Dual Sourcing in China: How to Build a Second Supplier That Is Actually Ready

Dual sourcing exists only when a second supplier can produce approved-equivalent output, pass the required evidence gates and accept a defined volume within the maximum interruption the business can tolerate. It must also remain available during the failure the business is trying to survive.

That last condition matters. Two Chinese suppliers may use the same subcontractor, mould shop, material producer, industrial park, port or parent company. If both are exposed to the same failure, the buyer has duplicated supplier administration without creating a usable production alternative.

The practical method is to define the disruption, map shared dependencies, qualify the second supplier against the same product truth, fund enough live readiness and agree an activation playbook before the primary supplier fails.

Two supplier names do not automatically create redundancy

The OECD's 2025 supply-chain resilience review treats redundancy, including backup suppliers, as one component of resilience alongside flexibility and an organisation's ability to respond. It also warns that dual sourcing and diversification can add complexity and cost and may not always improve resilience.

For an Australian importer, that turns the decision into a test rather than a slogan:

1. What failure should the second source absorb?
2. Would source B remain available if that failure occurred?
3. Can source B make the approved product now, not after an open-ended redevelopment project?
4. Can the business afford the one-time and recurring readiness cost?
5. What evidence authorises a volume transfer?

If those questions do not have recorded answers, the business has a candidate supplier, not a ready second source.

Start with the failure mode

The Australian Government's supply-chain resilience work starts with vulnerability, criticality and the residual ability to prevent, absorb, substitute, adapt or transform. Apply that logic to one product or component before asking for more quotations.

Name the event and its commercial consequence. A supplier-specific failure might be insolvency, chronic quality escape, loss of a key process, refusal to honour specifications or an irreparable relationship breakdown. A factory-level event might be fire, flood, utility interruption or loss of a licensed process. Broader events can affect a province, a port, a material network, a regulatory pathway or the entire China-to-Australia lane.

Then define the maximum tolerable interruption. A seasonal product with eight weeks of stock, an essential replacement part with two days of cover and a custom product that needs a new mould are not the same sourcing problem. The required response time determines how much of source B's capability must already exist.

The joint Australia-UK resilience initiative frames resilience as diversity of supply rather than total self-reliance and calls for responses tailored to the risk. A China-plus-China strategy can reduce exposure to one supplier or site. It does not automatically remove country, currency, customs, trade-remedy, port or upstream concentration. If the priority is a China-wide disruption, a second factory in the same country may be only one layer of the response.

Test whether the sources are genuinely independent

Map dependencies at the level that can stop the product. Do not stop at the selling entity.

This is not a demand for total independence. Some common dependencies may be acceptable or hard to replace. The purpose is to know which risk the second source does and does not cover, then price that residual exposure honestly.

Give both suppliers the same product truth

A second supplier cannot reproduce a product that exists mainly in source A's email history or technician memory. Establish one controlled buyer-owned definition before parallel qualification.

The definition should include the current drawing or specification revision, bill of materials, approved material grades and sources where relevant, tolerances, workmanship criteria, functional and safety tests, packaging and labelling, master data and change-approval rules. Use the existing guide to create a factory-ready product specification.

Where appearance, finish or assembly judgement matters, control an approved physical reference using the golden-sample method. Give each source an identified, equivalent reference and record custody. A sample at source A is not useful to source B during a disruption.

Do not hide source-specific process differences. Two factories can use different equipment or sequences and still meet the same output requirements, but any alternative process that affects safety, compliance, durability or appearance needs evidence and approval. The objective is equivalent controlled output, not forced imitation of every production step.

Qualify source B as if it were about to ship

The qualification standard should be driven by product and exposure, not by the fact that source B is called a backup. A dormant supplier with a passed introductory call is not a production option.

Australian Government supplier guidance recommends assessing reliability, quality and value rather than price alone and recording material commercial terms. Run the same commercial and technical hard gates used for the primary source. The supplier evaluation scorecard can record identity, evidence confidence, capability, quality systems, capacity, communication and commercial fit without allowing a high soft score to override a failed hard gate.

For consumer products, ACCC product-safety guidance tells Australian suppliers to check regular factory audits, raw-material testing, followed assembly processes, quality checks, qualified pre-shipment inspection and current staff training. The exact evidence set still depends on the product, applicable rules and risk. It can include:

• legal-entity and bank/payee reconciliation;
• capacity and process evidence for the exact product family;
• a risk-based audit or equivalent capability assessment;
• approved prototype, pre-production or production-reference samples as applicable;
• product-specific testing and compliance evidence;
• packaging and labelling approval;
• a pilot or controlled production lot;
• pre-shipment inspection against the same quality-control plan; and
• closed corrective actions before routine allocation.

Passing once is not permanent readiness. Record what changes trigger requalification: material, tooling, site, subcontractor, process, test method, key personnel, long inactivity or a serious quality event.

Model the cost before calling it resilience

Dual sourcing consumes cash and management capacity. The benefit is avoided or reduced disruption, but the readiness cost exists even when nothing goes wrong.

Separate one-time costs from recurring costs and activation costs. One-time lines may include discovery, audit, engineering transfer, samples, testing, pilot production and duplicate tooling. Recurring lines can include a unit-price premium from split volume, repeated testing, extra inspections, supplier management and higher logistics complexity. Activation may require expedited materials, revalidation, overtime, premium freight or a temporary capacity reservation.

The following example is hypothetical and demonstrates the ledger only.

Build low, base and high disruption scenarios. First estimate the contribution, recovery and expediting loss if source A is unavailable without a second source. Then estimate the residual loss after source B contributes only its evidenced ramp and surge volume. The difference is the avoidable disruption exposure in that scenario. Compare that range with readiness and activation costs over the same decision horizon; do not collapse uncertain inputs into one precise probability.

Dual sourcing is not automatically the cheapest control. Safety stock, a capacity reservation, repairable tooling, a tested substitute material or improved recovery terms may address the main failure at lower complexity. For a critical or fast-growing product, the second source may still be justified even when its normal-year accounting cost is visible.

Allocate volume to preserve readiness

There is no universal best split. An 80/20 or 70/30 rule may sound decisive, but the useful allocation depends on MOQs, process stability, supplier economics, demand, tooling, lead time and the volume required to keep source B's people and process current.

Set a minimum readiness workload instead. Ask what source B must produce, how often and with which test and inspection evidence to demonstrate that the line, tooling, materials and knowledge remain usable. Set that cadence from product lead time, process stability, change rate, MOQs and the supplier's evidenced restart needs. Intermittent production may create repeated setup variation or make the account commercially unattractive.

Then set an ordinary allocation band and a surge limit. The ordinary band keeps the relationship and process alive. The surge limit is the volume source B has evidenced it can accept within a stated time without unauthorised subcontracting, silent material changes or quality-control compression.

Keep allocation decisions separate from punishment. Moving volume in response to quality or delivery performance can be appropriate, but unexplained swings make capacity commitments and supplier economics harder to assess. Use recorded measures, notice periods and escalation rules.

Make tooling and data portable enough to activate

Custom tooling can make the apparent second source unusable. Record who owns each mould, die, jig, fixture, gauge, program and test asset; where it sits; whether it can be transferred; who maintains it; and what happens if the commercial relationship ends. The China tooling ownership guide explains the register and access controls to establish before funding an asset.

Portability does not always mean moving one mould between factories. Frequent transfers may be impractical or risky. The solution may be duplicate tooling, modular inserts, separately controlled soft tooling, spare critical components or a documented path to rebuild. Whatever the method, validate source B's actual configuration before treating it as ready.

Keep the current controlled files outside one supplier's systems. Preserve drawings, BOMs, approved deviations, software, test fixtures, packaging artwork, inspection criteria and validation results with access controls and revision history. A legal right to data is not the same as a technically complete transfer pack.

Define the activation playbook before disruption

Write the decision while the primary source is still operating. The activation record should name:

• the trigger, such as a defined capacity loss, delivery breach, safety concern or unresolved major quality failure;
• the decision owner and required technical/compliance sign-off;
• the evidence used to confirm source B's current capacity, materials and approved process;
• any sample, test, pilot, inspection or regulatory step that must be repeated;
• the initial volume and maximum ramp rate;
• inventory, freight and customer commitments during the transition;
• controls against unauthorised subcontracting or specification changes; and
• conditions for returning to ordinary allocation.

Test the playbook on a schedule derived from product criticality, lead time and change rate, and after any material change. Confirm contacts, lead times, tooling location, material availability, file revisions and realistic transport paths. A tabletop check will not prove production performance, but it can reveal an expired quote, missing drawing, unavailable mould or shared dependency before an actual disruption.

Parallel qualification and live allocations also affect working capital. Use the China sourcing cash-flow timeline to show deposits, tooling, tests, overlapping inventory and freight commitments rather than treating readiness as a unit-price-only decision.

What to do next

Choose one product where a supplier interruption would materially affect revenue, customers or compliance. Name the failure you want to absorb and the maximum tolerable interruption. Then complete a dependency map for the primary and proposed second source.

Do not call the strategy complete until source B has the same controlled product definition, has passed the required evidence gates, has produced acceptable output and has a commercially credible readiness workload. Record the year-one and recurring cost, the ordinary allocation band, the surge limit and the activation owner.

Dual sourcing is valuable when it creates an option the business can actually exercise. The goal is not two supplier names. It is two controlled, understood production paths—and an honest record of the risks they still share.

Sources

• OECD Supply Chain Resilience Review 2025
• Australian Government Office of Supply Chain Resilience
• Australia-UK Joint Supply Chain Resilience Initiative: introduction module
• business.gov.au: Suppliers
• ACCC Product Safety: How to source and test products

Sources retrieved 15 August 2026 Australia/Sydney. This article provides general operational information, not legal, customs, product-safety or financial advice.