Start with the production decision

Choose the network you need, then name the strategy

A sourcing review often begins with a map and ends with a country recommendation. That sequence skips the hard work. An automotive connector, a populated circuit board, and a finished consumer device can come from the same country while having different qualification requirements, transport economics, and failure consequences. The useful unit of analysis is a product family serving a defined customer market through a specific production and logistics network. Country selection is one input to that design.

Start by stating the business problem in measurable terms. Perhaps one site supplies every approved housing, a customer needs replenishment within a week, or an importer cannot confidently establish origin for a new assembly route. Each problem suggests different alternatives. Adding a factory may address the first; relocating final configuration may address the second; better process evidence may resolve the third. A company can incur substantial relocation expense while leaving its original problem untouched if the decision remains framed only as geographic diversification.

This article supplies a comparison method for procurement, operations, engineering, finance, and trade teams. It combines external research with original planning worksheets. The external research establishes context and data limitations; the worksheets show how to interrogate a particular investment. A connected supply-chain software workflow can maintain the evidence, but the decision still depends on demonstrated process capability, commercial commitments, and a realistic fallback when assumptions change.

Three strategies, overlapping networks

What China+1, friend-shoring, and regional sourcing actually change

For this comparison, China+1 means retaining some China-based activity while developing production in at least one additional country. It might duplicate final assembly, move one product family, or qualify an alternative component source. The phrase does not specify the additional country's distance from customers, political relationship, ownership structure, or upstream dependencies. An additional assembler that purchases the same sole-source board is a different investment from an independent board-and-assembly capability.

Friend-shoring means giving additional weight to sourcing locations with desired policy relationships or institutional compatibility. The relevant relationship depends on the buyer, customer destination, technology, ownership, and transaction. Define the actual benefit: a usable agreement, permissible technology transfer, predictable documentation, or a customer eligibility requirement. An undefined friendliness score invites opinions that cannot be audited and may change when a product crosses a different border.

Regional sourcing prioritizes proximity to the market being served. It can include nearby cross-border production rather than domestic production alone. One network can meet all three descriptions: an additional plant outside China, located in a preferred partner country, supplying a nearby customer region. Compare that physical design once. Do not count three strategy labels as three independent resilience benefits.

Strategy comparison: potential advantages require supplier- and lane-specific evidence
Decision dimensionChina+1Friend-shoringRegional sourcing
Primary design intentionAdd an option beyond existing China capacityFavor defined policy and institutional relationshipsPlace selected supply nearer customer demand
Potential operating benefitPreserve incumbent know-how while qualifying another routeAlign sourcing with market-access or customer requirementsReduce response distance for suitable products
Common hidden dependencyThe new site uses the incumbent's tooling or upstream sourceAligned countries still share critical materials or technologyNearby assembly imports long-lead components
Economic questionWhat ongoing volume keeps both sites capable?What measurable benefit justifies any premium?Does inventory and service value cover higher conversion cost?
Best evidenceIndependent production trial and switch exerciseTransaction-specific access and eligibility assessmentEnd-to-end replenishment and transport observations
Reason to deferThe alternative cannot make approved goods during disruptionThe preferred relationship provides no applicable benefitLocal capacity is unqualified or its upstream route dominates lead time

Read the evidence at the right scale

Public data can challenge a strategy without choosing your supplier

The OECD's 2025 supply-chain resilience analysis models broad relocalisation scenarios that reduce global trade by more than 18% and real global GDP by more than 5%, without a consistent improvement in resilience. Those are economy-wide model results, not a forecast for your factory or a measured penalty on every regional sourcing project. They challenge the assumption that moving everything closer automatically improves outcomes; they do not settle whether a particular regional line deserves investment.

The World Bank's LPI 2.0 provides 21 country-level logistics indicators, including six core indicators, using observed shipment movements. It replaces the former survey approach and does not offer a single overall country ranking. Use relevant indicators to investigate connectivity, time, and reliability, then obtain route-specific observations. A national indicator cannot establish the clearance time for one electronics importer or the actual truck transit to a particular assembly plant.

Keep the distinction between context and decision evidence visible on the worksheet. Public datasets can identify questions worth testing. Supplier trials, signed capacity reservations, customs analysis, and customer approvals answer different questions. Combining these evidence types is useful; silently treating them as interchangeable creates precision that the underlying information cannot support.

>18%Modelled decline in global tradeOECD 2025 broad relocalisation scenarios; this is a model result, not a supplier-level cost estimate.Source oecd-resilience
>5%Modelled reduction in global real GDPSame OECD scenarios; their scale and assumptions differ from a selective regional sourcing project.Source oecd-resilience
21Country-level LPI 2.0 indicatorsWorld Bank framework has six core and fifteen supplementary indicators, with no single overall ranking.Source world-bank-lpi

Evidence: oecd-resilience, world-bank-lpi

Worksheet 1 · Freeze the decision boundary

Write a comparable sourcing brief before requesting quotes

Give every candidate the same demand curve, product revision, acceptance criteria, packaging specification, destination, delivery obligation, and forecast horizon. Specify whether the quote includes components, customer-owned tooling, test equipment, warranty handling, and engineering support. A lower quote that excludes expensive obligations is not evidence of a more efficient country. Likewise, a larger factory is not necessarily offering more capacity to your program.

Separate committed base demand, plausible upside, and launch uncertainty. Define a smallest sustainable production allocation for an alternate source: enough recurring work to maintain operator experience, equipment settings, and current material records. Compare that allocation with the buyer's willingness to pay for it. A dormant supplier with an old drawing and an unsigned promise may be a candidate, but should not count as operating backup capacity.

Specify the service promise in operational terms. For example, the hypothetical program used later requires 100,000 good units per year, approval of production tooling, traceable component lots, and a proven route to the customer. Its service target is delivery reliability rather than the lowest quoted transit time. The brief also records who may approve deviations. Without that authority map, a launch team may trade away the very quality or resilience objective the board funded.

  • Scope: product family, engineering revision, bill of materials, test specification, and destination customers.
  • Demand: monthly base, peak, downside, minimum order sizes, and firm-order versus forecast responsibilities.
  • Service: required delivery windows, recovery target, packaging, traceability, and customer approval conditions.
  • Commercial basis: currency, payment timing, delivery terms, tooling ownership, warranty scope, and quote validity.
  • Decision ownership: named approvers for engineering, quality, trade, operations, finance, and final allocation.

Worksheet 2 · Establish eligibility

Use mandatory gates before weighted preferences

A weighted score works only after unacceptable choices are excluded. Set gates for lawful transactions, required customer approvals, product safety and quality requirements, verified capacity, and access to necessary technology. A gate has a test, an accountable reviewer, an evidence artifact, and a status. The statuses should distinguish passed, failed, and unresolved. An unanswered question does not become a pass because the sourcing deadline approaches.

Some gates concern the route rather than the supplier. A plant may be capable of producing the part, but the chosen process may not support the proposed origin claim, or the transfer of manufacturing files may require authorization. Record the exact process, entities, and destinations assessed. Broad labels such as approved country conceal too much. Changes to a sub-supplier, manufacturing step, or end customer can reopen a previously settled question.

Conditional selection is useful when the unresolved item can be bounded. A buyer might authorize sample tooling while prohibiting volume transfer until testing is complete. State the spending cap, expiry date, and stop condition. That creates a practical path to learn without representing an immature route as production ready. Once every mandatory gate passes, compare the options that remain using transparent preferences and economic analysis.

Worksheet 3 · Compare the complete cost

Calculate landed cost per accepted unit with consistent boundaries

Start with the contractual purchase price and add the costs that the buyer will actually bear: transport, insurance, customs and nonrecoverable duties, brokerage, handling, expected inspection and defect costs, inventory carrying expense, and recurring coordination. Treat recoverable taxes as a cash-timing item where appropriate rather than automatically an expense. Keep one-time tooling, qualification, and transfer spending separately visible before calculating a horizon-specific total. That distinction prevents a mature route and a new launch from being compared on incompatible bases.

A useful worksheet identity is annual recurring cost = accepted units × recurring cost per accepted unit. For the investment decision, discounted total cost = initial investment + sum of annual net cash costs divided by (1 + discount rate) to the relevant year. Define whether the model is nominal or real and use a matching discount rate. Avoid adding both a financing charge and a full capital cost for the same cash flow without a clear accounting reason.

The hypothetical examples here use a simpler undiscounted three-year comparison to keep the arithmetic inspectable. All three options deliver the same accepted quantity. They assume the displayed defect and inventory allowances already capture the chosen yield and stock policy, so those costs are not added again elsewhere. In a live investment case, finance should reconcile the unit model to purchasing, inventory, tax, and cash-flow records before interpreting a difference as savings.

Illustrative dollars per accepted unit; invented planning assumptions, not country benchmarks
Cost elementA · China+1 routeB · Partner-country routeC · Regional route
Purchase price$50.00$52.00$54.00
Transport and handling$2.00$1.60$0.80
Assumed duty and brokerage allowance$3.00$1.00$0.70
Inspection and defect allowance$1.00$0.80$0.60
Inventory carrying allowance$1.00$0.90$0.50
Recurring coordination$0.50$0.70$0.80
Recurring total$57.50$57.00$57.40

Read the cost result carefully

A unit-price winner can lose the investment comparison

At the assumed annual demand of 100,000 accepted units, route A costs $5.75 million annually, B costs $5.70 million, and C costs $5.74 million. Suppose initial qualification and tooling spending is $150,000, $400,000, and $600,000 respectively. Three years of recurring cost plus that initial spending gives $17.40 million for A, $17.50 million for B, and $17.82 million for C. B has the lowest recurring cost, yet A has the lowest undiscounted three-year total under these assumptions.

The incremental investment for B over A is $250,000, while the recurring saving is $50,000 a year. The simple payback is five years, before discounting, taxes, residual values, or changes in demand. C carries a $450,000 initial premium over A and saves only $10,000 a year in the baseline. These results do not make C irrational; they make its service, flexibility, or disruption benefit an explicit part of the decision rather than an unpriced assertion.

Recalculate using downside demand and the actual allocation. If an alternate receives only one-quarter of the program volume, it earns savings on that allocation, not on every unit in the global business. Conversely, minimum purchase commitments may force spending when customer demand disappears. Treat residual tooling value cautiously: ownership has little economic value if the tool cannot be removed, serviced elsewhere, or used without the incumbent's proprietary process knowledge.

Worksheet 4 · Capacity that can be called

Measure qualified output rather than building size

Ask the candidate to demonstrate the actual process at the intended staffing level, material supply, product mix, and inspection burden. Rated machine capacity usually omits at least some downtime, changeovers, reject production, staffing gaps, or capacity already promised to others. A credible offer shows the bottleneck resource and the allocation available to your program. It also explains which improvement assumptions remain unproven.

For a hypothetical line, 160 scheduled hours per week multiplied by 75% productive availability, 50 units per productive hour, and 96% accepted yield gives 5,760 accepted units per week. If the supplier commits only 60% of that output to your program, your illustrative share is 3,456 units. Do not multiply by yield again if the quoted rate already describes accepted units. Keep definitions explicit because double-counting efficiency reductions can be as misleading as ignoring them.

Evaluate peak recovery separately from normal capacity. During an incumbent outage, an alternate may need raw material, additional shifts, customer approval, and transport bookings before it can increase output. Compare the time to reach useful volume with inventory coverage at the affected customer. A larger eventual capacity has limited value if the buyer runs out of parts before the alternative becomes operational. The qualification plan should test that interval rather than assume a signed purchase agreement closes it.

Trace the dependencies beneath the country label

Two suppliers can still be one supply chain

Map critical dependencies by physical production site and process, not only the name on the invoice. Two contract manufacturers might share a semiconductor package source, an alloy specification, a toolmaker, a test laboratory, or a logistics gateway. Separate ownership is helpful information, but it does not establish independent operation. Prioritize the components and processes whose loss prevents shipment, even when their purchase value is small.

OECD TiVA measures value added across international production chains, which helps explain why final export origin and upstream economic exposure differ. It is an industry-level analytical resource, not a bill of materials for a named supplier. UN Comtrade offers reported merchandise trade flows, useful for exploring sourcing patterns. Neither dataset reveals your actual factory's unique sub-tier network. Use those sources to form diligence questions, then substantiate answers with supplier and engineering records.

A dependency worksheet should capture the component, producing site, approved substitutes, tooling ownership, qualification lead time, replenishment route, and evidence date. Leave undisclosed origins marked unknown. If an alternative assembler cannot disclose a critical dependency, the team should carry that uncertainty into its resilience assessment. Assuming independence because no shared supplier has been identified confuses missing information with evidence of separation.

Evidence: oecd-tiva, un-comtrade

Worksheet 5 · Concentration arithmetic

Calculate concentration at more than one tier

For a simple screening metric, calculate HHI as the sum of squared supply shares using decimal fractions. A 70/30 split between two assembly sites produces 0.70 squared plus 0.30 squared, or 0.58. An equal three-way split produces approximately 0.333. Lower values indicate a more even distribution across the entities counted. They do not establish that those entities can survive the same shock, nor do they establish sufficient replacement capacity.

Now suppose the 70/30 assemblers buy 90% of their combined critical-chip demand from one fabrication route and 10% from another. At that upstream level, the corresponding concentration is 0.82. The buyer improved assembly distribution while retaining a much more concentrated critical input. This hypothetical calculation demonstrates why the denominator and tier belong next to the number. A purchase-spend HHI may hide a cheap part that stops an expensive product.

Add a practical outage test beside the concentration metric: remove each critical node in turn and estimate how many approved units can still reach customers, by week. Include shared utilities, logistics nodes, and tooling where they matter. This is a scenario exercise, not a claim that every node failure has the same likelihood. Its purpose is to reveal a gap that a tidy supplier-count metric may conceal and to identify the smallest investment that closes it.

Trade measures require product-level work

Compare origin, classification, and duties as separate fields

A shipping country, a supplier's address, and a legally determined origin can be different. The European Commission explains that non-preferential origin uses its own rules, including product-specific provisions and substantial-transformation analysis where applicable; minimal operations do not confer origin. Preferential origin has a separate role in claiming agreement benefits. A sourcing proposal therefore needs a documented manufacturing process and applicable rule analysis before it claims a tariff advantage.

WITS is useful for screening historical tariff patterns, but its technical note warns that computed applied rates are indicative and can omit unilateral increases outside its MFN and preferential inputs. Do not paste a country average into the landed-cost model and call it the expected duty on a particular imported product. The live assessment requires the relevant national tariff line, origin, effective date, customs value, measures, and any supported preference claim.

Give the trade team a change log showing which manufacturing steps move, which materials retain their suppliers, and which entities buy and sell the goods. Store the conclusions and assumptions with the exact product revision. If the proposed advantage depends on an interpretation that remains unresolved, run the economics both with and without it. Our guide to AI tariff classification versus human broker review explains why preserving product facts and reviewer reasoning matters even when software accelerates the research.

Evidence: eu-origin, eu-preferential, wits-tariffs

Policy alignment needs a transaction test

Make friend-shoring benefits concrete and reviewable

Replace a broad political compatibility rating with a list of operational requirements. Can the proposed plant receive the necessary manufacturing technology? Can the intended end user receive the finished item? Does a customer procurement rule apply? Is there an agreement benefit for this product and process, and can the evidence be maintained? These questions make the strategy testable without pretending that a single label predicts every future policy decision.

BIS guidance illustrates the need for that specificity: U.S. export controls can apply to reexports and certain foreign-produced items, with analysis depending on the item, technology connection, destination, end user, and end use. Moving assembly across a border does not itself settle whether a transaction is subject to the EAR or whether authorization is needed. Assign that determination to the responsible trade specialist using current rules and transaction facts.

Build a policy register with an owner, review trigger, expiry or renewal date where relevant, and the affected products. A change to a restricted party, controlled technology, sourcing rule, or customer condition should prompt a targeted assessment. The useful output is an action: continue, obtain approval, change the process, qualify another route, or stop a transaction. A generalized country dashboard is less useful if it cannot identify which purchase orders need attention.

Evidence: bis-reexports

Regional economics depend on the whole route

Shorter distance is valuable when it shortens the decision cycle

Map the route from constrained input through production, border handling, consolidation, delivery, and acceptance at the customer. Nearby final assembly may still wait for an imported component. Conversely, postponing configuration near demand can improve response without relocating every upstream step. Compare alternatives at the point where a customer can actually change its order, because that is where proximity may create commercial value.

Use measured lead-time distributions instead of only average days. Separate waiting for production, transit, customs, receiving, and quality release. Record the observed sample size and seasonal coverage. If a proposed regional route has only three trial shipments, treat its apparent reliability as provisional. The World Bank's LPI redesign reinforces the value of observed movements, but its country indicators remain a context source rather than a substitute for these lane-level observations.

Value the shorter cycle through a named mechanism: lower inventory, fewer expedites, less obsolete stock, improved order acceptance, or reduced customer downtime. Avoid adding all of those benefits when they arise from the same event. For example, an inventory reduction and a financing benefit may describe the same released cash. Our supply-chain visibility guide connects transport events with production and customer promises so that service claims can be tested against the actual sequence.

Evidence: world-bank-lpi

Ramp time is part of the sourcing cost

Budget the learning curve, engineering work, and quality containment

A transferred drawing does not transfer every useful production habit. The incumbent may hold knowledge about fixture wear, material variation, machine settings, test interpretation, packaging damage, or difficult changeovers. Ask engineering and quality to identify which knowledge is documented, which is embodied in equipment, and which depends on particular people. Assign a transfer method to each important gap.

Structure the ramp around evidence milestones: tooling acceptance, material approval, process trials, measurement-system checks, representative production, and customer release where required. These are recommended planning categories, not a substitute for the customer's contractual or industry-specific process. Define acceptance thresholds before trials begin. Moving the threshold after poor results may conceal a capacity or quality issue that will reappear at volume.

Include temporary duplication in the business case. The company may need incumbent production, qualification batches, additional inspection, engineering travel, and buffer inventory simultaneously. A relocation plan that counts only the steady-state factory price misses this overlap. Track launch defects separately from mature-process defects and state when the cost model expects convergence. If the alternate remains expensive after the planned learning period, the decision should be reopened rather than indefinitely described as a temporary ramp problem.

Electronics lens

Separate assembly flexibility from component availability and IP access

For an electronics program, examine the approved component list before choosing the final assembly geography. The useful alternate may require compatible board fabrication, assembly capability, firmware loading, secure key provisioning, test coverage, and access to approved components. If the same constrained chip feeds every route, relocating the enclosure assembly may improve a shipping lane while doing little for product availability.

Compare who owns the test fixtures and software, who can update them, and whether a second site can reproduce the accepted results. Preserve golden samples, configuration records, calibration evidence, and change histories. A duplicate fixture that runs an outdated program can produce false confidence. Plan controlled access to design files and manufacturing data, and define how supplier-created process improvements may be used at another site.

Use product architecture to create options where it is economical. Engineering might qualify an alternate component, standardize an interface, or postpone market-specific configuration. Each change needs its own performance and approval work, so do not assume redesign is faster than sourcing. Compare the cost and time of a more transferable design against the cost of preserving a specialized incumbent process. The strongest result may be selective portability rather than a complete factory move.

Automotive lens

Tie regional content analysis to approved production processes

An automotive sourcing choice can be constrained by customer approval, traceability, process validation, tooling, delivery sequencing, and service-part obligations. Build the comparison around the actual component and vehicle program. A supplier capable of another customer's similar part is a candidate for qualification, not automatic evidence that it can ship the exact approved part at the required rate.

The USTR's 2024 USMCA automotive report describes a 75% regional value content requirement for passenger vehicles and light trucks, alongside separate requirements for specified core parts. That dated report illustrates why regional sourcing and origin calculations must be evaluated at the applicable product and program level. It is not a statement that every automotive component has a 75% rule, nor proof of a particular shipment's current tariff treatment. Use the governing current provisions and relevant official guidance for the actual transaction.

Allocate launch responsibilities explicitly. Identify who funds duplicate tools, approves process changes, retains traceability records, and supplies parts during the transition. Assess whether the customer must approve a site move and whether service commitments outlast the main production contract. A regional factory can offer valuable replenishment flexibility, but an unapproved process or inaccessible tool can eliminate that benefit when demand becomes urgent.

Evidence: ustr-auto-report

Worksheet 6 · Make preferences visible

Build an eight-factor scorecard with anchored ratings

After the gates pass, assign weights before scoring candidates. In the illustrative worksheet, total landed economics receives 25%, qualified capacity 20%, independence from shared failures 15%, trade-policy fit 10%, delivery response 10%, ramp and quality 8%, IP and process access 7%, and reversibility 5%. These are example preferences, not industry standards. An automotive launch with a fixed customer deadline may reasonably place more weight on approved capacity than a mature, low-variability electronics family.

Use a one-to-five scale anchored to observable outcomes. For capacity, a score of one might mean only minimum required volume is proven with no recovery headroom; three might mean the agreed peak is demonstrated; five might mean peak plus a defined recovery increment is demonstrated and reserved. The mandatory minimum still belongs in the gates. Scores describe differences among eligible options, while unresolved evidence appears as an interval or an explicit condition.

Calculate score out of 100 as the sum of each percentage weight multiplied by its rating divided by five. Retain the underlying measurements next to the ratings. A score is a structured expression of preferences, not a probability of success. Two reviewers should be able to reproduce it from the same evidence and identify exactly where they disagree.

Illustrative weighted worksheet; all ratings are invented for three eligible candidate networks
Criterion and weightA ratingB ratingC rating
Landed economics · 25%433
Qualified capacity · 20%434
Independent dependencies · 15%243
Trade-policy fit · 10%243
Delivery response · 10%335
Ramp and quality · 8%433
IP and process access · 7%343
Reversibility · 5%432
Weighted total out of 10066.666.467.0

Worksheet 7 · Challenge the ranking

A 0.4-point lead is a request for better evidence

The example regional route scores 67.0, only 0.4 above A and 0.6 above B. That difference is too small to justify a sweeping strategic conclusion. If C's capacity rating falls from four to three after a production trial, its score loses four points and becomes 63.0. If A's dependency evidence improves by one rating point, A gains three points. The apparent winner is therefore sensitive to questions that can be investigated directly.

Test weights as well as ratings. If the team moves five percentage points from delivery response to landed economics, A gains one point because its economics rating exceeds its delivery rating by one; C loses two points because its delivery rating exceeds its economics rating by two. A then scores 67.6 and C 65.0. This reversal does not invalidate the tool. It reveals that the decision depends on how much the buyer values response time relative to economics.

Use the sensitivity result to prioritize learning. A trial that resolves C's available capacity may be worth more than another broad country report. A tooling-rights review may resolve whether A deserves its reversibility score. Put a cost and duration against each investigation, and ask whether the result could change the choice. The team should buy evidence that affects the decision rather than accumulate documentation merely to make the recommendation look substantial.

Worksheet 8 · Stress the operating plan

Compare surviving output under a small set of explicit shocks

Run a baseline, a concentrated upstream outage, a transport interruption, a policy change, and a demand downturn. Describe each event precisely enough to calculate consequences: affected nodes, duration, available inventory, alternate capacity, transfer delay, and customer demand. Do not assign arbitrary probabilities simply to produce an expected-value answer. When probabilities are poorly supported, show scenario costs and service outcomes separately.

For each week, unmet demand can be calculated as the greater of zero and demand minus usable beginning inventory minus accepted receipts, with inventory carried forward consistently. Receipts must respect component availability, qualified capacity, and transport timing. Estimate consequences from the actual contract and commercial situation; distinguish lost contribution, delayed revenue, expedite spending, and customer penalties. Adding full lost sales revenue to lost margin for the same missed order would overstate damage.

The New York Fed's GSCPI combines transport-cost and manufacturing indicators into a broad supply-pressure gauge. It can prompt a review of assumptions, but it is not a forecast for a named supplier or route. A rising macro index should not automatically trigger a factory transfer. The relevant action depends on whether the observed pressure affects the constrained materials, transport capacity, or timing in the specific network under consideration.

Evidence: nyfed-gscpi

Preserve the option to change course

Price reversibility before signing the long commitment

A sourcing decision includes the ability to revise it. List the assets and permissions needed to switch: movable tooling, current drawings, process settings, test software rights, approved material alternatives, customer release, and capable staff. Assign each an owner and retrieval method. A contract stating that the buyer owns a tool is incomplete operating evidence if removing it requires months of dispute or if another plant cannot run it.

Compare the cost of keeping an option alive with the cost of recreating it later. An alternate may need periodic orders, fixture maintenance, software updates, training, and material qualification to remain useful. Model those recurring costs openly. An option that receives no work and no maintenance should not retain the same readiness rating year after year. Conduct a small transfer exercise using current data to discover what has decayed.

Use staged commitments where they fit the commercial relationship: sample development, bounded pilot volume, then larger allocation after evidence improves. Suppliers also need a viable business case, so do not demand unlimited capacity reservations without compensation or credible demand. Negotiate mutual triggers for expansion, price review, and exit. This makes reversibility a managed operating arrangement rather than a unilateral wish written into a sourcing presentation.

Implementation worksheet

Move from a country shortlist to an approved volume allocation

Sequence the work so that expensive commitments follow the evidence they require. Some diligence can run in parallel, but production qualification cannot be completed by a calendar label. The following sequence is a project structure, not a promise that every transfer fits a standard number of days. A complex automotive program may need a different schedule from a mature electronics accessory.

Keep an issue register that records each unresolved assumption, the consequence if it is wrong, the next test, and the person empowered to close it. Procurement coordinates the process but should not silently approve engineering, trade, or quality questions. The final allocation decision should include the remaining conditions, the forecast basis, the first review date, and the evidence that would cause the allocation to change.

  1. Define the product and service boundary

    Freeze the comparable brief, identify the customer promise, and agree on mandatory gates. Capture the incumbent's real cost and performance using the same definitions that candidates will face.

  2. Build candidate network records

    Collect site-specific capacity, tooling, sub-tier, logistics, and process information. Separate supplier statements from observed evidence and mark undisclosed dependencies explicitly.

  3. Resolve transaction and approval questions

    Confirm the relevant origin, classification, control, and customer conditions for each proposed route. Record which changes would require the assessment to be repeated.

  4. Run the smallest useful production tests

    Use representative materials, staffing, equipment, and product mix. Test quality and available capacity together; preserve results that explain both successful and unsuccessful runs.

  5. Reconcile economics and challenge scores

    Update landed cost, transfer spending, inventory, and realistic allocation. Apply the agreed weights, then vary the disputed assumptions and calculate whether the ranking changes.

  6. Release volume in evidence-based stages

    Agree on the starting allocation, ramp limits, stop conditions, and incumbent coverage. Increase volume only after the relevant acceptance evidence is available and the next stage is commercially supported.

  7. Exercise and maintain the fallback

    Keep alternative tooling and records current, run periodic orders or transfer tests, and review readiness after design changes. Measure how much approved output the network can restore within the customer recovery window.

Make the decision survive staff turnover

Keep a sourcing record that explains why the allocation exists

Store the decision as a versioned evidence package: demand assumptions, product scope, cost model, gates, scorecard, sensitivity analysis, approvals, and open conditions. Link each important score to a trial, contract, data extract, or specialist assessment. A future team should be able to tell whether an allocation was chosen for low cost, customer proximity, independent capacity, or a combination. Otherwise, the next negotiation may remove a deliberately purchased resilience benefit in pursuit of a narrower price target.

Define review triggers rather than relying only on an annual meeting. Examples include a material design change, an upstream source substitution, unexpected yield deterioration, a contract renewal, a trade measure change, or an alternate failing a readiness exercise. Match the review to the affected assumption. A minor packaging update need not reopen every country comparison, while a change to the sole critical component may warrant immediate reassessment.

A data-management layer can keep supplier identities, sites, product revisions, dependencies, and evidence dates consistent across systems. The useful dashboard shows accepted output, delivered cost, quality escapes, readiness, and unresolved conditions alongside the chosen allocation. A strategy remains credible when its assumptions are still true and its alternatives still work. The label on the map matters less than that continuing proof.

Explore the connected roadmap

Use these related service, technology, and industry pages to compare next steps and keep the topic connected to real implementation choices.

01

Supply Chain Software

Connect supplier evidence, allocations, and operational performance.

02

Supply Chain Visibility

Link transport events to production and customer commitments.

03

Data Management

Maintain consistent supplier, site, product, and evidence records.

Supply Chain Software

Connect supplier evidence, allocations, and operational performance.

Supply Chain Visibility

Link transport events to production and customer commitments.

Data Management

Maintain consistent supplier, site, product, and evidence records.

FAQ

Is China+1 the same as leaving China?

No. In this article, China+1 retains selected China-based activity and develops an additional production option elsewhere. The allocation can vary by product, customer, or process. The relevant test is what the added route can deliver when the incumbent is unavailable, including access to materials, tooling, approvals, and logistics.

Can one sourcing network be China+1, friend-shored, and regional?

Yes. An additional factory outside China can also be in a preferred partner country and near the customer market. Evaluate the physical network once, then identify its distinct benefits. Counting the three labels separately would exaggerate diversification if they describe the same capability.

Which strategy has the lowest total cost?

There is no universal ranking. Compare accepted-unit economics, initial transfer spending, realistic volume allocation, inventory, quality, and service requirements over the same horizon. A route with the lowest recurring cost may lose a shorter investment comparison because its tooling and qualification costs are higher.

How much volume should an alternate supplier receive?

Enough to sustain the capability you intend to rely on, subject to its commercial minimums and your economics. Determine that through process needs, staff continuity, maintenance, material refresh, and readiness tests. A fixed percentage chosen without those facts may either waste money or leave the alternate unable to respond.

Does assembly in a different country automatically change origin?

No. Origin depends on the applicable jurisdiction, product, and manufacturing facts. Preferential and non-preferential rules serve different purposes. Obtain a supported analysis of the proposed route, keep its process evidence current, and do not infer eligibility or duty treatment from a shipping address alone.

How do we compare options when supplier data is incomplete?

Mark the uncertain field, show a plausible score or cost interval, and test whether resolving it could change the decision. Fund targeted diligence where the result matters. Do not convert missing sub-tier information into a claim of independence or give an unresolved mandatory gate an automatic pass.

When is regional sourcing most compelling?

It becomes compelling when shorter response produces enough verified value to cover the complete cost and qualification effort. Relevant mechanisms include later configuration, lower obsolete inventory, reduced expediting, and customer service. Test the complete upstream route because nearby final assembly can still depend on distant constrained inputs.

What belongs in a board-level recommendation?

Show the actual network, passed gates, cost horizon, chosen weights, sensitivity results, scenario service outcomes, and staged commitments. State which benefits are measured and which remain assumptions. Include the conditions that would change the allocation and the ongoing cost of keeping the fallback capable.

Illustrative sourcing committee case

A controller manufacturer buys two different kinds of flexibility

Consider a fictional manufacturer of industrial controllers sold to equipment makers in North America and Europe. It currently assembles every unit at one China-based contract manufacturer. Its board asks for diversification, but the operations team identifies two separate problems: a common board supplier could interrupt the entire product family, while late customer configuration changes create expedited shipments. This is an invented decision scenario, not a client result or a claim about any real supplier.

The shortlist includes a second Asian assembler, a plant in a preferred partner country, and a regional configuration operation near one customer market. The initial map makes the regional option look most independent. The dependency review shows that all three still rely on the same board design and critical chip route. The regional operation can respond quickly to configuration changes but cannot replace board production during an upstream outage. The committee therefore refuses to credit it with a full manufacturing-resilience benefit.

Engineering proposes two bounded investments: qualify a second board route with appropriate component and test work, and pilot regional final configuration using common semi-finished inventory. Procurement obtains the economics for each separately. Finance compares the board investment against outage scenarios and the configuration pilot against actual expedite and obsolescence records. The regional pilot must demonstrate that shorter response time produces measurable customer and inventory value before it receives more volume.

The resulting decision is a portfolio rather than a slogan. The incumbent retains work for which its approved process remains competitive; an alternate board route develops a specific recovery capability; the regional operation earns expansion through observed service results. The committee stores the assumptions and reviews them after the production trials. It does not report savings or risk reduction before results exist. The lesson is to match each investment to the failure or delay it can actually change.

  • Separate the upstream outage problem from the late-configuration problem before comparing investments.
  • Avoid awarding geographic diversification credit to a route that retains the same critical dependency.
  • Tie each pilot to a measurable decision: approved alternate output or reduced response cost.
  • Preserve incumbent capability while testing alternatives under bounded commitments.

Research scope and calculation method

Researched on October 8, 2026 using primary publications and official data documentation from OECD, the World Bank, the United Nations, the New York Fed, the European Commission, BIS, and USTR. Public statistics provide context; none is used to claim that a particular country, supplier, or strategy universally wins. The OECD relocalisation figures are explicitly modelled macroeconomic outcomes. LPI 2.0 is described according to its current methodology, with no invented overall country ranking. The USTR automotive figure is attributed to its 2024 report and is used as a dated illustration of rule complexity, not a current shipment determination. All supplier costs, volumes, capacities, ratings, concentration examples, paybacks, and case details are hypothetical editorial calculations. Dollar examples use consistent accepted-unit boundaries and an explicitly undiscounted three-year horizon; a live investment requires its own cash-flow, tax, currency, timing, and uncertainty analysis. Section source references identify the factual context; the scoring framework and implementation recommendations are original analysis, not official guidance from the cited institutions.

Research ledger

Sources and further reading

  1. OECD Supply Chain Resilience Review: Navigating RisksOECD · 2025-06-02

    Primary source for modelled global trade and GDP changes under broad relocalisation scenarios; not a supplier-level forecast.

  2. About the Logistics Performance Indicators 2.0World Bank

    Defines the 21 indicators, observed movement basis, 2023–2024 data coverage, and absence of a single overall country ranking.

  3. Trade in value-addedOECD

    Explains how TiVA indicators trace value added across production chains and identifies the revised 2025 database.

  4. UN Comtrade Data Explorer: AboutUnited Nations Statistics Division

    Official description of reported merchandise trade statistics and partner-country concepts; not used as a supplier-specific dependency map.

  5. Technical note on bilateral applied tariffWorld Bank, World Integrated Trade Solution

    Explains aggregation and limitations of indicative applied tariff data, including unilateral increases outside MFN and preferential inputs.

  6. Non-preferential rules of originEuropean Commission, Taxation and Customs Union

    Official account of EU non-preferential origin concepts, product-specific provisions, minimal operations, and evidence responsibilities.

  7. Preferential rules of originEuropean Commission, Taxation and Customs Union

    Explains the role of preferential origin in EU trade agreements and unilateral arrangements; transaction-specific eligibility must be checked separately.

  8. What is a reexport?U.S. Bureau of Industry and Security

    Official guidance on reexports, foreign-produced items, and item, destination, end-use, and end-user considerations under the EAR.

  9. 2024 Report to Congress on the Operation of the United States–Mexico–Canada Agreement with Respect to Trade in Automotive GoodsOffice of the United States Trade Representative

    Dated primary report used for the 75% passenger-vehicle/light-truck regional value content illustration and separate core-parts requirements; not presented as a complete current tariff determination.

  10. Global Supply Chain Pressure IndexFederal Reserve Bank of New York

    Describes the combined transport-cost and manufacturing indicators used as a broad pressure gauge; the index is not an official Federal Reserve estimate.

Continue exploring

Related bizz insights

Compare adjacent approaches, implementation choices, and operating practices across these closely connected guides.

Supply Chain Strategy

Just-in-Time vs Just-in-Case vs DDMRP

A research-backed comparison of Just-in-Time, Just-in-Case, and DDMRP across working capital, service, obsolescence, shock absorption, and three transparent volatility scenarios.

36 min read
Supply Chain Strategy

Single Sourcing vs Dual Sourcing vs Multisourcing

Compare supplier concentration, qualification costs, available capacity, and correlated risk with a practical sourcing playbook for medical devices and other regulated products.

31 min read
Transportation and Logistics

AI Ocean-Rate Forecasting vs Index-Linked Buying and Forwarder Quotes

A myth-versus-data investigation of AI ocean-rate forecasts, index-linked buying, and forwarder quotes, with backtests, basis-risk controls, and procurement decision rules.

40 min read
Warehousing and Fulfillment

AS/RS vs AMRs vs Conveyor Automation

An operator-first comparison of AS/RS, autonomous mobile robots, and conveyor/sortation automation across throughput, density, flexibility, safety, maintenance, integration, expansion, and five-year TCO.

41 min read

Make sourcing decisions traceable to operating evidence

bizz builds supply-chain and data workflows that connect supplier capabilities, landed cost, approvals, and performance so teams can act on a shared view of the production network.

Explore supply-chain software