Quality Performance and PPM Disclosure Pages on Changzhou Manufacturer Websites: How Should They Be Accepted?

Direct answer: A quality-performance and PPM page cannot just say "defect rate below 50 PPM, customer satisfaction 98%." It should state how each metric is defined, what the numerator and denominator are, the period and sample scope, how it is split by customer/product/line, the basis for trends and comparisons, which system the data comes from, whether customer data is authorised, whether third-party citations have a source and scope, and how missed targets and anomalies are disclosed. Otherwise, when an auditor asks "is this PPM by piece or by defect, does it include rework, which customers, what period," the number does not stand.

This guide is for quality, production, sales, legal, supply-chain, and website teams at manufacturers in Changzhou and nearby industrial areas. It explains how to build and accept a quality-performance and PPM disclosure page. Information is current to 2026-11-02. It is a review of public standards, regulations, and project practice, not a quality audit, statistical, or legal opinion; specific definitions, data truthfulness, advertising-citation compliance, and customer requirements depend on the company's quality system, contracts, regulatory findings, and applicable law.

How does the quality-performance page relate to COA, complaint/8D, supplier-onboarding, and ESG pages?

The COA page answers "what was tested on this batch and what were the results"; the complaint/8D page answers "how complaints are received and corrected"; the supplier-onboarding page answers "how customers audit us"; the ESG page answers "company-level sustainability performance"; the quality-performance page answers "how quality performed over a period, on what basis, with what trend, compared with what." COA is batch evidence; quality performance is an aggregate. They cannot replace each other: a fully conforming batch does not mean a low overall PPM; a low overall PPM does not mean every batch is traceable.

The common break is: PPM is a single number with no denominator; "outgoing inspection defect rate" is presented as "process defect rate"; only large customers are counted and packaged as "overall"; reworked-and-passed batches are counted as first-pass; year-on-year is used to hide a basis change; customer case data is published without authorisation; third-party reports are cited without source or year. A page that states definition, basis, sample, source, and boundary gives a customer audit something to check.

This page map is editorial; quality objectives and monitoring are in GB/T 19001-2016, clauses 6.2 and 9.1.

How should PPM, FPY, and OTD be defined?

Standardise the terms first: PPM usually means "defects per million," but the denominator can be pieces, batches, orders, or defect opportunities; FPY (first-pass yield) is "units conforming first time ÷ units input"; OTD (on-time delivery) is "orders delivered on time ÷ total orders"; complaint rate is "complained batches/orders ÷ total batches/orders"; return rate is "returned quantity/value ÷ shipped quantity/value." Each metric needs numerator, denominator, period, and data source, or same-named metrics are not comparable.

The most common confusion is PPM versus DPMO: PPM is per piece, while a product with multiple defect opportunities is better measured by DPMO (defects per million opportunities); a batch pass rate cannot be converted directly into PPM. The page should provide a "metric definition table" with formula, unit, period, data source, and owner; where products/customers use different bases, list them separately rather than blending.

These definitions are editorial; specific formulas, periods, and data sources follow the quality system and customer requirements.

How should numerator, denominator, and period be set?

The denominator decides the meaning: by piece for single-unit products; by batch for batch release; by order for delivery performance; by defect opportunity for multi-process/multi-characteristic products. Fix the period (monthly/quarterly/annual/rolling 12 months) with explicit start and end dates; define "nonconforming" (customer return, process scrap, outgoing hold, upheld complaint) and do not mix severity levels. When the period changes, year-on-year and month-on-month become distorted.

"PPM fell" invites the questions: did the denominator change? the period? the nonconformity definition? the product mix? the customer mix? If the denominator switches from pieces to batches, or the period from 12 months to 3, the number can swing without any real quality improvement. The page should fix the basis and note changes; to compare, recalculate history on the same basis or state "basis changed, not directly comparable."

These requirements are editorial; quality-objective monitoring is in GB/T 19001-2016, clause 9.1.

How should sample and representativeness be explained?

State which customers, products, lines, plants, and periods the sample covers: full population or sampling? random or convenience? how large? Small samples (e.g., fewer than 30 batches/1,000 pieces) should be marked "limited sample, indicative only," not packaged as the overall level. If only large customers or mature products are counted, state the scope—do not call it "company-wide." Changes in customer mix (new customer ramp-up, incumbent growth) also affect PPM and need explanation in the trend narrative.

"We supply one large customer at low PPM" does not mean "all customers are low": product complexity, acceptance standards, and return policies differ, so PPM is not directly comparable across customers. The page can list by customer/product or explain weighted/simple averaging; where customer names appear, obtain authorisation or anonymise. Sample representativeness and data authorisation are the two most overlooked and most challenged points on a quality-performance page.

This explanation is editorial; specific sampling and representativeness follow the quality system and statistical method.

How should quality objectives be set and assessed?

Quality objectives should be measurable, achievable, time-bound, and aligned with customer, regulatory, and product requirements: for example PPM, FPY, OTD, complaint closure rate, and return rate, each with baseline, target, basis, owner, and assessment cycle. Lower is not always better: a PPM target must fit product complexity, customer requirements, and cost; an OTD target must consider capacity and supply chain. Decompose targets to workshops/teams/product lines, assess periodically and review; record any basis change.

The common problem is "targets every year, good data every year": no baseline, arbitrary achievement rates, targets only at company level, result-only assessment leading to data beautification, and targets changed three times a year distorting year-on-year. Put objectives in the quality-system documents and link them to monitoring and management review; when publishing, state "target + actual + period + basis," not an unsupported "exceeded target."

Objective management is editorial; quality objectives and planning are in GB/T 19001-2016, clause 6.2.

How can trends and comparisons avoid distortion?

Trends need a fixed basis, fixed period, fixed scope, and the number of data points; comparisons must state the benchmark (own history, target, customer requirement, public industry data) and must not compare different bases directly. Year-on-year/month-on-month needs the base period; where basis, product mix, or customer mix changed, explain it below the chart. Do not show a "continuously falling" curve without axes, sample size, and anomaly explanations.

"Industry average" or "benchmark level" needs a source and year, not an impression; citing third-party data requires source, scope, sample, and time to meet advertising-citation requirements. Charts can be styled but must not truncate axes, hide anomalies, or use inconsistent scales. The goal is to let customers judge stability, not to make the curve look good.

These practices are editorial; third-party citation requirements are in Article 11 of the Advertising Law.

Where does the data come from, and how should workpapers be kept?

Quality-performance data usually comes from QMS/MES/ERP/complaint systems/inspection records; state the system, extraction date, extractor, formula, and filters. Workpapers must allow recalculation: raw data, cleaning rules, exclusions, summary tables, charts. Before publication, the quality lead signs off; where the external basis differs from internal reports, explain the difference. A "PPM 50" without a workpaper is a slogan.

The common problem is "multiple data sets": sales, quality, and customer-audit numbers disagree; departments calculate differently; spreadsheets are edited repeatedly with no version control. Build one "external quality-data basis sheet," maintained by quality, used by sales, reviewed by legal/marketing; keep version and date on each update. When an auditor requests source records, trace from batch/order to inspection and complaint records.

Workpaper requirements are editorial; quality-record control is in GB/T 19001-2016, clause 7.5.

What advertising and citation rules apply to publishing quality data?

Quality data on a website is commercial promotion and may constitute advertising. Article 11 of the Advertising Law requires that data, statistics, survey results, abstracts, quotations, and other cited content be true, accurate, and sourced; where there is a scope of application and validity period, it must be clearly stated. The SAMR's 2026 Guidance on Enforcement of Advertising Citation Content distinguishes "cited content" (data generated by a party other than the advertiser) from "self-verified content" (data obtained by the advertiser's own experiments/measurements/statistics); in both cases the advertiser is responsible for truthfulness, accuracy, and legality and bears the burden of proof.

"Industry lowest PPM," "zero complaints," and "100% satisfaction" are high-risk: without source, scope, or sample, they may be found false or misleading. The Guidance also provides that cited content without a source, without clear scope/validity, or inconsistent with the materials provided may not be published or will be handled under the law. Use a "data + source + scope + period + sample size" template, and keep authorisation and the original report for third-party data.

Citation requirements are in Article 11 of the Advertising Law; details are in the Guidance on Enforcement of Advertising Citation Content (SAMR Announcement No. 20 of 2026).

How should missed targets, anomalies, and complaints be disclosed?

A quality-performance page need not report only good news: it can state the reasons for missing the target, corrective actions taken, and expected recovery, plus 8D closure for major complaints (redacted). That is more credible than "always on target" and matches the logic of continual improvement. Do not publish customer names, batch numbers, claim amounts, or pending disputes; for recalls/safety incidents, disclose as required by the regulator, not as "handled internally."

Anomaly disclosure needs boundaries: "one product line saw a short-term fluctuation during changeover and recovered through tightened inspection and process adjustment" is enough; specific customers and defect details are not required. "Complaint closure rate 95%" is fine if the denominator and period are stated. Auditors look at whether anomalies are identified, corrected, verified, and generalised. Reporting only good data makes customers doubt completeness.

These boundaries are editorial; recall and incident disclosure follow regulator requirements and company policy.

How should certifications, audits, and customer evaluations be used?

An ISO 9001 certificate can be stated as "certified," with the certification body, certificate number, scope, and validity, and with attention to status (valid/suspended/withdrawn); do not use an expired certificate or turn the "certification scope" into "all products." Customer audit reports, awards, and supplier ratings are customer data and cannot be published without authorisation; anonymised wording such as "an A-rated supplier to a global Tier-1 customer in 2025" is acceptable if consent or anonymisation is documented.

"Customer satisfaction 98%" needs the survey method, sample size, response rate, and scoring basis; without them, 98% is just a number. Third-party audit/certification reports may be summarised, but not quoted out of context; nonconformities that have been closed can be described as "rectified" without details. Certifications and evaluations are external evidence for quality performance, not decoration replacing internal data.

These uses are editorial; certificate status is verified with the certification body, and customer data requires consent.

What should the page say, and what must not be published?

A public page should include: a metric definition table, basis and period, sample scope, trend chart with sample size, targets and achievement, anomalies and corrections, certification/audit summaries, data-source notes, customer-data authorisation notes, and a contact route. Do not publish: customer names and batch numbers (without consent), specific claims/litigation, unclosed major quality incidents, internal process parameters, supplier lists, or full customer audit reports. Redacted cases and a basis note are fine; raw detail is not.

The page exists to let customers judge quality stability, not to prove "zero defects." Wording should be restrained: "counted on basis XX," "sample covers XX," "data as of XX"—not "industry-lowest PPM," "zero complaints," or "100% conforming." Once published, quality data must withstand customer audits and regulatory checks; quality and legal/marketing should sign off before release.

These modules are editorial; public scope, customer data, and wording follow confidentiality, legal, and marketing advice.

Go-live acceptance checklist

Ten checks: metric definitions, numerator/denominator, period, sample scope, trend and comparison, data workpapers, citation compliance, customer authorisation, anomaly disclosure, and public boundaries plus links. Each needs an owner, evidence, and review date. The pass condition is not "the page shows PPM," but that a customer can recalculate, see the sample and period, and find the source; internally, the workpaper exists.

Before go-live, run one real metric through: extract → clean → calculate → review → chart → label scope/period/sample → legal/marketing review → publish → retain workpaper. Any step that fails is a gap in the page or the data.

This checklist is editorial; specific basis, data truthfulness, advertising compliance, and disclosure follow the company system, contract, and regulator requirements.

Who should read this, who should not, and what decisions does it support?

It suits manufacturers in Changzhou and nearby areas with customer quality audits, supplier ratings, tenders, or export business—especially auto parts, electronics, equipment, medical devices, and precision machining; it also suits buyers assessing supplier quality stability. It does not suit businesses with no quality data, pure trading companies, or teams wanting one page to replace quality-system/legal advice.

Read it if you are in quality, production, sales, legal, supply chain, or a website vendor; you are building a quality-performance page, preparing for a customer audit, publishing PPM/FPY/OTD, or citing third-party data; or you need to answer "how is PPM calculated, what is the denominator, is the sample enough, how to compare trends, can the data be public, what citation is needed." Skip it if you treat batch conformance as the overall level, cherry-pick customer data, or want a page claiming "industry-lowest/zero complaints." It supports self-review of definitions, basis, sample, trends, workpapers, citations, consent, and disclosure.

Lingyi Information is a website development and enterprise AI project provider, not a quality certification body, statistical body, or law firm. This article is not a substitute for quality-audit, statistical, or legal advice. Changzhou projects are delivered remotely or on site as confirmed per project; Lingyi is one candidate, does not claim a local office or team, and does not promise audit outcomes, data results, conversion, or inquiries.

FAQ

How is PPM calculated, and what is the denominator?

PPM = defects ÷ total × 1,000,000, but "total" can be pieces, batches, orders, or defect opportunities and must be stated. By piece suits single-unit products; by batch suits batch release; by opportunity is better for multi-process products (then it is DPMO). A PPM without a denominator cannot be compared.

Can PPM be averaged across customers and products?

Simple averaging is not advisable. Acceptance standards, product complexity, and batch sizes differ, so averaging hides differences; list separately or explain the weighting. If only some customers are counted, state the scope—do not call it "company-wide."

Can the website say "PPM below 50"?

Yes, but state the basis (pieces/batches/orders), period, coverage, sample size, data source, and cut-off date; otherwise it may be treated as unsourced or misleading. Quality data is commercial promotion subject to Article 11 of the Advertising Law: true, accurate, sourced, with scope and validity stated where applicable.

Do customer case data need consent?

Yes. Customer names, batches, claims, and audit reports are customer data and should not be published without consent; anonymise (e.g., "a global Tier-1 customer") and keep the anonymisation basis. Customer evaluations and awards also need a confirmed public scope.

Should we publish bad numbers?

Not everything need be public, but do not report only good news. State the reasons for missed targets, actions, and recovery, plus 8D closure for major complaints (redacted). Reporting only good data makes customers doubt completeness; recalls and incidents are disclosed per regulator requirements.

What should be stated when citing third-party data?

State the issuer, report name, year, statistical scope, sample, and time; state the scope of application and validity period where applicable. The 2026 Guidance on Enforcement of Advertising Citation Content requires cited content to be true, accurate, and lawful; the advertiser bears responsibility and the burden of proof, and must retain the original materials after publication.

One-sentence takeaway

A Changzhou manufacturer's quality-performance and PPM page should not sell "good numbers." It should turn metric definitions, numerator/denominator, period, sample scope, trends and comparisons, data workpapers, citation sources, customer consent, anomaly disclosure, and public boundaries into a verifiable checklist—so customers can recalculate and find the source.

Method and limitations

This is a review of public standards, regulations, and project practice, not a certification or legal opinion on any company's quality data, statistical basis, advertising compliance, or customer audit. No undisclosed customer quality data, audit reports, or claim information were used. Metric definitions, sample notes, checklists, and the acceptance checklist are editorial suggestions, not quality, statistical, or legal conclusions. Specific basis, data truthfulness, advertising citation, and disclosure depend on the quality system, contract, regulatory findings, and applicable law.

Sources and scope

Information is current to 2026-11-02. The following are public official/primary pages; their scope, version, and wording control.

1. GB/T 19001-2016, Quality management systems—Requirements: National Standards Full-Text Public System. Quality objectives and planning (6.2), monitoring, measurement, analysis and evaluation (9.1), control of documented information (7.5), control of nonconforming outputs (8.7), improvement (10).
2. Advertising Law of the PRC (2021 amendment), Article 11: MOFCOM Global Law Database (legal text). Advertising that uses data, statistics, survey results, abstracts, quotations, and other cited content must be true, accurate, and sourced; where there is a scope of application and validity period, it must be clearly stated.
3. Guidance on Enforcement of Advertising Citation Content (SAMR Announcement No. 20 of 2026, issued and effective 12 June 2026): MOFCOM Global Law Database (announcement text). Definitions of cited content and self-verified content, truthfulness/accuracy/legality, advertiser responsibility and burden of proof, source and scope/validity labelling, and handling of unsourced or inconsistent content.