PPQ (Process Performance Qualification): Stage 2 Process Validation
Understand how qualification studies connect process knowledge, commercial-scale execution, sampling evidence and documented decisions.

On this page · 11 sections
- 1PPQ (Process Performance Qualification): Stage 2 Process Validation
- 2FDA guidance, U.S. law and EU Annex 15: compare scope first
- 3Where PPQ fits between process design and continued verification
- 4Write a protocol that makes the final decision testable
- 5Design sampling around variability, not a convenient sample total
- 6Justify study size with independent evidence and coverage
- 7Connect execution records, results and exceptions without losing context
- 8Recognize failure modes before they distort the conclusion
- 9Carry unresolved questions into a defined verification plan
- 10Review the conclusion against the protocol and actual evidence
- 11Software support: assemble evidence, retain human decisions
How does PPQ apply to your shop floor?
Pick your industry and scale — Ask V5 rewrites the definition in your context, gives a worked example, and shows what V5 does on day one.
01PPQ (Process Performance Qualification): Stage 2 Process Validation
What PPQ establishes—and what it does not
PPQ (process performance qualification) is the commercial-scale evaluation of a defined manufacturing process to establish whether it can reproducibly produce product meeting predefined quality criteria.
In FDA’s nonbinding 2011 guidance, Section IV.C.2 places PPQ as the second element of Stage 2, process qualification. A successful PPQ confirms the process design and demonstrates that the commercial manufacturing process performs as expected. PPQ is therefore part of qualification—not another name for the entire validation lifecycle.
“PPQ batches” means the batches included in that qualification study, not a separate product category. The following distinctions reflect pharmaceutical validation practice; they separate the objects evaluated and decisions made rather than establish regulatory equivalence.
| Term | Object evaluated | Main question | Evidence or decision |
|---|---|---|---|
| Process validation | Process across its lifecycle | Can it consistently deliver quality product? | Accumulating design, qualification and production evidence |
| Process qualification | Commercial process design and readiness | Is reproducible commercial manufacturing demonstrated? | Qualification-stage conclusion |
| PPQ | Defined commercial manufacturing process | Does the process perform as expected? | Protocol results and study conclusion |
| Equipment performance qualification | Equipment under intended operating conditions | Does the equipment perform its intended function? | Equipment performance evidence |
| Continued process verification | Process during routine production | Does it remain in control? | Monitoring results and follow-up actions |
| Individual batch release | One manufactured batch | Is this batch acceptable for distribution? | Authorized batch disposition |
Keep these evidence boundaries explicit: process validation spans the lifecycle; equipment qualification addresses equipment readiness and performance; batch release concerns disposition of an individual batch. Passing equipment tests does not establish process reproducibility, and an acceptable batch result alone does not answer the study’s broader question. A PPQ conclusion should stay within the process conditions and variability actually evaluated.
02FDA guidance, U.S. law and EU Annex 15: compare scope first
Begin a PPQ comparison with the product, manufacturing activity and destination market—not the study label. FDA’s Process Validation: General Principles and Practices is guidance containing nonbinding recommendations. Applicable U.S. drug-CGMP regulations are binding requirements; the guidance and the regulations should not be treated as interchangeable authorities.
| Jurisdiction or product context | Reference | Status in this article | Applicability question |
|---|---|---|---|
| U.S. commercial drug manufacturing | FDA’s 2011 process validation lifecycle guidance | Nonbinding recommendations | Does the product and manufacturing activity fit the guidance’s scope? |
| Applicable U.S. drug-CGMP operations | 21 CFR 211.100(b) and 211.110(d) | Binding regulations within applicable scope; specific duties belong in the records and failure-control discussions | Which operations fall within the applicable drug-CGMP requirements? |
| EU medicinal-product operations | EU GMP Annex 15 | Current-text and applicability check | What do the current text and applicable product and market commitments mean for the proposed study? |
| Clinical-trial or treatment IND material | FDA’s commercial manufacturing process definition | Excluded from that glossary definition, not a statement that the entire guidance is irrelevant to development | Is the material being discussed as development material or commercial product? |
| Supplements, food, devices and chemicals | Product- and market-specific framework | Practical applicability assessment, not a universal pharmaceutical PPQ obligation | Which framework applies, and what qualification terminology and evidence are appropriate? |
For EU work, check Annex 15’s current text and applicable product and market commitments before comparing study scope, batch strategy, acceptance criteria or documentation. This comparison does not establish identical terminology, mutual acceptance or a shared numerical batch minimum.
As a practical planning method, record the applicability decision before adapting a process validation template. Keep binding cGMP obligations, guidance recommendations and locally chosen study controls distinguishable. Software support belongs in a separate category: organizing study evidence is not the legal basis for the study or a substitute for human approval and release.
03Where PPQ fits between process design and continued verification
FDA’s lifecycle discussion describes three stages in its nonbinding guidance. Treat them as connected evidence questions, not simply consecutive dates on a validation schedule.
- Stage 1 — Process design: What commercial process is proposed, and why should its controls work? FDA describes defining the commercial manufacturing process and documenting control decisions. In practical terms, the output is a proposed process with a reasoned control strategy, not merely a collection of successful development runs. Explain how selected critical process parameters relate to quality and operating limits.
- Stage 2 — Process qualification: Can the designed process support reproducible commercial manufacture? FDA’s glossary defines qualification as “Confirming that the manufacturing process as designed is capable of reproducible commercial manufacturing.” Its Stage 2 discussion includes facility design, equipment and utility qualification, and PPQ. PPQ therefore sits within qualification; it is not a synonym for the entire validation lifecycle.
- Stage 3 — Continued process verification: Does the process remain in control during routine production? FDA describes ongoing assurance at this stage, rather than treating qualification as the end of validation.
For a practical readiness gate before PPQ, check the process version, material inputs, equipment status, measurement methods and personnel readiness against the planned study. A failed gate sends the affected work back to investigation, preparation or process design; a scheduled start date is not evidence of readiness.
Read the workflow diagram as evidence-dependent decisions: design → readiness review → qualification including PPQ → routine verification, with return paths wherever evidence leaves a question unresolved.
- Define the processConnect development knowledge, operating conditions and quality attributes.
- Check readinessConfirm equipment, materials, methods and personnel readiness.
- Authorize the ProtocolSet sampling, criteria, analysis and exception handling.
- Execute the studyRecord conditions, samples, results and unexpected events.
- Reconcile evidenceResolve identity gaps and evaluate departures from plan.
- Assess performanceCompare complete evidence with predefined decision criteria.
- Transfer monitoringAssign follow-up questions, indicators and review responsibilities.
04Write a protocol that makes the final decision testable
FDA’s PPQ Protocol section describes a written protocol specifying manufacturing conditions, controls, testing and expected outcomes. It also recommends tests and acceptance criteria for each significant processing step. These are guidance recommendations; the practical template below translates them into operational fields, not a mandatory FDA document format.
| Protocol component | Practical content |
|---|---|
| Purpose and scope; product/process version | State the qualification question, covered products, process boundaries, exclusions and controlled manufacturing-document versions. |
| Scientific rationale; readiness references | Connect prior knowledge and risks to the study design; identify the readiness records supporting execution. |
| Execution conditions; sampling and analytical methods | Specify intended operating conditions, permitted ranges, sampling instructions, method versions and sample-to-result identifiers. |
| Acceptance logic; analysis plan | Define criteria, analysis populations, statistical methods and how individual results contribute to batch and study conclusions. |
| Deviation handling; responsibilities | Define exception pathways, assessment responsibilities, amendment authority and decision owners. |
| Report structure; handoff to continued verification | Plan criterion-by-criterion conclusions, limitations, unresolved questions and assigned follow-up monitoring. |
Make each attribute decidable
For every important attribute, specify the measurand—the quantity being measured—its unit, method version, sampling location and time, acceptance criterion, analysis population and decision owner. Connect critical quality attributes to the process questions they test. Use quality risk management to explain why particular attributes or conditions receive greater attention.
Keep individual observations, within-batch summaries and across-batch conclusions separate. A satisfactory pooled average does not itself establish that every unit or sampling stratum met its criterion. Write acceptance criteria with an explicit level of application, for example:
- Each tested unit must meet the referenced unit-level limit.
- Each defined sampling stratum must independently satisfy its stated criterion.
- Each batch must meet its specified within-batch criterion.
- The combined study must meet the predefined across-batch criterion using the stated analysis population.
These are drafting examples, not approved product specifications; replace each reference with the applicable controlled criterion and method.
Set exception rules before execution
Define how missing observations affect evaluability; who may invalidate a measurement and on what basis; and how unexpected process conditions trigger assessment. Specify amendment authorization, timing and impact review, preserving the original criteria and reasons for change. State when unresolved investigations prevent a conclusion and how remaining questions will be tracked. Keep study acceptance separate from individual lot-release decisions, as covered in the comparison table above.
05Design sampling around variability, not a convenient sample total
FDA’s PPQ sampling recommendations call for enough samples to provide sufficient statistical confidence in quality within a batch and between batches. The guidance also recommends more extensive sampling during PPQ than during routine production. These are guidance recommendations, not a fixed numerical sampling minimum.
In practical study design, connect the sampling plan to plausible sources of variation. Select risk-relevant positions and time windows: for example, different discharge locations and early, middle and late production. Retain those identities in the analysis. A pooled average can hide a low result at one position offset by a high result elsewhere; comparing strata helps expose that pattern of process variability.
Worked case 1: count units without overstating independence
Assume an illustrative study samples three production time windows per batch, four positions per window and two distinct units per position, across four study batches.
- Per batch: 3 × 4 × 2 = 24 sampled units.
- Across four batches: 4 × 24 = 96 sampled units.
These are unit counts, not a recommended study size. The 96 unit observations are not 96 independent batch observations. Units from the same batch, position or time window may share conditions and therefore be correlated. Clustering weakens an independence assumption when shared conditions make results more alike than independently selected results would be.
Measuring one unit twice still samples one unit. Repeated measurements can help assess measurement repeatability, but they do not add coverage of additional product units or production conditions.
Worked case 2: all-pass does not establish low risk
Assume independent binary unit results, perfect defect detection and a constant 10% unit nonconformance probability. Each sampled unit then has a 90% probability of passing. For ten sampled units, the probability that all pass is 0.9^10 = 0.3486784401, approximately 34.9%.
Thus, an all-pass sample can occur despite a materially defective process: under these assumptions, it happens about one-third of the time. The assumed 10% rate is an illustration, not an acceptable defect rate. This calculation is not the probability that a process is satisfactory after observing ten passes.
Confidence is conditional on the sampling design and model, not a certificate attached to any all-pass dataset. Document assumptions about independence, detection and stability before interpreting results. Preserve time order for later statistical process control, rather than relying solely on pooled summaries. Study duration and batch coverage need their own rationale, addressed next.
06Justify study size with independent evidence and coverage
FDA’s rationale for the PPQ approach recommends sound science, product and process understanding, and demonstrable control. This is guidance for designing the study, not a numerical regulatory minimum inferred from that recommendation.
Treat a proposed three-batch study as a proposal requiring justification, not a legal conclusion. What independent batch-to-batch information will each run add? What uncertainties will remain? What evidence would change the decision to accept the study size or extend it? Repeated runs under nearly identical conditions may demonstrate repeatability while leaving important sources of variation unchallenged.
As a practical design tool, build a coverage matrix with proposed batches as rows and scientifically relevant factors as columns: material lots, equipment configurations, operating conditions, campaign position and shifts. Include a factor only when there is a reason it could affect performance. Record covered conditions, gaps and supporting evidence; do not assume every possible combination needs testing. A planned batch comparison can help structure the review of differences between runs.
Distinguish evidence by the question it answers. Development studies can investigate mechanisms and operating boundaries, including edge-of-failure mapping. Pilot work can investigate scale-up hypotheses. Commercial-scale observations address performance under intended manufacturing conditions. Experience with sufficiently similar processes can inform expectations, provided the relevance and limitations are explained. These are complementary roles, not interchangeable proof.
FDA specifically recommends that the decision to begin commercial distribution be supported by data from commercial-scale batches. Pilot results therefore should not be presented as independently establishing PPQ success.
Finish the justification with six elements: uncertainty to resolve; existing knowledge; proposed independent coverage; analysis and decision criteria; residual uncertainty; and planned follow-up. Use the preceding sampling cases to distinguish sample-level confidence from batch-level coverage, without repeating their arithmetic.
07Connect execution records, results and exceptions without losing context
For U.S. drug manufacturing subject to Part 211, 21 CFR 211.100(b) requires production and process-control activities to be documented at the time of performance, and deviations from written procedures to be recorded and justified. A PPQ evidence chain should make both execution and departures reconstructable.
As an implementation example—not a mandated schema or a particular application database—use this conceptual data model: protocol version → batch execution → sample → result → investigation → final assessment. Link batch execution to material lots and equipment references; link each result to its sample and analytical-method version. Use foreign keys referencing immutable record identifiers, with explicit versions for controlled documents. An identifier stays fixed; a descriptive label, such as an equipment display name, may change. Filenames alone cannot reliably express these relationships.
Connect the electronic batch record and laboratory information management system through defined interface contracts. For potential MES, laboratory-system, historian or ERP interfaces, specify shared identifiers, units and conversions, original event time separately from receipt time, reconciliation totals for records sent, accepted and rejected, duplicate-detection rules, and an owner for each error queue. These interface controls are general implementation considerations, not statements of V5 functionality. Reconciliation should establish that every expected observation is present or explicitly unresolved.
For delayed results, unavailable instruments or failed imports, preserve the original event and identify the gap. On recovery, reconcile identifiers, counts, units and timestamps; retain the earlier value and the correction relationship rather than silently overwriting it. Distinguish a recovered original result from a new measurement or replacement sample. Connect deviation management to the affected records and document the effect on the study. These controls support data integrity during exception recovery.
Worked case 3: Assume 24 planned sampled units, each requiring one result, with one lost sample and one reported failing result. Reconciliation is 22 passing + 1 failing + 1 missing = 24 planned units; only 23 results were observed. Any replacement adds a separately identified observation, not a retroactive original. Do not report 24 passing results after substitution.
FDA’s recommendation on nonconforming PPQ data is not to exclude data from PPQ consideration without documented, science-based justification. In this example, retain the failing result and missing-observation status for assessment rather than treating substitution as justification for exclusion.
Protocol
Defines the versioned study and decision rules.
- protocol_id
- protocol_version
- acceptance_rule_id
- approval_status
Study Batch
Connects execution to its governing protocol version.
- batch_id
- protocol_id
- protocol_version
- work_order_id
- equipment_set_id
Sample result
Preserves observation identity, method and measurement context.
- result_id
- batch_id
- sample_id
- sampling_point
- measured_value
- unit
- method_version
Deviation
Links unexpected events to affected study evidence.
- deviation_id
- batch_id
- result_id
- impact_assessment
- status
Study Assessment
Records the reviewed conclusion and accountable approver.
- assessment_id
- protocol_id
- protocol_version
- conclusion
- reviewer_id
- approved_at
08Recognize failure modes before they distort the conclusion
These practical review checks separate weaknesses in the PPQ evidence from decisions about affected material. An exception can limit the study’s conclusion without automatically requiring batch rejection. Where relevant, connect the event to an out-of-specification investigation; determine its significance rather than treating every anomaly as a completed rejection decision.
| Failure mode | What the reviewer sees | Immediate control | Decision affected |
|---|---|---|---|
| Retrospective criteria changes | Acceptance limits revised after results became available. | Compare dated approvals with result timestamps; assess against original criteria. | Validity of the pass conclusion. |
| Unrepresentative sampling | Planned locations or operating periods absent. | Reconcile planned versus actual sample locations and times. | Coverage of process variability. |
| Pooled averages masking local variation | Acceptable means alongside extreme individual results. | Inspect individual observations by location, time and batch. | Uniformity and consistency conclusions. |
| Incorrect units or method versions | Summaries conflict with laboratory records. | Check units, conversions and executed method revisions against the protocol. | Comparability of results. |
| Untracked process changes | Executed settings or instructions differ from the protocol. | Compare executed process and protocol versions; trace change control. | Applicability to the intended process. |
| Missing evidence | Expected observations or supporting records cannot be traced. | Reconcile the evidence inventory; use the preceding records section’s checks for missing observations and data exclusions. | Completeness of the study. |
| Rejected in-process material not effectively segregated | Rejected containers remain available for unsuitable use. | Match rejection records to container identity, location and access restrictions. | Effectiveness of material control. |
For drug manufacturing subject to 21 CFR 211.110(d), rejected in-process materials must be identified and controlled under a quarantine system designed to prevent their use in manufacturing or processing operations for which they are unsuitable. This requirement applies when material has been rejected; an anomalous result alone does not establish that disposition.
09Carry unresolved questions into a defined verification plan
As covered above, PPQ is a lifecycle milestone, not the end of validation. FDA’s glossary definition of continued process verification sets the objective: assurance that the process remains in a state of control during routine production. The practical handoff to continued process verification should turn the study’s conclusions and remaining uncertainties into an owned monitoring plan.
FDA’s discussion of heightened monitoring after PPQ recommends continuing increased scrutiny, testing and sampling through process verification as appropriate. Completing the protocol is therefore not, by itself, a reason to reduce monitoring.
A practical handoff should identify:
- Monitored attributes and process parameters, with their data sources and links to relevant PPQ findings.
- Review frequency, responsible production and technical roles, and the quality reviewer.
- Investigation triggers, including adverse patterns detected through trend analysis, and responsibility for follow-up.
- Residual uncertainties, the observations needed to resolve them, and when their status will be reassessed.
- The evidence and rationale needed for any later change in sampling locations, intensity or frequency.
For a proposed adjustment, reviewers should answer: what additional process knowledge makes the proposed routine plan adequate? Document which uncertainty has narrowed and why the revised coverage can still detect meaningful deterioration. An adjustment supported by that reasoning differs from an unexplained reduction in evidence. Where uncertainty remains, assign targeted monitoring and a review point rather than treating the PPQ conclusion as closure.
10Review the conclusion against the protocol and actual evidence
Use three decision gates for cross-functional review. This practical checklist and report structure are implementation aids, not a legally mandated template. Assign review and approval authority through the applicable site quality system.
1. Ready to execute?
- Confirm the correct process and protocol versions, defined acceptance criteria, and specified test and analysis methods.
- Check traceable readiness references, including relevant equipment qualification, training and method status.
- Confirm justified study coverage and named reviewers for execution, analysis and the final conclusion.
2. Evidence complete enough to assess?
- Reconcile planned versus actual samples; explain missing, additional, replaced or excluded results.
- Confirm units, method versions and intact sample, batch and equipment relationships against the batch execution history.
- Account for every exception and confirm that analyses represent the intended populations rather than only convenient subsets.
3. Conclusion and follow-up adequately justified?
- Evaluate every criterion; explain relevant within-batch and between-batch behaviour and account for contrary evidence.
- Identify limitations; resolve open issues or document their management, owners and decision implications.
- Assign verification follow-up with measures, review dates and escalation triggers.
Ask process engineering: does the conclusion fit process behaviour? Manufacturing: was execution representative? Laboratory/statistical specialists: are methods, populations and analyses appropriate? Quality: are exceptions, limitations and proposed actions adequately addressed?
Organize the final report as: scope; executed study; departures from plan; results and analysis; investigations; criterion-by-criterion assessment; limitations; conclusion; and follow-up. Link each assessment to its supporting records so reviewers can reconstruct the reasoning without relying on the summary alone.
11Software support: assemble evidence, retain human decisions
Use that assembled record as an input to the evidence review described above, not as the PPQ conclusion itself. Designated humans assess the study against its protocol and authorize batch release; assembling execution records does not confer either approval.
Frequently asked questions
Q.Can existing equipment PQ evidence replace any part of a PPQ study?+
Existing equipment PQ evidence can support the equipment-readiness portion of the study rather than trigger duplicate testing. As a study-design practice, map that evidence to the intended product, load, operating range, configuration and current equipment condition. Identify changes or gaps needing additional work. Do not treat successful equipment operation alone as evidence of reproducible product quality: the PPQ assessment still needs process-level evidence connecting materials, operating conditions, controls and quality results.
Q.How should a team evaluate a proposed three-batch study when process knowledge is limited?+
Treat three batches as a proposed design, not its scientific justification. FDA’s PPQ protocol guidance recommends sampling sufficient to support confidence both within and between batches. As a planning practice, list unresolved variability sources, then ask whether the proposed batches and sampling locations can distinguish their effects. Where critical material attributes, operating ranges or scale effects remain poorly understood, consider further development work before qualification rather than simply increasing the batch count.
Q.Why might every batch meet its release specifications while the overall PPQ conclusion remains unresolved?+
Release results answer whether tested batch attributes meet their specified limits; a PPQ assessment also asks whether the collected evidence supports reproducible process performance. FDA’s PPQ protocol guidance recommends criteria addressing both intra-batch and inter-batch variability. For example, passing results could coexist with unexplained drift across filling time or substantial differences between batches. A batch comparison can help reviewers identify patterns that individual certificates of analysis do not resolve.
Q.Do repeated tests on the same sample provide the same confidence as testing additional sampled units?+
No. As a statistical design principle, repeated measurements mainly characterize measurement variability for that sample; additional appropriately selected units can reveal variation across the material or process. Assuming ten tablets are sampled separately and each is measured twice, the dataset contains twenty measurements but only ten sampled tablets—not twenty independently sampled units. The analysis should preserve that hierarchy. Use a stratified sampling design when location or processing time is a plausible source of variation.
Q.What should happen when a planned sample is lost or a laboratory result arrives after the scheduled review?+
As a records practice, log the event, identify the affected protocol objective and assess whether the remaining evidence supports it. A lost sample creates a coverage gap; a delayed result remains pending, not passing. Follow the protocol’s deviation provisions and document any replacement-sampling rationale. FDA’s PPQ protocol guidance recommends documented, science-based justification for excluding data. Maintain traceability through the sample handoff record, and postpone any conclusion that depends on unavailable results.
Q.Can the same PPQ dataset support both an FDA-focused study and an EU Annex 15 assessment?+
A shared dataset can be a practical starting point, but it is not a declaration that the two assessments are interchangeable. Build a crosswalk from each assessment question to the protocol, execution records, sampling coverage, analyses and conclusions. Check the current text of EU Annex 15 for applicable scope and expectations. As a documentation practice, record gaps explicitly and obtain additional evidence where needed rather than relabelling an FDA-focused report as an EU assessment.
Q.What evidence should reviewers consider before reducing enhanced monitoring after qualification?+
As a risk-based monitoring practice, review accumulated variability, trends, deviations, measurement reliability and how well the observations cover routine operating conditions. Ask whether the proposed lower sampling frequency could miss a meaningful change, particularly across material lots, shifts or equipment configurations. Record the rationale, remaining uncertainties and triggers for restoring enhanced monitoring in the CPV plan. A quiet period alone is weak justification if production has not exercised the variability the monitoring was intended to detect.
Q.How should the final assessment handle open investigations without confusing study completion with batch release?+
List each open investigation, its affected data and its potential effect on the study conclusion. FDA’s concurrent-release guidance says process conclusions depend on full protocol execution and data evaluation, and expects concurrent release to be rare. As a reporting practice, distinguish execution status, evaluation status and authorized batch disposition. If an investigation could change the conclusion, identify that dependency explicitly rather than recording unqualified success. Human reviewers make and document the respective decisions.
Primary sources
Explore this topic
PPQ sits inside 4 overlapping topic clusters in our glossary. Every neighbour is one click away.
Electronic records, signatures, audit trail and ALCOA+ data-integrity principles.
ICH Q8/Q11/Q12 toolkit — Quality by Design, design space, control strategy, CPV and lifecycle management.
URS-through-PQ lifecycle, GAMP 5 categorisation and CSA's modern alternative.
Want to see how PPQ could fit into your own records and workflows? Explore the related V5 pages or talk to our team about what applies to your operation.
