Home/Resources/ Factory audit checklist

Free checklist / Excel + CSV / No signup

Electronics and NPI factory audit checklist

102 checks across 14 areas, from the quality system and EVT/DVT/PVT discipline to SMT, test coverage, traceability and IP protection. Each check says what evidence to ask for and why it matters.

Download checklist · Excel Checklist · CSV ↓Use it on this page

Before you start

How to run a factory audit

34 checks are marked critical: a Fail there usually means the factory should not move to tooling or production for your product until it is resolved. Adjust the flags to your product and its risks.

Set the scope
Name the product, the processes it needs and the exact site that will build it. Mark sections that do not apply as N/A rather than deleting them, so the record shows they were considered.
Send the evidence list ahead
Share the “evidence to request” column before the audit. What the factory sends, and how quickly, is the first finding. A desk review of those documents tells you what to look for on site.
Walk the product, not the showroom
Follow one product from receiving through SMT, assembly, test and packing. Compare what is on the floor with what the documents say, at the line that will actually build your product.
Choose your own samples
Pick a shipped serial number, a component lot or a recent corrective action yourself and ask for its records. Records chosen by the factory show its best work, not its normal work.
Record evidence, not impressions
Note the document, revision, date or observation behind each status. A Pass without a reference is an opinion.
Close the loop
Give every Concern and Fail an owner and a date. Agree which critical items must be closed before tooling, a pilot build or a purchase order, and check them again.

Desk audit or on-site audit?

Use both, in that order. A desk audit filters candidates and focuses the visit; an on-site audit verifies that what was described is what happens on the line.

What a desk audit and an on-site audit can each establish
Area	Desk audit	On-site audit
Certificates and procedures	Confirms they exist, their scope and expiry dates	Confirms they are used at the line that will build your product
Process controls	Shows how controls are described	Shows whether they are followed, at which stations, by whom
Records and traceability	Uses the samples the factory chooses to send	Lets you choose a serial number or lot and watch it being traced
People and engineering	Names on an organization chart	Meeting the engineers who would work on your project
Capacity	Stated figures	Lines, utilization and shared equipment you can see
Effort	Remote; can be repeated as documents arrive	Travel and time on site, usually after a desk review has narrowed the questions

The checklist

Mark each check as you go. Your marks and notes stay in this browser; nothing is sent anywhere. Print the page or download your audit as a CSV when you are done.

PassEvidence seen and adequate for your product.
ConcernPartly in place or evidence incomplete. Needs follow-up, but not a blocker on its own.
FailMissing or inadequate. On a critical check, resolve it before committing.
N/ADoes not apply to this product or site. Note why.

01 · 8 checks

Quality management system

Whether the quality system is real, covers the right site and is used day to day.

QMS-01Does a current quality management system certificate cover the site and processes that will build your product?Critical
Evidence to request
Certificate (for example ISO 9001) showing the certification body, site address, scope statement and expiry date, plus the date of the latest surveillance audit.
Why it matters
A certificate for a sister company, another address or a narrower scope says nothing about the line that will build your product.
QMS-02Does document control make sure the line works to the current drawing and instruction revision?Critical
Evidence to request
Document control procedure and master list; pick a work instruction at a station and compare its revision with the master list.
Why it matters
Out-of-date drawings or instructions at the workstation are a common source of build errors, especially after an engineering change.
QMS-03Are nonconforming parts and products identified, segregated and dispositioned under a documented process?Critical
Evidence to request
Nonconforming material procedure, the physical hold or quarantine area, and recent material review (MRB) disposition records.
Why it matters
Without physical segregation and a recorded decision, rejected material tends to find its way back into production.
QMS-04Does quality report independently of production, with authority to stop a line or hold a shipment?
Evidence to request
Organization chart showing the quality reporting line and a written statement of who can stop shipment; ask for a recent example of a hold.
Why it matters
If production can overrule quality, nonconforming product ships whenever the schedule is under pressure.
QMS-05Are internal audits carried out to plan and are their findings closed?
Evidence to request
Internal audit schedule for the last 12 months, two recent audit reports and the closure status of their findings.
Why it matters
Audits that run on schedule and close findings show the system is used, not just certified.
QMS-06Does management review quality data such as yield, complaints and returns, and act on it?
Evidence to request
Minutes or slides from the latest management review, with quality indicators and the actions agreed.
Why it matters
It shows whether leadership sees quality trends and assigns owners, or only reacts to escalations.
QMS-07How are customer-specific requirements recorded and flowed down to the line?
Evidence to request
The mechanism for capturing customer requirements (special characteristics, packaging rules, labeling) and a redacted example from an existing customer.
Why it matters
Requirements that live only in an email thread are lost when people or shifts change.
QMS-08If your product is regulated, does the site hold the relevant sector certification, such as ISO 13485 for medical devices or IATF 16949 for automotive?Critical
Evidence to request
Certificate with scope and expiry, and records showing how the sector requirements are applied on the line that will build your product.
Why it matters
Regulated products need sector-specific controls; a general quality certificate may not meet your regulatory obligations. Mark N/A if your product is not regulated.

02 · 8 checks

Engineering and NPI process

How the factory takes a design through EVT, DVT and PVT, and who does the engineering.

NPI-01Does the factory run a defined NPI process with stage gates (for example EVT, DVT and PVT) and written exit criteria?Critical
Evidence to request
NPI procedure, a gate checklist and a redacted example of a completed gate review with open issues and owners.
Why it matters
Gates with exit criteria stop a design moving into tooling or production while known issues are still open.
NPI-02Who are the named engineers for your project, and which disciplines are in house?Critical
Evidence to request
Names and roles of the project, mechanical, electrical, test and quality engineers, and which engineering work is subcontracted.
Why it matters
Capability claims often describe group resources that will not be available to your project.
NPI-03Does the factory give written DFM and DFA feedback before tooling is released?
Evidence to request
A redacted DFM report showing the issues raised, the changes proposed and how each was closed.
Why it matters
Design issues found before steel is cut are far cheaper to fix than mold changes after the first trial (T1).
NPI-04Is there a formal engineering change process (ECN/ECO) that requires your approval?Critical
Evidence to request
Change procedure and a recent change record showing the impact assessment, customer approval and effective date or serial number.
Why it matters
Uncontrolled changes, including component substitutions, are a frequent cause of field problems in electronics.
NPI-05Are golden samples and first article inspection (FAI) controlled?
Evidence to request
How golden samples are labeled, signed, stored and retired; an FAI report with measured values against the drawing.
Why it matters
A signed reference sample and a measured first article settle disputes about what acceptable means.
NPI-06Are pilot builds measured and turned into an action list before ramp?
Evidence to request
A pilot run report template or example with yield by station, defect Pareto, open issues, owners and dates.
Why it matters
A pilot that is not measured cannot tell you whether the line is ready for volume.
NPI-07Are process FMEAs or equivalent risk reviews done for new products?
Evidence to request
A process FMEA or risk register for a comparable product, with actions tied to the highest-risk steps.
Why it matters
It shows whether the factory anticipates failure modes or waits for them to appear on the line.
NPI-08How are test specifications and fixtures developed and validated before production?
Evidence to request
Fixture development plan, a test specification example and validation records such as known-good and known-bad sample checks.
Why it matters
Unvalidated fixtures produce false passes and false failures, and both cost money.

03 · 7 checks

Incoming quality (IQC)

How components and parts are checked, stored and sourced before they reach the line.

IQC-01Is there an incoming inspection plan for each critical component, with defined sampling levels?Critical
Evidence to request
IQC procedure, the sampling plan (for example AQL levels by part category) and inspection records for a recent lot.
Why it matters
A defective part caught at receiving costs a fraction of the same failure found after assembly.
IQC-02Where are electronic components bought, and how is counterfeit risk managed?Critical
Evidence to request
Approved distributor list, policy on independent brokers, certificates of conformance and the checks applied to broker-sourced parts.
Why it matters
Counterfeit or remarked components can pass basic tests and fail later in the field.
IQC-03Are moisture-sensitive devices controlled according to their moisture sensitivity level (MSL)?Critical
Evidence to request
MSD handling procedure (for example based on IPC/JEDEC J-STD-033), dry cabinets or bags, floor-life tracking and bake records.
Why it matters
Moisture absorbed by component packages can cause cracking or delamination during reflow.
IQC-04Are incoming lots linked to the supplier lot and date code?
Evidence to request
Receiving labels and system records that connect supplier lot and date code to the internal lot number.
Why it matters
Without lot linkage you cannot limit a containment when a component issue appears.
IQC-05Is there an approved vendor list (AVL), and does purchasing enforce it?
Evidence to request
AVL for your product type and the control that stops purchasing from buying an unapproved source.
Why it matters
Unapproved alternates are a common source of silent product changes.
IQC-06How are rejected incoming lots handled and fed back to sub-suppliers?
Evidence to request
Supplier corrective action requests (SCARs) issued in the last 12 months and their closure status.
Why it matters
It shows whether the factory manages its own supply base or simply absorbs defects.
IQC-07Are housings, plastics, metal parts and finishes inspected against drawings and cosmetic standards?
Evidence to request
Dimensional inspection records against drawings, written cosmetic criteria and signed limit samples.
Why it matters
Plastics and finishes vary from lot to lot; without limit samples, cosmetic acceptance becomes an argument.

04 · 9 checks

SMT and PCBA

Printing, placement, reflow and inspection on the surface-mount line.

SMT-01Is solder paste stored, thawed and used within controlled time limits?
Evidence to request
Paste storage temperature logs, thaw and open-time labels on jars or cartridges, and expiry tracking.
Why it matters
Paste used outside its working life causes poor wetting, voids and weak joints.
SMT-02Is solder paste inspection (SPI) used, and is there a reaction plan when results drift?
Evidence to request
SPI equipment on the line, recent SPI data and the documented reaction to an out-of-limit measurement.
Why it matters
Many SMT defects start at printing; SPI catches them before parts are placed.
SMT-03Is feeder loading verified so the wrong reel cannot be placed?Critical
Evidence to request
Feeder verification system (for example barcode scanning at loading), splice controls and first-piece check records.
Why it matters
A wrong reel on one feeder can build an entire lot with the wrong component.
SMT-04Is a reflow profile defined for each product and verified with a profiler?Critical
Evidence to request
Profile specification by product, a recent profiler measurement on a board with thermocouples, and the verification frequency.
Why it matters
An unverified profile can leave cold joints or overheat temperature-sensitive components.
SMT-05Is automated optical inspection (AOI) used after reflow, with programs maintained for each product?
Evidence to request
The AOI program for a current product, its false-call rate and records of how escapes are reviewed.
Why it matters
A poorly tuned AOI program is either ignored by operators or lets defects through.
SMT-06If your board has BGA, QFN or other hidden joints, is X-ray inspection available?
Evidence to request
X-ray equipment or a subcontract arrangement, the sampling plan and the void acceptance criteria.
Why it matters
Joints under BGA and bottom-terminated packages cannot be inspected optically.
SMT-07Are stencils managed for design, cleaning, tension and life?
Evidence to request
Stencil register, cleaning schedule and tension check records.
Why it matters
Worn or dirty stencils degrade print quality slowly, which makes the cause hard to spot.
SMT-08Is a workmanship standard agreed (for example IPC-A-610 and its class), and are inspectors and rework operators trained to it?Critical
Evidence to request
The agreed standard and class, training or certification records for inspectors and rework operators, and the rework procedure.
Why it matters
Without an agreed standard, an acceptable solder joint means different things to you and to the factory.
SMT-09If your product needs conformal coating, underfill or potting, are those processes controlled?
Evidence to request
Process instructions, the coverage inspection method (for example UV inspection) and cure records.
Why it matters
Coverage and cure problems are invisible without a defined inspection method.

05 · 10 checks

Assembly and process control

Final assembly, in-process checks, packaging and outgoing inspection.

ASM-01Are work instructions at each station current, visual and in the operators' language?
Evidence to request
Walk the line and compare the instruction at each station with the master revision.
Why it matters
Operators follow what is in front of them, not what is in the document system.
ASM-02Are critical process parameters such as screw torque, welding, dispensing and sealing defined and monitored?Critical
Evidence to request
Process parameter sheets, monitoring records and calibration labels on the tools used.
Why it matters
Water resistance, drop performance and safety often depend on a handful of parameters.
ASM-03Is a first-piece check done and recorded at line start and after every changeover?
Evidence to request
First-piece records for recent shifts and changeovers, signed by quality.
Why it matters
It catches setup errors before a whole shift is built.
ASM-04Does in-process quality control (IPQC) patrol the line with defined checks and frequencies?
Evidence to request
IPQC check sheets, the patrol frequency and records of issues escalated.
Why it matters
Early detection limits how much product needs rework or sorting.
ASM-05Is process capability measured for key characteristics, for example with Cpk or control charts?
Evidence to request
Recent capability studies or SPC charts for key dimensions or process parameters.
Why it matters
It shows whether a process is stable and centered, not just inside its limits today.
ASM-06If the product has lithium cells, are battery handling, assembly and storage controlled?Critical
Evidence to request
Incoming cell checks, storage conditions, insulation and welding controls, fire precautions, and who provides the UN 38.3 test documentation.
Why it matters
Battery defects create safety risks and shipping restrictions. Mark N/A if there is no battery.
ASM-07Is mistake-proofing (poka-yoke) used at error-prone steps?
Evidence to request
Examples of fixtures, sensors or interlocks that prevent wrong orientation, missing parts or skipped steps.
Why it matters
Inspection finds errors after they happen; mistake-proofing prevents them.
ASM-08Is rework controlled, recorded against serial numbers and followed by re-test?
Evidence to request
Rework procedure, rework logs linked to serial numbers and re-test records.
Why it matters
Uncontrolled rework can hide damage and break traceability.
ASM-09Is final packing verified for contents, labels, regulatory markings and language?
Evidence to request
Packing checklist, label verification method and packing audit records.
Why it matters
Missing accessories or wrong labels are among the failures end customers notice first.
ASM-10Is outgoing quality control (OQC) performed to an agreed sampling plan before shipment?Critical
Evidence to request
OQC procedure, agreed AQL levels by defect class, recent OQC reports and lot rejection records.
Why it matters
Outgoing inspection is the last check under the factory's control before goods leave.

06 · 8 checks

Test (ICT/FCT and burn-in)

What is tested, how well the tests cover the design, and what happens to the data.

TST-01Is board-level test coverage (ICT or flying probe) defined, and are the untested areas known?Critical
Evidence to request
A coverage report listing tested and untested nets and components for a comparable board.
Why it matters
Defects escape where coverage is missing; you should know where those gaps are.
TST-02Does functional test (FCT) check every function the user relies on, including radio, sensors, charging and firmware version?Critical
Evidence to request
FCT specification mapped to product requirements, and version control of the test station software.
Why it matters
A unit can pass in-circuit test and still fail at what the user actually does with it.
TST-03Is firmware programmed and verified under version control?Critical
Evidence to request
Firmware release process, checksum or version verification at programming, and change records.
Why it matters
Shipping the wrong firmware version is easy to do and expensive to fix in the field.
TST-04Are test results stored per serial number and retrievable on request?
Evidence to request
Database or logs linking serial numbers to test results, and the retention period.
Why it matters
Per-unit test data is your evidence in any field-failure investigation.
TST-05Are test failures analyzed and reported as first-pass yield and a failure Pareto?
Evidence to request
First-pass yield by station and failure analysis records for a recent month.
Why it matters
First-pass yield shows process health better than final yield, which hides rework.
TST-06Are fixtures maintained and checked with known-good and known-bad samples?
Evidence to request
Fixture maintenance logs, daily verification records and the replacement schedule for wearing parts such as pogo pins.
Why it matters
Worn fixtures create false failures or, worse, false passes.
TST-07Where reliability risk justifies it, is there a defined burn-in or run-in test?
Evidence to request
Burn-in conditions (duration, temperature, load), sample size and failure records.
Why it matters
Burn-in screens early-life failures; whether it is worth the cost depends on the product and its risk.
TST-08Can reliability and environmental tests such as drop, temperature, humidity and vibration be run in house or through a named lab?
Evidence to request
Equipment list or third-party lab arrangement, and a sample test report.
Why it matters
DVT reliability testing needs equipment and discipline; know in advance who will run it and who pays.

07 · 6 checks

ESD and handling

Electrostatic discharge control and the physical handling of parts and finished goods.

ESD-01Is there a documented ESD control program, for example based on ANSI/ESD S20.20 or IEC 61340-5-1?Critical
Evidence to request
ESD program document, the named coordinator and recent ESD audit records.
Why it matters
ESD damage can be latent: a unit passes test and fails later in the field.
ESD-02Are wrist straps, footwear and workstation grounding tested and recorded daily?
Evidence to request
Daily wrist strap and footwear test logs and ground check records.
Why it matters
Controls that are not tested drift out of specification without anyone noticing.
ESD-03Are ESD protected areas (EPAs) marked and controlled, including for visitors?
Evidence to request
EPA signage and boundaries, entry checks, and smocks or heel straps for visitors.
Why it matters
An unmarked boundary means nobody knows where the rules apply.
ESD-04Are sensitive parts and PCBAs stored and moved in ESD-safe packaging?
Evidence to request
Shielding bags, conductive totes and racks; walk the work-in-progress storage areas.
Why it matters
Boards are often well protected at the station and exposed between stations.
ESD-05Are humidity and ionizers monitored where the process needs them?
Evidence to request
Humidity logs for production areas and ionizer balance check records.
Why it matters
Low humidity increases static charge generation.
ESD-06Are cosmetic surfaces protected from scratches and contamination during handling?
Evidence to request
Glove use, protective films, trays and handling instructions for finished surfaces.
Why it matters
Cosmetic rejects often come from handling between stations rather than from the process itself.

08 · 6 checks

Traceability and records

Whether a shipped unit can be traced back to its parts, process and test results.

TRC-01Can the factory trace a finished serial number to its component lots, PCBA, test results and build date?Critical
Evidence to request
A live demonstration: choose a shipped serial number and ask for its records.
Why it matters
Traceability determines how small a containment or recall can be.
TRC-02Does the factory's system enforce the process route so a unit cannot skip a test station?
Evidence to request
MES or equivalent system screens showing route control and what happens when a step is missing.
Why it matters
Route control stops untested units from reaching packing.
TRC-03Are serial numbers, MAC addresses and device IDs allocated and checked for duplicates?
Evidence to request
Allocation process, duplicate checks and label printing controls.
Why it matters
Duplicate identifiers cause field failures that are hard to diagnose.
TRC-04Are shipments linked to serial number ranges and lots?
Evidence to request
Packing lists and carton labels showing serial ranges or lot numbers.
Why it matters
You need to know which customers received which units.
TRC-05How long are production, test and inspection records kept?
Evidence to request
Record retention policy, compared with your own warranty and regulatory requirements.
Why it matters
Records may be needed years after shipment.
TRC-06Is it agreed which build and quality data you can see, in what format and how quickly?
Evidence to request
Written agreement or proposal on data access, reporting format and turnaround time.
Why it matters
Access to data needs to be agreed before an issue, not negotiated during one.

09 · 6 checks

Calibration and equipment

Measurement, test and production equipment that the quality results depend on.

CAL-01Is all measuring and test equipment on a calibration schedule with traceable certificates?Critical
Evidence to request
Calibration register, certificates traceable to national standards, and in-date labels on equipment at the line.
Why it matters
Measurements from uncalibrated equipment cannot support an acceptance decision.
CAL-02What happens when equipment is found out of calibration?
Evidence to request
Procedure for assessing product measured since the last good calibration, with an example.
Why it matters
Out-of-tolerance equipment may have accepted bad product; there should be a look-back.
CAL-03Is there preventive maintenance for production equipment such as SMT lines, reflow ovens and molding machines?
Evidence to request
Maintenance schedule, completion records and downtime logs.
Why it matters
Unplanned downtime turns into missed ship dates and rushed work.
CAL-04Have measurement systems for key characteristics been validated, for example with gauge R&R studies?
Evidence to request
Gauge R&R or measurement system analysis results for key measurements.
Why it matters
A measurement system with too much variation cannot tell good parts from bad.
CAL-05Are torque drivers, dispensers and similar tools verified at defined intervals?
Evidence to request
Daily or weekly verification records at the stations.
Why it matters
These tools drift with use, and the parameters they control are often critical.
CAL-06Is the equipment capable of your design, for example the smallest component package and finest pitch on your board?
Evidence to request
Equipment list with models and stated capability, and examples of comparable boards built.
Why it matters
Very small packages and fine pitch need equipment that can place and inspect them reliably.

10 · 6 checks

Corrective action (8D and CAPA)

How problems are contained, solved at the root and prevented from returning.

CAR-01Is there a structured corrective action process, such as 8D, that requires root cause analysis?Critical
Evidence to request
Two recent, redacted 8D reports from customer complaints, including the root cause and the verification of effectiveness.
Why it matters
Without root cause analysis, the same defect comes back under a different name.
CAR-02How fast does the factory contain an issue: stop shipment, sort stock and notify the customer?Critical
Evidence to request
Containment procedure, committed response times and a recent example.
Why it matters
Slow containment multiplies the number of affected units.
CAR-03Are corrective actions verified as effective before they are closed?
Evidence to request
Closure records with data showing the issue did not recur.
Why it matters
Closing an action on paper is not the same as fixing the problem.
CAR-04Are lessons fed back into FMEAs, work instructions and new projects?
Evidence to request
An example of an FMEA or work instruction updated after a corrective action.
Why it matters
It shows whether the factory learns across products or only fixes one customer's issue.
CAR-05How are customer returns (RMAs) analyzed and reported?
Evidence to request
RMA process, failure analysis turnaround time and return trend reports.
Why it matters
Return analysis is how field problems reach the production line.
CAR-06Is there a defined escalation path to management for repeat or serious issues?
Evidence to request
Escalation matrix with names, roles and time limits.
Why it matters
You need to know who to call when the normal channel is not working.

11 · 7 checks

Supply chain and subcontractors

Who else is involved in making your product, and how materials are bought and controlled.

SCM-01Which processes are subcontracted (for example SMT, molding, painting, cables or packaging), and to whom?Critical
Evidence to request
Process map showing in-house and subcontracted steps, with subcontractor names and locations.
Why it matters
Your product quality depends on sites you may never have seen.
SCM-02How are subcontractors and key suppliers qualified and audited?
Evidence to request
Supplier audit schedule, recent audit reports and the approved supplier list.
Why it matters
The factory's supplier controls become your supplier controls.
SCM-03Will the factory notify you and get approval before changing a sub-supplier, material or component source?Critical
Evidence to request
Written change-notification commitment and examples of notifications sent to other customers.
Why it matters
Silent changes are one of the most common causes of unexpected field failures.
SCM-04Are long-lead, single-source and allocated components identified and managed?
Evidence to request
Component lead-time and risk list, buffer stock policy and end-of-life alert process.
Why it matters
One unavailable component can stop the whole line.
SCM-05Is it clear whether materials are turnkey (bought by the factory), consigned (supplied by you) or a mix?
Evidence to request
Proposed material model, any material markup and the factory's responsibilities for consigned parts.
Why it matters
The model decides who carries inventory risk and who controls sourcing.
SCM-06How are excess and obsolete materials handled when forecasts change?
Evidence to request
Policy on cancellation windows, non-cancellable non-returnable (NCNR) parts and liability for leftover stock.
Why it matters
Forecast changes can leave you paying for material you never use.
SCM-07Are restricted-substance requirements such as RoHS and REACH flowed down to suppliers and documented?Critical
Evidence to request
Material declarations, supplier compliance statements and test reports where applicable.
Why it matters
Noncompliant materials can block market access or force a recall.

12 · 7 checks

Capacity and planning

Whether the factory can build your volumes, on which line, and how it plans for them.

CAP-01What realistic capacity is available for your product, and how much of it is already committed?Critical
Evidence to request
Capacity calculation for your product (units per hour, shifts, uptime) and current utilization of the relevant lines.
Why it matters
Quoted capacity is often total site capacity, not what is free for you.
CAP-02Which line and shift will build your product, and is it shared with other customers?
Evidence to request
Named line, layout and the scheduling rule when several customers share it.
Why it matters
Shared lines get reprioritized when a larger customer surges.
CAP-03How are materials and production planned?
Evidence to request
ERP or MRP system, planning cycle and how your forecast becomes purchase orders.
Why it matters
Weak planning shows up as shortages, expediting costs and missed dates.
CAP-04Is there a ramp plan from pilot to full rate, including staffing and training?
Evidence to request
Ramp plan with weekly volumes, headcount and operator training plan.
Why it matters
Ramping faster than people can be trained usually costs yield.
CAP-05Is the factory financially stable, and how concentrated is its customer base?
Evidence to request
Business registration, years in operation and the approximate share of revenue from its largest customers.
Why it matters
A factory that depends on one customer may struggle, or change priorities, if that customer changes.
CAP-06How are holiday shutdowns, such as Chinese New Year, and labor turnover planned for?
Evidence to request
Holiday calendar, pre-build plan and staff retention approach.
Why it matters
Output and quality can dip for weeks after a long holiday while new staff are trained.
CAP-07What lead times are committed for samples, pilot builds and production orders?
Evidence to request
Lead times by stage, with component lead times shown separately.
Why it matters
A schedule built on unconfirmed lead times will slip.

13 · 7 checks

IP protection and information security

Agreements, tooling ownership and the practical controls that protect your design.

IP-01Is a confidentiality agreement in place that can be enforced where the factory operates, such as an NNN agreement for work in China?Critical
Evidence to request
Signed agreements, the governing law and language, and the counterparty's registered entity name. Take local legal advice on the agreement itself.
Why it matters
An agreement that cannot be enforced where the factory operates offers limited protection.
IP-02Is ownership of tooling, molds, fixtures and design files documented?Critical
Evidence to request
Tooling agreement, tooling register and asset tags on your molds and fixtures.
Why it matters
Unclear ownership makes it hard to move production or recover tools later.
IP-03Who can access your design files, firmware and data?
Evidence to request
Access controls on project folders, policies on USB drives and personal email, and who is on your project team.
Why it matters
Access that is not limited cannot be audited.
IP-04Are your samples and product area kept away from visitors and other customers?
Evidence to request
Visitor policy, separated project areas and sample storage controls.
Why it matters
Unreleased products are most exposed during development and pilot builds.
IP-05How are rejected branded units, overruns and surplus packaging destroyed?
Evidence to request
Scrap and destruction procedure and records, including for branded packaging.
Why it matters
Uncontrolled scrap and overruns can reach the gray market under your brand.
IP-06For connected products, how are signing keys and device credentials protected during programming?
Evidence to request
Key handling process, who can access keys and how programming stations are controlled.
Why it matters
Leaked keys or credentials can compromise every device in the field. Mark N/A if not applicable.
IP-07In ODM work, is it clear which parts of the design belong to you and which belong to the factory's platform?
Evidence to request
Contract terms on design ownership, exclusivity and use of your changes for other customers.
Why it matters
If the base platform belongs to the factory, define what is yours before development starts.

14 · 7 checks

EHS and social compliance

Environment, health, safety and labor practices at the site.

EHS-01Does the factory hold a valid business license and the environmental permits its processes require?Critical
Evidence to request
Business license, environmental approvals and discharge permits as applicable to its processes.
Why it matters
A site operating without required permits can be closed at short notice.
EHS-02Are working hours, wages and overtime within local law and your code of conduct?Critical
Evidence to request
Time and payroll records for a sample of workers, and a recent independent social audit (for example SMETA, amfori BSCI or RBA) with its corrective actions.
Why it matters
Labor violations are a legal, ethical and brand risk, and excessive overtime also affects quality.
EHS-03Is there any sign of child labor, forced labor or withheld identity documents?Critical
Evidence to request
Age verification records, policy on identity documents and, where possible, worker interviews without management present.
Why it matters
These are non-negotiable for most brands and are illegal in most markets.
EHS-04Are fire safety measures in place, including clear exits, extinguishers, drills and safe storage of flammables and batteries?Critical
Evidence to request
Walk the site; fire drill records and fire inspection certificates.
Why it matters
Blocked exits and poor storage are life-safety issues.
EHS-05Are workers protected from chemical, noise and machine hazards?
Evidence to request
PPE in use on the floor, safety data sheets at points of use and machine guarding.
Why it matters
Basic safety practice is also a good indicator of overall discipline.
EHS-06How is hazardous waste such as solder dross, chemicals and batteries handled?
Evidence to request
Contracts with licensed waste handlers and waste transfer records.
Why it matters
Illegal disposal can lead to fines and site closures.
EHS-07If your customers require it, can the factory provide responsible minerals reporting?
Evidence to request
Conflict Minerals Reporting Template (CMRT) or Extended Minerals Reporting Template (EMRT) and smelter lists.
Why it matters
Many brands and retailers require this data from their supply chain.

Frequently asked questions

What is a factory audit?

A factory audit is a structured review of whether a manufacturer can build your product to the required quality, in the volumes you need, under acceptable conditions. For electronics it typically covers the quality system, engineering and NPI process, incoming inspection, SMT and assembly controls, test coverage, ESD, traceability, corrective action, subcontractors, capacity, IP protection and social compliance, and records the evidence behind each conclusion.

What is the difference between a desk audit and an on-site audit?

A desk audit reviews documents and records the factory sends: certificates, procedures, sample reports and organization charts. An on-site audit checks whether those controls are actually used on the line that will build your product, lets you choose your own samples and meet the engineers. A desk audit is a good filter and makes the on-site visit more focused; it is not a substitute for it before a significant commitment.

How long does a factory audit take?

It depends on the product, the number of processes and how prepared the factory is. A desk review can start as soon as documents arrive. An on-site audit of a single electronics assembly site commonly takes one or two days, longer when several sites or subcontractors are involved. Allow time afterwards to follow up on concerns and to re-check critical items.

Who should carry out the audit?

Someone who understands electronics manufacturing and your product: typically a quality or manufacturing engineer, your own or an independent one. Social compliance audits such as SMETA, amfori BSCI or RBA are carried out by approved third-party auditors. Regulated products may also need an audit against the relevant sector standard.

Where Zefinity helps

Factory audits and supplier qualification are part of our Connect work: requesting relevant evidence, coordinating interviews, sample reviews or on-site audits where in scope, and recording what was verified and what remains unconfirmed before you select a partner.