A factory that has supplied the same product for six years will still ship a bad lot in the seventh. The difference between a buyer that recovers the loss and one that absorbs it is rarely the quality of the relationship. It is whether the relationship was documented well enough that a bad month has a consequence.

Supplier relationship management in China is usually presented as a matter of trust, visits and quarterly meetings. Those things matter, and they are not a remedy. The metrics that make a business review meaningful only bind the factory if they appear in the agreement, with a defined data source and a stated consequence. A review that produces a chart and an apology produces nothing that survives the next order.

The practical conclusion is that every KPI worth tracking should exist twice: once as a dashboard line, and once as a contractual obligation with a remedy attached. This article covers the metrics that matter, how to write them into the agreement, what remedy each one should carry, how to run the reviews that keep them honest, and where the cultural dimension of Chinese supplier management intersects with the legal record.

Why a relationship still needs a contract behind it

Factories respond to customers who are worth keeping. Volume, payment reliability, forecast accuracy and personal regard all influence how a factory allocates its best line, its best engineers and its scarce capacity in a tight quarter. That is real, and it is why the relationship deserves investment.

It is also unstable. The account manager who values the relationship leaves. The factory takes on a larger customer and the priority shifts. A new owner buys the plant and reviews every contract for margin. In each case the goodwill that carried the relationship does not transfer, and what remains is the document.

A weak contract produces a specific pattern of loss. Quality problems are met with a promise to improve rather than a corrective action plan. Delivery slips are explained rather than compensated. Cost reduction commitments evaporate at the annual review. The buyer has no lever except to move the business, which is expensive, slow and often impossible in the middle of a product cycle.

A contract that carries the metrics converts each of those conversations into a defined process. The factory knows what is measured, how it is measured, and what happens when the number is missed. The buyer knows what it can ask for without renegotiating the relationship. Most suppliers prefer this arrangement, because it removes ambiguity and because a well-run customer is easier to plan around.

There is a second reason to put the numbers in the agreement. Chinese commercial litigation and arbitration are document-driven, and a claim for poor performance requires a standard that was agreed in advance. A buyer that complains about a ninety-two per cent on-time delivery rate under a contract that promises only “timely delivery” has a complaint. The same buyer under a contract that defines on-time delivery, sets a target, requires monthly reporting and provides a service credit has a claim.

Measuring performance: the metrics that matter

What gets measured gets done, provided the measurement is defined the same way on both sides. Five families of metric cover most of what a buyer needs.

Quality.

KPIFormulaTypical target
First-pass yieldGood units first time divided by total unitsAbove 95 per cent for consumer electronics, above 98 per cent for mature products
Defect rate in parts per millionDefective parts per millionBelow 5,000 is acceptable, below 1,000 is strong
Incoming inspection pass rateLots passing incoming inspection divided by total lotsAbove 95 per cent
Field failure rateField failures divided by total shippedBelow 1 per cent in the first year
Customer complaintsNumber of quality-related complaintsA declining trend
Corrective action closure rateCorrective actions closed on time divided by total requestsAbove 90 per cent on time

Delivery.

KPIFormulaTypical target
On-time delivery rateOrders delivered on or before the promise date divided by total ordersAbove 95 per cent
Lead time accuracyActual lead time divided by quoted lead timeWithin ten per cent of the quote
Order fulfilment rateQuantity shipped divided by quantity orderedAbove 98 per cent
BacklogDays of past-due ordersNil
Delivery responsivenessTime to confirm a rush orderWithin 24 hours

Cost.

KPIFormulaTypical target
Price competitivenessThe buyer’s price against a market benchmarkAt or below the benchmark
Cost reduction achievementActual reduction divided by the target reductionAt or above the target
Quote accuracyFinal price divided by quoted priceWithin five per cent
Total cost of ownershipPurchase cost plus quality cost plus logistics costA declining trend

Responsiveness.

KPIHow it is measuredTypical target
Quote response timeFrom request for quotation to quoteWithin three business days
Email response timeTime to respond to an enquiryWithin 24 hours
Issue resolution timeFrom issue report to resolutionWithin 48 hours for critical issues
Sample lead timeFrom request to sample deliveryTo the agreed schedule

Choosing the metrics. Start with five to seven. Too many metrics dilute attention and produce a dashboard nobody reads. Focus on what the business depends on: a quality-led product needs first-pass yield and field failure rate; a time-to-market product needs new product introduction speed and delivery accuracy. Every metric needs a data source that can be collected consistently, a target that is stated as a number, and a review rhythm — monthly for operational metrics, quarterly for strategic ones. Share the results with the factory, because a metric the supplier never sees cannot change behaviour.

Two design choices decide whether the numbers can be trusted.

The first is who measures. A factory reporting its own on-time delivery rate is reporting against a promise date it set itself, and the figure will improve when the promise dates get longer. The buyer should either measure from its own purchase order date or require the factory to report against a date recorded in a system the buyer can see.

The second is how the metric is defined. Every KPI has a gaming route. On-time delivery can be improved by shipping partial orders, so it must be paired with order fulfilment. Defect rate can be improved by reclassifying defects as cosmetic, so the definition must list what counts. First-pass yield can be improved by excluding rework that happens before the formal inspection, so the measurement point must be specified. The definition, not the target, is where the argument will happen.

From metric to obligation: drafting the clause

A KPI becomes enforceable when it is written as an obligation with six components. Missing any one of them turns the metric back into a management aspiration.

  1. The defined metric, with the formula, the unit, the measurement point and the exclusions.
  2. The data source, naming who produces the data, in what format, and whether the buyer may audit it.
  3. The measurement period, whether monthly, quarterly or per lot.
  4. The target and the tolerance, distinguishing the target from the minimum acceptable level, because a single number gives no room for a bad month.
  5. The reporting obligation, with a due date and a defined report.
  6. The consequence, which is the remedy section below.

A worked example shows how much of the value sits in the definition rather than the number.

MetricContract languageData sourceRemedy
On-time delivery“On-time delivery means delivery to the carrier named in the purchase order on or before the confirmed ship date, where the confirmed ship date is the date stated in the supplier’s order acknowledgement and may not be later than the lead time in Schedule 2. Partial shipments count as on time only if the quantity shipped is at least the ordered quantity.”Buyer’s purchase order system and the carrier’s bill of ladingService credit per late order; expedited freight at the supplier’s cost; escalation after two consecutive months below the minimum
Defect rate“The defect rate is the number of units failing the inspection standard in Schedule 3 divided by the number of units inspected, measured per lot at the supplier’s final inspection and verified at the buyer’s incoming inspection. Cosmetic defects within the limits in Schedule 3 are excluded; functional defects are not.”Supplier’s final inspection record, verified by the buyer’s incoming inspectionRework or replacement at the supplier’s cost; sorting cost recovered; corrective action plan; price adjustment where the defect is accepted
First-pass yield“First-pass yield is the number of units passing final inspection without rework, repair or re-test divided by the number of units entering final inspection, reported weekly by production line.”Supplier’s manufacturing execution system, subject to auditJoint root cause analysis; process audit; capacity held pending improvement
Corrective action closure“A corrective action is closed on time when the containment action is implemented within 48 hours and the root cause and permanent corrective action are verified within the period stated in the request.”Corrective action log maintained jointlyEscalation to the factory’s general manager; charge for repeat containment; review of new orders
Cost reduction“The supplier will deliver the cost reductions in Schedule 5, measured against the baseline prices in that schedule, with the mechanism and the effective date agreed at the preceding quarterly review.”Invoiced prices against the baselineWithholding of the productivity share; benchmarking; re-sourcing right
Response time“The supplier will acknowledge any quality notification within one business day and provide an interim containment plan within two business days.”Email or portal timestampsEscalation; charge for third-party containment where the supplier does not respond

Three drafting habits make the difference between a clause that works and one that reads well.

Use objective standards, not efforts. “The supplier will use its best endeavours to deliver on time” is unenforceable in practice, because it sets no standard against which failure can be measured. “The supplier will deliver on time in accordance with the definition in Schedule 2” is measurable. Where a softer standard is genuinely needed, express it as a process obligation — notification within a period, a recovery plan within a period — rather than as an effort.

Put the detail in a schedule. The body of the agreement should carry the obligation and the remedy. The definitions, the targets, the inspection standard and the reporting template belong in schedules, which can be updated by agreement without re-executing the contract. A schedule that is revised annually, signed and dated, keeps the metrics aligned with the product as it changes.

Tie the obligation to the buyer’s own conduct. Many delivery failures are caused by late forecasts, late purchase orders, late approvals or late payments. The agreement should state what the buyer must provide, by when, and that the supplier’s obligations are adjusted where the buyer is late. A delivery metric that ignores the buyer’s inputs will be disputed at the first review, and the supplier will be right.

What remedy attaches to each metric

A metric without a consequence is information. The consequence should be proportionate, escalating, and defined in advance so that nobody has to negotiate it during a bad quarter.

The ladder below is the standard architecture. Not every metric needs every rung.

RungWhat it isWhen it appliesDrafting point
ReportingA monthly or weekly report against the metricEvery metric worth trackingFix the template and the due date
Corrective actionA written plan with containment, root cause, permanent action and verificationAny breach of the minimum levelDefine the response deadline and the verification standard
Cost recoveryReimbursement of sorting, rework, replacement, freight and inspection costsQuality failures and late delivery that require expedited freightRequires records; state that costs are recoverable on production of invoices
Service creditA credit calculated by formula against the invoice valueRepeated or material missesSet the rate per day or per lot, and a cap
Price adjustmentA reduction in price where the buyer accepts non-conforming goodsQuality failures accepted for useAgree the valuation method in advance
Volume reallocationMovement of volume to another approved sourcePersistent failure after corrective actionRequires a qualified second source to exist
Step-inThe buyer or a third party performs the supplier’s obligation at the supplier’s costLine-down situationsDefine the trigger and the cost recovery
Termination for causeTermination on notice after a defined failurePersistent failure, or a single critical failureDefine “persistent” numerically

Service credits and liquidated damages. A service credit is a liquidated damages clause by another name, and it is enforceable in China subject to the same adjustment risk: a court or tribunal may reduce an amount that is excessively higher than the actual loss, with thirty per cent above the proven loss as the practical reference point. Three consequences follow.

First, a credit that is a genuine pre-estimate of the administrative cost of a late shipment survives review better than a round number chosen to look severe. Second, the clause should state whether the credit is the exclusive remedy for the failure it addresses, or whether the buyer may also claim its additional loss. Exclusivity protects the supplier and lowers the price; it also caps the buyer’s recovery in the one case where the loss is catastrophic. Third, delay credits and quality credits should have separate caps, so that a bad month on one does not exhaust the ceiling for the other.

Corrective action as a contractual obligation. The corrective action request is the workhorse of supplier management in China, and it is usually operated informally. Making it contractual changes the dynamics: a defined acknowledgement period, a defined containment requirement, a defined root-cause deadline, and a defined consequence for a repeat failure. The eight-discipline format used across the automotive industry is widely understood by Chinese factories and provides a ready structure that both sides recognise.

Termination for persistent failure. Termination for cause should not depend on an argument about whether the failure was material. Define it numerically: performance below the minimum level for a stated number of consecutive measurement periods, or three repeat failures of the same corrective action within a stated period, or a single failure that stops the buyer’s line for more than a stated period. A defined trigger is easier to exercise and harder to challenge than a general materiality standard.

Termination for convenience. Where the buyer wants the ability to leave for commercial reasons, a termination for convenience clause with a notice period and a defined exit payment is cleaner than manufacturing a cause. It should be paired with transition obligations: continued supply during the notice period, transfer of tooling, release of design files and reasonable assistance to a successor.

Remedies that do not work. Three clauses appear often and deliver little. A right to “discuss” performance has no content. A penalty expressed as a percentage of annual spend, unrelated to any loss, will be reduced. And a termination right that requires the buyer to prove a material breach in an arbitration will never be used, because the cost of proving it exceeds the value of leaving.

Business reviews

Regular reviews are the mechanism that keeps the metrics alive. They also create the record that a later termination or claim will rely on, which is a reason to run them properly rather than a reason to run them formally.

Monthly operational reviews cover current orders, quality issues and delivery performance. They are short, thirty to sixty minutes, attended by the buyer’s sourcing contact and the factory’s account manager, and suited to active production programmes.

Quarterly business reviews cover all the KPIs, the root causes of major issues, and the forward view on forecasts, new projects and improvement plans. They are attended by management on both sides and suit strategic or key suppliers.

Annual strategic reviews cover long-term planning, the year’s performance against goals, market and technology trends, and are attended by senior leadership. They suit genuine partnerships.

The frequency should follow importance. Strategic suppliers warrant quarterly reviews plus monthly operational updates; important suppliers warrant a semi-annual review plus quarterly check-ins; commodity suppliers warrant an annual review with issues addressed as they arise. New suppliers should be met monthly for the first three to six months, because that is when the misunderstandings are formed.

A quarterly review agenda that works: the previous quarter’s KPIs against target; the major issues and their root causes; operational updates on orders, capacity and any change in equipment, personnel or process; the status of improvement projects; the forward view on forecast, new product pipeline and expected constraints; and the action list with owners and dates.

Six practices separate a review that changes behaviour from one that produces a slide deck. Send the data and the agenda in advance, so that the factory arrives prepared and cannot dispute the numbers in the room. Lead with data rather than opinion. Balance recognition with criticism, which matters more in China than in most markets. Focus on the fix rather than the fault. Record every action with an owner and a date. And follow up between reviews, because an action list that is never checked teaches the factory that the review is theatre.

One structural point is worth adding. Where the agreement makes the review mandatory — a quarterly meeting, a data pack due ten business days before it, minutes circulated within five business days and signed by both sides — the minutes become the contemporaneous record of what was agreed and what was promised. That record is what a tribunal reads when the buyer later says the supplier committed to a corrective action and did not perform it.

Working with the factory: culture, conduct and the record

Chinese factory relationships run on personal trust to a degree that surprises foreign buyers, and the cultural dimension is real. It also creates legal risk where the personal and the corporate are not separated.

Relationships and face. Business in China is built on relationships, people prefer to deal with those they know, and time invested in the relationship returns in cooperation. Face matters: public criticism causes loss of face and produces defensiveness rather than correction, and private conversations are more effective than confrontation in a meeting. Hierarchy is more pronounced than in Western firms, decisions are often made at the top, and seniority and age carry weight. Factories tend to take a long view, value repeat business and will sometimes invest in a relationship at thin margins.

Building the relationship. Visit at least once or twice a year for key suppliers, meet management as well as the day-to-day contact, and bring engineers so that engineer-to-engineer relationships form. Respond to emails promptly. Use formal channels for commitments and informal channels for speed, and keep the two distinct. Pay on time, provide forecasts, give reasonable lead times and leave the factory a workable margin. Give the factory opportunities to grow with you.

What not to do. Do not restrict contact to problems. Do not change requirements constantly. Do not push for price reductions every quarter. Do not make threats you will not carry out, because an unenforced threat teaches the factory that your deadlines are negotiable. Do not mistake directness for effectiveness; what reads as efficient in Frankfurt reads as rude in Dongguan.

Four legal points sit underneath the cultural ones.

WeChat is a record, and a weak one. Chinese factories negotiate on WeChat, and much of the operative communication happens there. Messages are admissible, but authenticity is challenged and the original device may be needed. Commitments that matter should be confirmed by email or in signed minutes, and the WeChat thread retained as corroboration rather than relied on as the primary record.

Authority is formal. An account manager’s assurance is not a variation of the contract unless the person has authority and the contract allows variation that way. Chinese companies generally act through the company seal or a signed contract, and a clause requiring variations to be in writing and signed by an authorised signatory is enforceable and useful. Ask for the signatory list at the start of the relationship.

Gifts and hospitality carry bribery risk. Commercial bribery is regulated under the Anti-Unfair Competition Law, and a factory that pays a “commission” to a buyer’s employee creates exposure for the buyer as well as for the individual. A written gift and hospitality policy, applied to factory visits and festivals, protects the relationship rather than damaging it, because it gives the local team a reason to decline.

Payment and invoicing are part of performance. Late payment damages a relationship faster than any quality dispute, and irregular invoicing creates tax exposure on both sides. The agreement should state the payment trigger, the invoice requirements and the consequence of a defective invoice, and the finance team should be part of the quarterly review rather than a separate conversation.

Compliance is a supply risk. Labour and environmental breaches at the factory are the factory’s legal problem and the buyer’s delivery problem. Working hours, overtime pay, social insurance contributions and the statutory cap on the proportion of dispatched workers are inspected, and a remediation order can stop a line for weeks. Environmental permits, emission limits and waste handling carry the same risk, with production suspensions available for serious violations. Neither topic belongs in the quarterly review as a moral question; both belong there as a continuity question, with the factory confirming its permits, its inspection record and any pending remediation.

Performance improvement and escalation

When performance falls short, the sequence matters more than the severity of the response.

Verify that the expectation was clear. Did the buyer communicate the standard, does the factory understand it, and is it reasonable? A surprising share of “poor performance” is a communication failure, and a supplier that did not know the requirement will not improve in response to pressure.

Raise it promptly. Issues that fester become normal, and a complaint about a practice that has continued for a year reads as a change of terms rather than a correction.

Use data. Show the trend, the defect data and the delivery record. Ask for the factory’s own analysis, because a supplier that has diagnosed its own problem is far more likely to fix it.

Ask for a corrective action plan. Root cause, containment, permanent action, verification, and a date for each. Reject plans that describe activity rather than outcome.

Set improvement targets. Specific, measurable, time-bound, and tied to the review calendar so that progress is visible.

Escalate in defined steps. An informal discussion, then a formal written concern, then a business impact conversation, then a reduction of business or a switch of supplier. The contract should support the ladder: a notice of breach that triggers a cure period, a right to reallocate volume, and a termination right that activates on the defined trigger.

Two habits improve the outcome. The first is to separate the commercial conversation from the contractual one. A letter that recites the breach, the cure period and the consequence is a legal document; it should not be softened into a suggestion, and it should not be the first time the factory hears of the problem. The second is to keep the record of improvement. A factory that fixes a problem deserves the record to show it, both because recognition is motivating and because a later dispute about the same issue should start from the corrective action that was closed.

Where improvement genuinely happens, the relationship is usually stronger than before, because the factory has learned that the buyer’s standards are real and that meeting them is rewarded with volume.

Managing a portfolio of suppliers

A buyer with more than a handful of suppliers needs a system, because attention is finite and the suppliers that get it are the ones that improve.

SegmentDescriptionManagement approach
StrategicHigh value, high risk, critical to the businessDeep partnership, quarterly reviews, joint planning, dedicated resource
KeyImportant but not criticalRegular reviews, performance management, development
TacticalMedium value, alternatives availableStandard procurement, periodic review, competitive pricing
CommodityLow value, easily replacedTransactional, price-driven, minimal management

Segmentation should drive the contract as well as the calendar. A strategic supplier’s agreement should carry the full metric set, the service credits, the audit rights, the capacity commitments and the exit terms. A commodity supplier’s agreement can be a purchase order and a set of standard terms, provided the standard terms contain the quality standard and the right to reject.

Supplier development. For suppliers with potential but weak systems, development beats replacement. Training in quality management and lean methods, shared best practice, engineers working alongside the factory’s engineers, and a structured improvement programme with milestones. The contract should support it with an audit right, an agreed improvement plan and a defined consequence if the milestones are missed.

Concentration risk. Strong relationships are good; dependence on a single supplier is not. Keep backup suppliers for critical products, maintain at least a small relationship with an alternative so that it stays qualified, avoid any supplier representing more than fifty to seventy per cent of production where possible, and test the market periodically to confirm that the pricing is fair. The common failure is a “second source” that exists on paper, has never run a production lot, and cannot be activated when the primary fails. Qualification is a continuing obligation, not a one-off event.

The cost of dual sourcing. Two sources mean two sets of tooling, two qualification cycles, two quality systems and, usually, a higher unit price at each. The decision should be made on the value of the products at risk rather than on a general preference for redundancy, and the agreement with each source should make clear what volume commitment, if any, the buyer is giving.

Practical checklist

  1. Choose five to seven metrics that match the product and the stage, and drop the rest from the dashboard.
  2. Write each metric into a schedule with a formula, a measurement point, a data source, a target, a minimum level and a reporting date.
  3. Name the party that produces the data, and give the buyer an audit right over it.
  4. Attach a consequence to each metric, escalating from reporting to corrective action to cost recovery to credit to reallocation to termination.
  5. Express credits as a formula tied to a real cost, keep delay and quality caps separate, and state whether the credit is the exclusive remedy.
  6. Define termination for persistent failure numerically, and pair it with transition obligations.
  7. Make the quarterly review contractual: data pack due date, minutes, signed action log and an escalation matrix with named roles.
  8. Record commitments in email or signed minutes rather than WeChat alone, and require variations to be signed by an authorised signatory.
  9. Adopt a written gift and hospitality policy before the next factory visit.
  10. Qualify a second source before it is needed, and keep it qualified with periodic small orders.

Supplier relationship management is one of the highest-return activities in manufacturing, and the return comes from the combination of goodwill and enforceability rather than from either alone. A factory that trusts the buyer will warn it about a problem early. A contract that measures the problem and attaches a consequence ensures that the warning is followed by a fix. The firm’s commercial contracts team drafts supplier agreements with the metric schedules and remedy architecture described here, and the manufacturing support team handles the audits, corrective action disputes and terminations that follow. Three related guides cover the adjacent ground: quality control in China hardware, supply chain risk management, and multi-supplier strategy.

Frequently Asked Questions↑ Back to article

What are the most important KPIs for tracking Chinese factory performance?

The most important KPIs fall into five categories: (1) Quality — first-pass yield, defect rate (PPM), incoming inspection pass rate, customer returns, (2) Delivery — on-time delivery rate, order fulfillment rate, lead time accuracy, (3) Cost — price competitiveness, cost reduction achievement, quote accuracy, (4) Responsiveness — response time to inquiries, issue resolution time, (5) Flexibility — ability to handle order changes, rush orders, new product introductions. Start with 5-7 key metrics that matter most to your business. Too many KPIs become overwhelming.

How often should I do business reviews with my Chinese suppliers?

The frequency depends on the importance of the relationship: (1) Strategic/key suppliers: quarterly business reviews (QBRs), plus monthly operational updates, (2) Important/secondary suppliers: semi-annual business reviews, plus quarterly check-ins, (3) Tactical/commodity suppliers: annual reviews, with issues addressed as they arise. For new suppliers, meet more frequently at first (monthly for the first 3-6 months) to build the relationship and address teething issues. Reviews are most effective when they happen regularly, not just when there is a problem.

How do I motivate a Chinese factory to improve performance?

Factories are motivated by a combination of: (1) Volume and growth — the promise of more business is the strongest motivator, (2) Fair pricing — paying a reasonable margin gives factories incentive to invest in quality, (3) Long-term commitment — multi-year agreements or forecasts give factories confidence to invest, (4) Recognition and relationship — Chinese business culture values respect and face; public recognition and good relationship matter, (5) Shared savings — if you propose cost reductions, share the savings (e.g., 50/50 split), (6) Clear expectations — factories want to know what is expected and how they are doing. The worst approach is only contacting the factory when something is wrong — positive reinforcement matters too.

What is the best way to handle poor supplier performance in China?

When a factory's performance is poor: (1) First, make sure your expectations are clear and communicated — sometimes the factory doesn't know what you expect, (2) Bring the issue up promptly and professionally — don't let problems fester, (3) Focus on facts and data, not blame — show them the numbers, ask for their analysis, (4) Ask for a corrective action plan — what will they do to improve, by when, (5) Set improvement targets and timelines — make it specific and measurable, (6) Follow up and track progress — hold them accountable to the plan, (7) Escalate gradually — verbal warning, formal warning, business impact discussion, then moving business if no improvement. Give factories a chance to improve — most want to do better, but they may need clear direction.

How do I build a strategic partnership with a Chinese factory?

To build a strategic partnership: (1) Be a good customer — pay on time, provide forecasts, treat people with respect, (2) Share information — give them visibility into your plans, volumes, and roadmap, (3) Commit volume — multi-year agreements or volume commitments show you are serious, (4) Invest in the relationship — visit regularly, meet their team, attend important events, (5) Involve them early — bring them into new product design for DFM input, (6) Grow together — as your business grows, help them grow with more orders and new opportunities, (7) Be fair — don't squeeze them on price every quarter; share both risks and rewards. Strategic partnerships take time to build — usually 1-2 years of consistent, positive interactions.

What quality KPIs should I track for my Chinese factory?

Key quality KPIs include first-pass yield (the percentage of products that pass inspection the first time without rework), defect rate measured in parts per million (PPM), incoming inspection pass rate from third-party inspections, customer return rate for defects found after delivery, and corrective action effectiveness (how well and how quickly the factory resolves quality issues). For most consumer products, a first-pass yield of 95%+ is good, and a defect rate under 1,000 PPM for major defects is a reasonable target. Track these over time rather than focusing on single batches, as trends reveal more about factory performance than individual data points.

What delivery KPIs matter most for Chinese manufacturing?

The most important delivery KPIs are on-time delivery rate (the percentage of orders shipped by the committed date), order fulfillment rate (the percentage of ordered quantity actually delivered), lead time accuracy (whether actual lead time matches quoted lead time), and responsiveness to rush orders or order changes. For most Chinese factories, an on-time delivery rate of 90%+ is a reasonable target for stable products, though new products or custom parts often have lower rates initially. Delivery performance tends to improve as the relationship matures and the factory gains experience with your product. Track delivery KPIs alongside quality KPIs because faster delivery should never come at the expense of quality.

Geason Law Offices AI Industry Research Group | Reviewed by He Qianggao (Lawyer License No. 14403201310092337, 10+ years of practice)