How to Specify Supplier Pre-Shipment Functional Tests and Acceptance Criteria

A supplier pre-shipment functional test is only useful when it supports a defined release decision. The purchase requirement must identify the exact product configuration, test conditions, method, measurement, pass/fail limit, records, failure handling, and person authorized to release shipment. A video of equipment running or a report marked “PASS” is not, by itself, sufficient evidence that agreed requirements were met.

Define the release decision before writing the test

First decide what the pre-shipment activity is intended to prove. A routine release check, a witnessed factory acceptance test (FAT), design validation, regulatory assessment, and site acceptance test serve different purposes. Combining them under a vague requirement such as “test before shipping” creates disputes because neither party knows what result is required for release.

For most purchased equipment, the practical question is narrower: may this identified product be released for packing and shipment against this purchase order? The test plan should be written around that question.

Before agreeing individual test cases, identify:

• Purchase order number, line item, and applicable revision.
• Supplier legal entity and actual test location.
• Product model, configuration, serial numbers, and software or firmware revision where relevant.
• Approved drawings, bill of materials, wiring diagrams, control narratives, and specification revisions.
• Quantity in the order, quantity completed, quantity presented for test, and any excluded units.
• Whether testing is performed on every unit, a defined sample, or one integrated system.
• The documents that control if requirements conflict.
• Required test witnesses, reviewers, and shipment-release authority.

A test can be technically well executed and still be commercially inadequate if it is performed on the wrong configuration or on an unclear population. For example, a successful test on one prototype does not automatically release a production lot unless the contract defines it as representative and controls changes from the approved unit.

Inspection guidance for buyers commonly emphasizes confirming the supplier, site, lot, product quantity, specification revision, sample plan, and exceptions before release. Currawong Web’s pre-shipment inspection guidance makes the related point that an inspection report records what was checked under its stated scope; it does not guarantee every unit, future production, or destination-market compliance.

Set the release status in advance. A simple structure is often sufficient:

The supplier may prepare the records and recommend a status, but the buyer or an authorized reviewer should own the final commercial release decision unless the contract assigns that authority differently.

Translate required functions into observable test cases

“Operates normally” and “works correctly” are not testable requirements. Convert each required function into a test case that another competent person can repeat and review.

Start with the approved product requirements. Depending on the equipment, relevant functions may include motion sequence, start/stop response, output rate, control logic, alarm response, interlocks, electrical input compatibility, interface communications, pressure retention, valve operation, or load handling. Do not assume a duty condition merely because it is typical for similar equipment. Use the signed technical specification, approved operating instructions, or engineering-approved test requirement for the exact configuration.

Each test case should answer the following questions:

1. What function or failure mode is being checked? State the intended operation in specific terms.
2. What is the test article? Record model, serial number, configuration, revision, and software version where applicable.
3. What is the initial condition? Define assembly condition, setup, utility supply, ambient conditions, and pre-test status where these affect results.
4. What input or simulated duty is applied? Identify voltage, pressure, temperature, load, material, speed, command signal, or other relevant input.
5. What sequence is followed? State the steps, timing, operator actions, and any required stabilization period.
6. What is measured or observed? Name the output, response, condition, or event that determines conformity.
7. What method and equipment are used? Identify fixtures, instruments, test software, and measurement locations.
8. What result passes or fails? Give a numerical limit, tolerance, or unambiguous observable condition.
9. What evidence is retained? Specify raw readings, screen captures, photographs, video, completed checklist, or data files.
10. What happens after a failure? Define segregation, investigation, repair, retest, and approval requirements.

For example, “verify motor operation” is not an adequate test case. A more usable requirement could identify the correct motor and drive configuration, specified supply conditions, commanded direction, required speed range, expected alarm state, duration, instrument or controller readout used, and permissible result. The actual values must come from the approved technical package, not from a generic article or supplier preference.

The same approach applies to custom machinery. “Demonstrate automatic cycle” should be divided into the relevant sequence steps: safe start condition, sensor recognition, actuator movement, control response, alarm or interlock behavior, and end-of-cycle state. This does not mean every feature must be tested under full production duty at the factory. It means the agreed scope must state which conditions are simulated, which are demonstrated, and which remain for site commissioning.

Sourcing All’s functional sample acceptance guidance usefully frames a functional specification around the characteristic, test article and condition, method, equipment, sample selection, pass/fail limit, retest rule, evidence deliverable, and change trigger. Treat this as practical structure rather than a universal test protocol; the content of each field remains product- and contract-specific.

Set measurable pass/fail criteria and test conditions

A test result is not objectively acceptable unless both the acceptance limit and the test condition are specified. “No excessive vibration,” “acceptable leakage,” or “stable output” leaves too much to individual judgment unless the parties define what those terms mean for the particular product.

Use the applicable approved drawing, purchase specification, validated performance curve, or contractually invoked standard as the source of the limit. Record the document number, revision, clause, and relevant operating condition beside the test result. Avoid transferring limits from a different model, material grade, pressure class, or control version.

A complete acceptance criterion normally includes:

• Measured characteristic and unit, such as pressure in kPa or bar, speed in rpm, current in A, temperature in °C, or dimensional displacement in mm.
• Required value, permitted range, or maximum/minimum limit.
• Test input and environmental condition.
• Test duration, cycle count, or stabilization period where material to the result.
• Measurement point and calculation method.
• Number of units or samples to be tested.
• Instrument range and resolution suitable for the limit being checked.
• Instrument identification and calibration status when measured values are contractual evidence.

Do not write “use calibrated equipment” as the entire measurement-control requirement. Where the measurement supports release, ask the supplier to identify the instrument or data acquisition channel used, its identification number, and calibration due date or applicable status. A calibration certificate alone does not prove that the correct instrument, range, setup, or method was used for a particular result.

Some criteria are appropriately binary. For example, a specified alarm may be required to activate when a defined input condition is applied, or a guarding interlock may be required to prevent a commanded motion under a stated condition. Even then, define the setup and observation point. A binary pass/fail statement is reliable only when the required behavior is precise.

Sampling also needs explicit control. There is no universally appropriate sample size, test duration, tolerance, or retest population. These depend on lot structure, product risk, the consequences of failure, whether defects are detectable through functional testing, contractual terms, and any applicable standard. If every unit must meet a critical electrical, pressure, or safety-related functional requirement, a sample demonstration may not provide adequate evidence. That decision requires product-specific engineering and quality review.

Keep functional verification separate from inspection and qualification

Functional verification is one part of a broader conformity plan. It should not be presented as proof of every quality, safety, or compliance requirement.

Visual and dimensional inspection checks physical attributes: finish, marking, packaging, dimensions, material markings, assembly completeness, and visible defects. It may confirm that a measured feature matches an approved drawing, but it does not necessarily show that the product will perform under duty.

Functional verification confirms an agreed operation under stated factory conditions. It can demonstrate that a control sequence runs, a valve responds, an actuator moves, or a machine achieves an agreed output during the defined test. It does not automatically establish long-term reliability, full field performance, or integration with the buyer’s site.

Formal qualification or compliance testing may require a separate plan, qualified personnel, independent review, or a named laboratory. Examples can include type testing, destructive testing, material verification, pressure testing, electrical safety assessment, certification review, and validation of safety functions. Whether such work is required depends on the product, jurisdiction, contract, and governing standard.

Site acceptance testing (SAT) addresses conditions that cannot be reproduced adequately at the factory: installed utilities, process materials, facility controls, foundations, piping, operator setup, upstream and downstream equipment, and actual duty cycle. A successful FAT reduces risk before shipment, but it does not make SAT unnecessary when site variables are material.

Use the quality plan or inspection and test plan to map each requirement to an appropriate verification method: document review, visual inspection, dimensional measurement, functional demonstration, formal test, certification review, or site test. Tests involving stored pressure, electrical hazards, lifting, rotating equipment, high temperature, chemicals, or safety functions need product-specific procedures and competent engineering review. Do not ask a supplier to improvise hazardous test conditions simply to create a more impressive pre-shipment video.

Specify the records required for remote review

Remote review is strongest when it is based on controlled records, not presentation material alone. Require the evidence package before shipment release, with enough detail for the reviewer to determine what was tested and whether the applicable criteria were met.

A practical functional-test record package includes:

• Approved test procedure number and revision.
• Purchase order, product, configuration, serial-number, and lot identification.
• Test date, location, and supplier personnel roles.
• Test setup description, including relevant utilities, fixtures, loads, and simulated duty.
• Test-case checklist with referenced requirements and pass/fail limits.
• Raw readings, data logs, controller screenshots, or calculation sheets where applicable.
• Instrument identification and calibration-status evidence when measurements support acceptance.
• Photographs or video showing identification, setup, and key test stages when useful.
• Nonconformance report, corrective action, retest evidence, and deviation disposition.
• Signature or electronic approval fields for supplier, witness, and authorized release reviewer as applicable.

Video can be valuable supporting evidence. It can show the product identity, physical setup, motion sequence, indicator states, and a portion of the test in real time. However, it rarely establishes compliance on its own. A video may not show the actual input condition, measurement method, full duration, instrument reading, product revision, or population tested. Treat it as corroboration of a defined test method and recorded result, not as a substitute for them.

A structured FAT record is useful because it captures measurable expectations and retained results for later review. Process Navigation’s FAT template discussion similarly emphasizes documented performance expectations and test results rather than a generic checklist. The buyer should still define the contractual content, document format, record-retention period, and timing for submission.

Control failures, retests, deviations, and shipment release

A failed test is not resolved by overwriting a result, repeating the test without explanation, or issuing a new report marked “PASS.” The record should preserve the original failure and show the disposition.

The required workflow should state:

1. Identify the failed criterion, affected serial numbers or lot, and test condition.
2. Place affected units on hold or otherwise maintain clear identification and segregation.
3. Record immediate containment and preliminary cause investigation.
4. Define the proposed correction, repair, replacement, or process action.
5. Decide whether retest is allowed, which test cases must be repeated, and whether the retest population expands beyond the original unit or sample.
6. Review whether the failure triggers an engineering change, document revision, or renewed approval of the configuration.
7. Obtain written concession approval if shipment with a deviation is proposed.
8. Keep shipment on hold until the authorized reviewer issues a documented release decision.

A retest rule should not allow selective replacement of failed units followed by an unexplained re-run on only the repaired product. The appropriate retest scope depends on the failure mechanism and the possibility that the issue affects other units. For a lot-wide process concern, the buyer may need broader containment and engineering review. For an isolated assembly error with verified traceability, the affected population may be smaller. The quality agreement or inspection and test plan should establish who decides.

The Sinospect FAT checklist guidance describes a FAT as a controlled hold point with agreed criteria, documented results, nonconformity follow-up, and a release-or-hold decision. That is a useful procurement principle: if an authorized criterion is missing, preserve the observed result but do not declare conformity until the acceptance basis is agreed.

Minimum pre-shipment functional test-plan checklist

Before approving a supplier test plan, confirm that it answers all of these questions:

• Is the exact order, configuration, revision, quantity, and test population identified?
• Does each required function have a defined method, input condition, observable output, and pass/fail limit?
• Are units, durations, load or simulated duty, and stabilization requirements stated where relevant?
• Are instruments, measurement locations, and required calibration records identified?
• Is the boundary between inspection, functional testing, formal qualification, and site testing clear?
• Are required records, submission timing, witnesses, and signatories defined?
• Does the plan state how failures, retests, deviations, and corrective actions are controlled?
• Is the person authorized to hold or release shipment identified in writing?

Frequently asked questions

What is the difference between a pre-shipment inspection and a functional acceptance test?

A pre-shipment inspection commonly checks quantity, appearance, packaging, markings, documentation, and selected physical characteristics against a defined scope. A functional acceptance test checks whether the identified product performs agreed functions under stated test conditions. A single pre-shipment activity can include both, but each requirement should retain its own method and acceptance criterion.

What should a supplier pre-shipment functional test record include?

At minimum, include product identification, applicable document revisions, test procedure, test conditions, test equipment, raw results, acceptance limits, pass/fail status, nonconformances, retest evidence, and release signatures or approvals. The required level of detail should match the product risk and contractual requirements.

Can a supplier video be accepted as proof that equipment passed a functional test?

Video can support review, especially when it shows product identity, setup, and an operating sequence. It is not sufficient proof by itself unless the agreed test method and acceptance criteria make it sufficient. It normally cannot replace raw measurements, defined test conditions, controlled records, or a documented pass/fail decision.

Who should approve shipment when a functional test has an unresolved deviation?

The buyer or an explicitly authorized reviewer should approve any release with an unresolved deviation, following the purchase order, quality agreement, and deviation-control process. The supplier can propose a concession, but should not unilaterally convert a failed or incomplete requirement into an accepted condition.

The bounded selection rule is simple: release equipment only when the exact product and test population are traceable, each release-critical function has been checked against an agreed method and criterion, required records are complete, and any deviation has written approval from the party authorized to accept it. Everything else should remain a hold point, an engineering-review item, or a site-commissioning obligation.