a-620
2. IPC — J-STD-001, Requirements for Soldered Electrical and Electronic Assemblies (context of ICT and process criteria). https://www.ipc.org/j-std-001
3. Regulation (EU) 2021/821 — control of exports of dual-use items. https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:32021R0821
4. European Commission — Ecodesign for Sustainable Products Regulation (ESPR) / Digital Product Passport (traceability context). https://green-forum.ec.europa.eu/implementing-ecodesign-sustainable-products-regulation_en
5. Mordor Intelligence — Europe Electronic Manufacturing Services Market (system-integration context). https://www.mordorintelligence.com/industry-reports/europe-electronic-manufacturing-services-market
6. ASSEL — final assembly / system integration. https://asselems.com/en/pcb-assembly
7. ASSEL — data security in contract manufacturing (OEM–EMS shared-responsibility model). https://asselems.com/en/data-security-contract-manufacturing-oem-ems