Summary
In short
- Acceptance criteria must be defined before the trials run. Criteria set afterwards describe what happened rather than what was required, and the exercise proves nothing.
- Validation matters most where output cannot be verified afterwards: sterilisation, welding, sealing, heat treatment, cleaning and coating, which are the classic special processes.
- Challenge conditions rather than ideal ones. A process validated at nominal settings has demonstrated it works when everything is right, which is not the question.
- The worst case is the point of validation: minimum time, maximum load, coldest incoming temperature, oldest equipment, least experienced operator within the permitted range.
- Revalidation triggers should be defined in the record: change to equipment, material, method, product or a defined period since the last validation.
- A validation with no failures at any condition tested has usually not tested the edge of the operating window, and therefore has not located it.
What it is
What it is
What is validation?
Objective evidence that a process is capable of consistently achieving its intended result, established before the process is relied upon and re-established when it changes. It is required particularly where the resulting output cannot be verified by subsequent monitoring or inspection.
How is it different from verification?
Verification confirms that specified requirements have been fulfilled on a given occasion. Validation confirms that requirements for a specific intended use can be consistently met. Verification is a check on an output; validation is a statement about a capability, and the second requires a protocol, trials and acceptance criteria set in advance.
When to use it
When to use it, and when not to
This records a validation exercise. Routine verification and monitoring follow from it.
Use it for
- Processes where the output cannot be fully verified by subsequent inspection or testing
- New processes, equipment, products or materials before production release
- Changes to a validated process, including equipment, method, material or supplier
- Cleaning validation, including allergen removal and changeover methods
- Establishing critical limits for a control point, where validation is what makes the limit defensible
Not for
- Routine verification and monitoring, which confirm the validated process is operating as intended
- Equipment qualification, which establishes the equipment is installed and operating correctly and precedes validation
- First article inspection, which verifies a production process through a produced part
- Calibration, which addresses the measurement rather than the process
- Capability studies, which describe ongoing variation once the process is validated and running
Standards
What it is built against
Validation is required by quality standards generally and prescriptively in regulated sectors.
| Clause | Requirement | Where it lands |
|---|---|---|
| ISO 9001 cl.8.5.1(f) | Validation and periodic revalidation of processes where resulting output cannot be verified by subsequent monitoring | Header |
| ISO 9001 cl.8.3.4 | Design and development controls including validation that outputs meet requirements for the intended use | Protocol |
| 21 CFR 820.75 | Process validation for medical devices, with documented procedures, approved protocols and qualified personnel | Protocol |
| 21 CFR 117.160 | Validation of process preventive controls, providing scientific evidence the control is capable | Conclusion |
| Codex principle 3 | Establish validated critical limits, distinguishing validation from monitoring and verification | Conclusion |
| BRCGS Issue 9 | Validation of processes and control measures, with records and defined revalidation triggers | Trials |
| AS9100 cl.8.5.1.3 | Production process verification, related to but distinct from validation | Header |
| ISO 13485 cl.7.5.6 | Validation of processes for production and service provision, including software validation | Protocol |
What it does not cover
- Routine verification and monitoring, confirming the validated process is running as intended.
- Equipment qualification, establishing correct installation and operation before validation.
- First article inspection, verifying a production process through a produced part.
- Calibration, addressing the measurement system rather than the process.
- Capability studies, describing ongoing variation once the process is validated.
Filling it in
Filling it in well
Write the protocol first, challenge the worst case, and define what would trigger doing it again.
What is being validated, against what requirement, under what conditions, how many replicates, what measurements, and what result constitutes acceptance. Approved and dated before the first trial. This sequence is the entire basis on which the conclusion carries weight, and reconstructing it afterwards produces a record that passed by construction.
Shortest time, lowest temperature, heaviest load, largest pack, coldest incoming product, oldest equipment in the fleet, and the operator conditions actually permitted rather than the best available. A process validated at nominal settings has demonstrated something true and not useful, because production does not run at nominal.
A validation that located the boundary of the operating window is more useful than one that passed everywhere, because it tells you where the edge is. Where nothing failed at any condition tested, the conditions were probably not challenging enough, and the record should say what range was explored.
Change to equipment, material, supplier, method, product or packaging, and a defined maximum period. Written into the validation record itself, these become findable when someone later asks whether a change requires revalidation, which is a question usually asked after the change has been made.
Audit findings
Common audit findings
Validation findings concentrate on sequence and on what was actually challenged.
| Finding | Clause | What fixes it |
|---|---|---|
| Acceptance criteria defined after trials were completed. | 21 CFR 820.75 | Approve the protocol before running; criteria derived from results prove nothing. |
| Validation conducted at nominal rather than worst-case conditions. | 21 CFR 117.160 | Challenge the limits of the permitted operating window. |
| Verification records presented as validation. | ISO 9001 cl.8.5.1(f) | They answer different questions; verification cannot establish capability. |
| No revalidation triggers defined. | BRCGS Issue 9 | Write them into the record; the question arises after the change is made. |
| Number of replicates not justified. | 21 CFR 820.75 | State why the sample size supports a conclusion about consistency. |
| Validation not repeated after an equipment or material change. | ISO 9001 cl.8.5.1(f) | Change invalidates the evidence; link validation to management of change. |
| Cleaning method verified but never validated. | BRCGS Issue 9 | Verification confirms the method was followed; validation establishes it works. |
| Critical limits adopted from a code of practice without validation for the process. | Codex principle 3 | Validate against your product and equipment; borrowed limits are assumptions. |
| Protocol approved by someone without the technical basis to approve it. | 21 CFR 820.75 | Approval requires qualified personnel; a signature is not the requirement. |
| Validation record does not identify the equipment or configuration used. | ISO 13485 cl.7.5.6 | The evidence applies to what was validated; record it precisely. |
Worked case
Case in point: validated at the easy setting
A sealing process was validated on a heat sealer set to a dwell time and temperature in the middle of the permitted range, with product at ambient and the newest of four machines. Twenty samples were tested, all passed, and the process was released.
Production ran across all four machines, on chilled product from the packing hall, at the lower end of the dwell range when line speed increased. Seal failures appeared intermittently and were investigated as a machine fault.
The validation had established that a well-maintained machine seals warm product at nominal settings, which nobody doubted. The conditions under which the process actually ran had not been part of the exercise.
Definitions
Definitions and key terms
- Validation
- Objective evidence that a process can consistently achieve its intended result, established before reliance and re-established on change.
- Verification
- Confirmation that specified requirements have been fulfilled on a given occasion.
- Protocol
- The approved plan defining what will be tested, under what conditions, how many times and what constitutes acceptance.
- Worst case
- The most challenging combination of conditions within the permitted operating window, which is where validation should be performed.
- Special process
- A process whose output cannot be fully verified by subsequent inspection, such as welding, sterilisation, sealing or heat treatment.
- Revalidation
- Repeating validation following change or after a defined period, with the triggers ideally written into the original record.
- Equipment qualification
- Establishing that equipment is correctly installed and operating as intended, preceding process validation.
- Operating window
- The range of conditions within which the process is permitted to run, and the range validation should challenge.
FAQ
Frequently asked questions
What is the difference between validation and verification?+
Verification confirms that requirements were met on a given occasion, by inspecting or testing the output. Validation establishes that the process is capable of consistently meeting requirements, which is a statement about capability rather than about one result. Verification cannot substitute for validation, particularly where the output cannot be fully inspected.
Which processes need validation?+
Those where the resulting output cannot be verified by subsequent monitoring or measurement, which ISO 9001 states explicitly. In practice: sterilisation, welding, sealing, heat treatment, coating, cleaning and allergen removal. You cannot inspect a seal to know it will hold, or a cleaned surface to know the allergen is gone, without destroying or testing every unit.
Why must acceptance criteria come first?+
Because criteria written after the results are known describe what happened rather than what was required, and the exercise then passes by construction. The protocol, including conditions, replicates and acceptance criteria, has to be approved and dated before the first trial for the conclusion to carry any weight.
What conditions should be used?+
The worst case within the permitted operating window: shortest time, lowest temperature, heaviest load, coldest incoming product, the oldest equipment in the fleet. Production runs across the whole window and the whole fleet, and a validation at nominal settings demonstrates something nobody was in doubt about.
When does validation need repeating?+
On change to equipment, material, supplier, method, product or packaging, and after a defined period. Writing those triggers into the validation record itself makes them findable when someone later asks whether a change requires revalidation, which is usually asked after the change has already been made.
The agents
What the agents do with it
The record establishes capability. What fails is criteria written afterwards and a window nobody challenged.
Holds the approved protocol with its date, records trials against pre-set criteria, and carries revalidation triggers on the record itself.
Connects equipment, material and method changes to the validations that depend on them, so revalidation is raised before production continues.
Links validated processes to the specific equipment and configuration used, since validation evidence applies to what was validated.
Confirms that those approving protocols and interpreting results have the technical basis the standards require.
This template lives in KnowQuality — quality and food safety. HACCP, nonconformance, traceability, laboratory and customer complaints.
Sources
Sources
- ISO 9001:2015 clauses 8.3.4 and 8.5.1
- 21 CFR 820.75, process validation, FDA
- 21 CFR 117.160, validation of preventive controls, FDA
- ISO 13485:2016 clause 7.5.6, validation of processes for production and service provision
- Codex Alimentarius CXG 69, guidelines for the validation of food safety control measures