An APQP control plan is the working document that spells out, for every manufacturing operation, exactly which product and process characteristics get measured, how they get measured, how often, and what happens when a result falls outside tolerance. It’s built collaboratively by engineering, quality, and production, it evolves through three successive versions as a product moves from prototype toward full production, and once approved it becomes a binding commitment to the customer for the entire production life of the part.
The Three Versions of a Control Plan
A single part goes through three control plans across its development, each one denser and more disciplined than the last.
Prototype
The prototype control plan covers the earliest physical builds, when parts may come off temporary tooling or non-production equipment. Its job is to verify that the basic design and material choices meet functional requirements. Inspection at this stage is often lab-based or manual, and characteristics tend to be dimensional measurements, material tests, and performance checks against prototype samples. You’re learning whether the design works, not whether a line can repeat it thousands of times.
Pre-Launch
Once production-intent tooling and equipment are in place, the pre-launch control plan takes over. Inspection frequencies go up, sample sizes increase, and the team watches for process issues that only surface at higher volumes. This version bridges proving the design and proving the process. Many manufacturers also layer a Safe Launch plan on top during this stage, adding a period of heightened inspection and containment activity to catch early production problems before parts reach the customer.
Production
The production control plan takes effect once the process has demonstrated statistical stability and adequate capability. Inspection frequencies typically drop compared to pre-launch because the process has earned that reduction, but every characteristic still has a defined monitoring method. This is the version that lives on the shop floor for the life of the part. It governs the SPC charts operators watch, the gauges they use, and the steps they follow when something drifts out of tolerance.
What Goes Into Each Row
A control plan uses a column-based layout where each row represents one operation or process step. The header carries administrative information: part number, part name, engineering change level, supplier code, and the quality team responsible for the document. The body then works left to right through a fixed sequence of columns for every operation.
- Part or process number and operation description, matching the numbering used in the Process Flow Diagram so the two documents stay synchronized.
- Machine, device, jig, or tool performing the operation — the specific press, CNC machine, weld cell, or fixture.
- Product characteristics, meaning measurable features of the part itself such as a bore diameter, surface finish, or material hardness.
- Process characteristics, meaning the parameters that influence those product features, such as furnace temperature, injection pressure, or weld current.
- Special characteristic class, flagging whether the characteristic is safety-critical, significant, or otherwise elevated.
- Specification and tolerance, pulled directly from engineering drawings or material specifications.
- Evaluation or measurement technique — the specific gauge, sensor, test method, or visual standard used to check it.
- Sample size and frequency, driven by the risk levels from the PFMEA.
- Control method, whether an SPC chart, automated sensor alarm, error-proofing device, or visual inspection checklist.
- Reaction plan, giving explicit instructions for the operator when a result falls outside tolerance, typically covering containment of suspect material, notification of supervision, and documented corrective action.
Every measurement system referenced in the plan needs validation through a Measurement System Analysis study, commonly a Gauge Repeatability and Reproducibility study. If the gauge itself introduces too much variation, the data it produces is unreliable, and decisions built on that data can mask real process problems or reject good parts. This is one of the more frequently failed elements during audits, because teams invest in controlling the process but neglect to prove their measurement tools are up to the job.
Special Characteristics and Error-Proofing
Special characteristics are the features most likely to affect safety, regulatory compliance, or fit and function. They get flagged in the Special Characteristic Class column with symbols or codes defined by the customer, and each OEM uses its own designation system. A steering column bolt torque, a brake pad friction coefficient, or a seatbelt anchor weld strength would all qualify. These demand tighter controls: more frequent sampling, mandatory SPC charting, and often dedicated error-proofing devices.
Error-proofing devices, sometimes called poka-yoke, deserve their own line in the control plan because they can fail silently. A sensor that detects a missing component, a fixture that prevents a part from being loaded backwards, or a vision system that checks label placement — each is mechanical or electronic, and each can be defeated by wear, miscalibration, or someone disabling it during a jam. The plan should document not just the device but a verification method and frequency: how the team periodically confirms the device still catches the defect it was designed to prevent, and what happens if it doesn’t.
How the Control Plan Connects to the Other Core Tools
The control plan doesn’t stand alone. It pulls from and feeds every other AIAG Core Tool.1AIAG. Quality Core Tools
- The Process Flow Diagram provides the operation sequence. Every row in the control plan should trace back to a step in the flow diagram; if a step exists in one document but not the other, something was missed.
- The Process FMEA identifies what can go wrong at each step, how severe the consequences are, and how likely the failure is. Its risk rankings directly determine which characteristics need tighter inspection frequencies and which control methods to apply.
- Measurement System Analysis validates the gauges and test methods listed in the control plan. Without an acceptable MSA study, the measurement technique column is a wish list.
- Statistical Process Control provides the charting and monitoring methods referenced in the control method column. SPC data also feeds back into the PFMEA and control plan during reviews; a process showing tight statistical control may justify reduced inspection frequency over time.
- PPAP packages the control plan along with supporting evidence for customer approval before production begins.
These documents must stay in sync. When the PFMEA gets updated because a new failure mode is discovered, the control plan needs a corresponding new row or revised control method. When an MSA study shows a gauge is inadequate, the measurement technique in the control plan has to change. Teams that treat these as separate paperwork exercises get caught flat-footed during audits.
When a Family Control Plan Works
Not every part number needs its own control plan. When multiple parts share a common manufacturing process, meaning the same equipment, the same sequence, and the same key parameters, a family control plan can cover them all. This is particularly common with bulk materials or geometrically similar parts that differ only in size or minor features.2AIAG Blog. Core Tools and IATF 16949:2016 Overview
The catch is that the team must define how much variation between parts is acceptable before grouping them. If two part numbers share a machining process but one requires an additional heat treatment, that added operation needs its own controls, and at that point you’re better off with separate plans. A family plan references the process name rather than individual part numbers and must document the range of specifications that apply across the family. Auditors look closely at these to make sure the “family” label isn’t being stretched to skip the work of writing proper individual plans.
Getting the Plan Approved Through PPAP
The Production Part Approval Process is the formal mechanism for proving to the customer that your process can consistently produce conforming parts. The control plan is one of up to 18 elements submitted as part of the PPAP package.3AIAG. Production Part Approval Process PPAP uses five submission levels that dictate how much documentation the supplier physically sends versus retains on file. At the most common Level 3, the control plan is submitted directly to the customer along with sample parts and supporting data. At Levels 1 and 2, the supplier retains the plan but must make it available on request.
Customer review practices vary by OEM. Some require a formal signature on the control plan before mass production can begin. General Motors explicitly does not require its signature on the plan but expects the supplier to demonstrate through measurement and inspection data that the plan’s requirements are being met.4General Motors Company. IATF 16949 GM Customer Specific Requirements Either way, once the plan is approved, what you documented is what you’re expected to follow for every production run going forward.
When You Have to Update It
A control plan is never finished. IATF 16949 clause 8.5.1.1 requires a review and update whenever any change occurs that affects the product, the manufacturing process, measurement systems, logistics, supply sources, production volume, or the risk analysis in the FMEA. The most common practical triggers:
- Engineering changes, meaning a revised drawing with new dimensions or tighter tolerances.
- Equipment or tooling changes, such as replacing a machine, modifying a fixture, or adding automation to a previously manual operation.
- Material or supplier changes, including switching to a new raw material grade or sourcing from a different supplier even when the specification is nominally the same.
- Quality escapes or warranty claims, since a defective part reaching the customer signals that existing controls were insufficient and the team must add or strengthen the corresponding control.
- Process relocation, whether to a different facility or a different area within the same plant.
- Volume changes, since a significant increase or decrease can affect process stability and may require adjusted sampling frequencies.
Internal audits routinely compare what the control plan says to what’s actually happening on the floor. A plan that specifies hourly SPC checks while operators are checking every four hours is a nonconformity, and a common one. Outdated control plans are one of the fastest paths to a major audit finding.
How Long You Have to Keep It
IATF 16949 requires organizations to retain quality records, including the control plan and its supporting documentation, for the length of time the product is active for production and service requirements plus one calendar year. For parts with long service lives, such as replacement brake rotors or transmission components, that retention period can extend decades beyond the end of regular production. Some OEM customer-specific requirements go further, mandating retention periods of 15 to 20 years regardless of whether the part is still in production.
As teams manage control plans in digital quality management systems, the integrity of those electronic records matters. Audit trails must capture who changed what, when, and why. Version control is essential; an auditor needs to see not just the current plan but every prior revision and the reason it was updated. Organizations that still manage control plans in standalone spreadsheets frequently struggle here, because spreadsheets don’t inherently track revision history or restrict unauthorized edits.
What Happens When Control Plans Fail
The consequences scale fast in automotive. At the supplier level, a major nonconformity on an IATF 16949 audit triggers a formal decertification process starting from the closing meeting of that audit. Losing IATF 16949 certification effectively locks a supplier out of most OEM supply chains; it’s the industry’s minimum entry requirement for doing business.
When quality failures result in safety defects that reach the field, the costs compound. Federal civil penalties for failing to report safety defects or comply with motor vehicle safety standards can reach $27,874 per violation per day, with a cap of approximately $139.4 million for a related series of violations.5Federal Register. Revisions to Civil Penalty Amounts, 2025 NHTSA has pursued penalties in the tens of millions; Ford Motor Company paid $65 million in January 2025 as part of a $165 million total penalty related to untimely recall reporting.6National Highway Traffic Safety Administration. Civil Penalty Settlements Those figures don’t include recall costs themselves, which can run hundreds of dollars per vehicle across millions of units.
A well-maintained control plan with meaningful controls, validated measurement systems, and honest reaction procedures catches most problems before they leave the plant. The gap between what the plan documents and what actually happens on the floor is where risk accumulates.