A quality manager called me recently, two weeks out from a VDA 6.3 audit. His team had only ever done ISO 9001 system audits, and a customer was sending three auditors for a two-day assessment. He assumed it would follow the familiar pattern: show the procedures, demonstrate compliance, get the certificate.

It does not work that way. ISO 9001 asks whether you have a procedure. VDA 6.3 asks how your process actually runs and demands you prove it at the station. The difference is between asking whether speed limits exist and asking why you are currently doing 50 in a 30 zone, how you know that is the correct speed, and what happens when you exceed it.

VDA 6.3 is the process audit standard issued by Verband der Automobilindustrie. It is a de facto entry requirement for supplying BMW, Mercedes, VW, Porsche, and Audi. Without a successful assessment, you simply do not ship parts. The standard organises your manufacturing reality into six process elements, P1 through P6, and grades the results with brutal arithmetic.

P1 to P3: Potential, Planning, and Preparation

P1 (Potential Analysis) happens before you even start production. The customer asks whether you have the QMS infrastructure, personnel capacity, equipment, and logistics to deliver what you promise. Many suppliers are surprised to be audited before a single part is made. P1 is a rapid filter that decides whether the customer will even begin the onboarding process with you.

If P1 passes, P2 (Process Planning) audits how you designed the manufacturing process for the specific product. The auditor looks for your process flow chart, PFMEA, and control plan. Critical characteristics must be identified and flowed down. In one mock audit I ran, the engineering team had copied a control plan from a different product family without transferring the special characteristics. That copied document immediately triggered a zero-score finding.

P3 (Process Preparation) shifts to physical reality. You have a plan, but did you actually execute it? The auditor checks equipment installation and qualification, personnel training records, and gauge calibration. I have audited plants with pristine calibration databases where the operator on the line was measuring a critical dimension with an instrument that expired three months earlier. P3 accepts evidence, never explanations.

Process audits live or die at the station: documentation means nothing if the operator cannot demonstrate the reaction plan.
Process audits live or die at the station: documentation means nothing if the operator cannot demonstrate the reaction plan.

P4 to P6: Execution, Supply Chain, and Feedback

P4 (Process Realisation) is where the audit gets physical. The auditor stands at the line, watches the machine parameters, and questions the operator. They verify SPC data, check whether results fall within specification, and test deviation handling. When the auditor asks what happens when a parameter breaches the upper limit, the operator must state the standardised reaction. Answering "I call the supervisor" instead of containment and adjustment drops your score immediately.

P5 covers supplier management. The auditor does not just evaluate your process; they evaluate how you control your sub-tier. You need documented selection processes, incoming quality monitoring, and supplier claim management. I recall an injection moulding supplier with excellent internal processes whose resin provider changed material lots without notification. The result was sporadic surface defects that the customer only caught at final assembly. The VDA auditor found this gap in 45 minutes by tracing incoming lot traceability.

P6 (Feedback and Improvement) looks at what happens after the product leaves your dock. This section assesses customer satisfaction monitoring, warranty data analysis, and your CAPA system. Teams often underestimate this element, assuming that passing inspection means the job is done. VDA 6.3 takes a longer view, evaluating whether your organisation systematically learns from field failures and drives continuous improvement through structured lessons learned.

The Scoring Mechanism and Star Questions

VDA 6.3 uses a specific scoring matrix. Each question is rated 10 points (fully compliant, evidence is clear), 4 to 6 points (partially fulfilled, gaps exist), or 0 points (no evidence or systemic failure). This quantitative approach removes subjectivity and forces auditors to base ratings on what they can verify at the gemba.

VDA 6.3 Performance Classifications

≥90%Grade ACompliant. Clear evidence and systematic, sustainable processes.
75-89%Grade BConditionally compliant. Corrective actions required and verified.
<75%Grade CNon-compliant. New business on hold; current deliveries at risk.
Composite scores determine your supplier status. Below 75% triggers delivery suspension risk across German OEM contracts.

The scoring mechanic that catches most teams off guard is the star question. Certain critical questions are marked with an asterisk. If you score zero on any star question, the entire process element automatically fails, regardless of how well you performed on the other questions. You can have full compliance on every other item, but miss a star question and that chapter drops entirely.

Star questions are the brakes of the VDA 6.3 audit: miss one, and the whole process element fails regardless of your other scores.

If you land at 89 percent, the specific location of the gap matters. A documentation gap in P2 can be closed with a revision. A P4 execution gap means the customer may send a resident engineer to your floor within two weeks to monitor your production. The score is just a number; the risk lies in where you lost the points.

How to Prepare: Rules from the Floor

Do not stage the floor for audit day. I have seen plants print procedures the night before and rush calibration checks in a panic. Experienced auditors spot this within ten minutes of walking the line. They read the machine counters, check the actual wear on the fixtures, and talk to the operators. If your standard work does not match the reality on the line, the audit will expose the disconnect.

Prepare your operators, not just your managers. The most common failure mode is a quality manager who can answer every question while the line operator cannot identify the product's special characteristics or state the reaction plan. VDA 6.3 questions the people doing the work. If the operator does not know what to do when a parameter drifts, your documented control plan is functionally useless.

Mock Audit Preparation Cycle

  1. 01Document ReviewVerify PFMEA, control plan, and flow chart alignment offline.
  2. 02Gemba WalkTest operator knowledge of special characteristics and reaction plans.
  3. 03Traceability CheckTrace a finished part back to the raw material lot.
  4. 04Gap ClosureAssign 8D owners to every zero-score finding discovered.
Running an internal assessment against the actual VDA 6.3 questionnaire exposes gaps months before the customer arrives.

Run a full internal mock audit annually using the actual VDA methodology. This means assessing the process against the exact P1-P6 questionnaire, not a generic ISO checklist. You will find the gaps before the customer does. Treat the audit as a continuous standard, not a one-time event. Customers return every twelve months, and they expect to see measurable improvement in your scores and your processes.

Executing the Turnaround

I ran a two-day mock audit for a supplier expecting their first customer assessment. The initial result was 68 percent, Grade C. The management team was shocked, but to their credit, they made no excuses. They accepted the findings and treated the two weeks before the real audit as a hard deadline to fix the systemic issues driving the low score.

In fourteen days, they rewrote the control plan to reflect actual line conditions rather than a generic template. They qualified all measuring instruments on the shop floor, trained every operator on special characteristics, and established material traceability back to the supplier lot. They also revised the PFMEA to address the actual failure modes the mock audit had uncovered rather than theoretical risks copied from a textbook.

The actual customer audit scored 86 percent, Grade B. It was conditionally compliant, but the customer recognised the speed and depth of the corrective actions. Two months later, after verification of the implemented fixes, the supplier achieved 91 percent, Grade A. More importantly, their scrap rate dropped and their production stabilised. The audit did not just generate paperwork; it forced the operational discipline that improved the line.

VDA 6.3 in an Industry 4.0 Environment

Process audits remain entirely relevant in operations running digital twins, predictive analytics, and automated inspection. Technology replaces tools, not thinking. VDA 6.3 does not evaluate which sensor you use; it evaluates whether you understand your process risk and manage it systematically. That requirement applies identically to a manual assembly cell and to a fully automated Industry 4.0 production line.

Digital data does not automatically equal process quality. You can stream SPC data to the cloud in real time, but if nobody interprets the trend or triggers a containment action when Cpk drifts below 1.33, the monitoring system is just an expensive dashboard. The human factor remains critical. VDA 6.3 emphasises competence, responsibility, and quality culture. No software platform replaces an operator who understands why a specific parameter is critical to the final product.

VDA 6.3 is a mirror. It shows an organisation exactly how it operates, not how it believes it operates. For hundreds of suppliers in the automotive chain, it is an annual test of operational maturity. For quality professionals, it reinforces one fundamental truth. Quality is not what you write in the procedure. Quality is what happens on the line when nobody is watching.