Dual sourcing is presented as resilience, and it is also duplication: two quality systems, two certification sets and two sets of tooling. The decision is worth making deliberately, with the cost of duplication on the table beside the risk being mitigated.

This article covers the motives, the components that are hardest to move, and a qualification sequence that avoids splitting volume before the second site is proven.

When is dual sourcing worth the cost?

When the risk outweighs duplication.

Dual sourcing pays when tariff exposure, single-site concentration or customer requirements justify the cost of a second tooling set, certification set and audit.

Three motives dominate: tariff and trade policy exposure, resilience against a single-site disruption, and customer requirements that specify a second region. Each has a different cost profile, and the calculation should be done per product line rather than per company.

Where the motive is a customer requirement, the cost is usually passed through in the contract. Where the motive is internal risk management, the duplication cost has to be recovered from the programme’s margin.

The ecosystem gap: components, tooling and PCBA

A charger is more than an assembly: it depends on a component ecosystem for switching devices, controllers, magnetics, connectors and PCB fabrication. A second site in a different region inherits a different ecosystem, which affects lead time, cost and the qualification work.

Tooling is the second constraint. Housings, moulds and test fixtures are often built around one site’s process; duplicating them is a capital cost that only makes sense above a volume threshold.

Qualifying a second site without splitting volume

The sequence that works is: qualify the second site on a defined platform, run a pilot batch against the same golden sample, compare measured performance against the first site’s records, and only then allocate a share of production.

Managing quality drift across two plants

Two sites produce two distributions of results, even with the same design, because the processes differ. The controls that keep them aligned are a shared golden sample, a common test specification, and a comparison of batch records rather than of certificates alone.

Where the two sites are in different countries, drift also appears in component sourcing, which is why the BOM should specify acceptable alternatives rather than a single supplier part for each position.

Cost of duplication: tooling, certification and audits

Duplication costs appear in four places: tooling and fixtures, certification samples and test reports, audit and qualification effort, and inventory because two sites require two safety stocks. Adding those lines to the landed cost model shows the true premium of dual sourcing.

Efficiency and compliance documentation has to be re-issued per site where the market requires it (Regulation (EU) 2019/1782 and 10 CFR Part 430), and EMC evidence follows the destination market’s framework (47 CFR Part 15).

What a second site does not fix

Dual sourcing does not remove exposure to component shortages that affect the whole industry, and it does not reduce the certification burden for a new market. It addresses site-specific risk: a single factory, a single port, or a single regulatory regime.

Framing the decision that way prevents the common outcome where a company pays for a second site and still has a single point of failure in a critical component.

Documentation to keep for both sites

For each site, keep the quality system records, the golden sample, the batch test reports, the certification set for each market and the approved BOM with alternatives. Where the two sets differ, the difference should be documented rather than discovered.

WECENT runs its own manufacturing and quality system in Dongguan, with ISO9001 certification, multi-gate QC including 100% functional testing, hi-pot pressure testing and load aging, and per-batch inspection reports from the quality control gates. Programme documentation is available to buyers under NDA through the OEM/ODM programme, and the platform range is listed in the product catalogue.

A further consideration is the platform choice: products built on a standard platform with limited customisation qualify faster at a second site than heavily customised assemblies, because the tooling and test fixtures are simpler to duplicate. Where a programme needs both, sequencing the standard platform first gives the second site a gentler learning curve.

Documentation discipline across sites follows the same principle. A shared test specification, a common golden sample and per-batch records allow a buyer to compare two sites on evidence rather than on assurance (ISTA covers transit performance for packaging, and 10 CFR Part 430 sets the adapter efficiency framework in the US market).

A third consideration is certification timing. A second site needs its own samples and reports for each market, and those are produced after the pilot batch rather than in parallel with it. Building four to six weeks into the programme for that step keeps the commercial launch date realistic, particularly where the destination market requires a notified-body route.

A practical way to test the decision is to model a single-site disruption: how many weeks of supply would be lost, and what that would cost in missed retail windows. If the answer is smaller than the duplication premium, the programme is better served by a robust single site with inventory than by two under-loaded ones.