Engineering teams routinely perfect a process in a controlled environment, dialling in parameters until they produce a small sample of perfect parts. The resulting Cpk values are flawless, the control charts look textbook-ready, and management receives a presentation with green arrows pointing upward. Everyone shakes hands and signs off the submission.
Three weeks later, the production line is down and scrap is pouring out of the machine. Operators are standing idle because the tooling has failed, the fixtures have loosened, or the material has shifted out of specification. The quality engineer who certified the process has moved to another project, leaving the production team to manage the fallout.
The gap between the pristine sample run and the production disaster is simple. Nobody asked the process to prove it could survive a real production shift. Nobody mandated a structured Run@Rate at actual cycle times, with trained operators, using real material from standard supply chains.
Sample Runs Prove Possibility, Not Probability
A sample run proves that conforming parts are physically possible under controlled conditions. It tells you nothing about whether the process can reliably produce those parts over an eight-hour shift. Run@Rate exists to prove probability. It is the only valid mechanism for transitioning a process from the engineering lab to the shop floor.
I have audited plants that submitted 30 perfect sample parts for a PPAP package, only to record a 40 percent scrap rate during serial production. The process had never been tested at actual cycle time. Thermal drift that was entirely invisible across 30 parts became catastrophic across 2,000. The fixture held tolerance for the initial trial but loosened after 200 production cycles.
This is the core difference between a capability study and a Production Trial Run. The capability study isolates variables to measure mathematical potential. The Run@Rate introduces the chaotic variables of real manufacturing to measure actual reliability. Skipping the Run@Rate means launching on unverified assumptions.
The Competence Assumption and Material Reality
When process engineers run a trial, they bring a deep understanding of machine parameters and the patience to diagnose issues in real time. Production operators, who may have received two hours of training on a Thursday, do not have this context. If a process only yields conforming parts when an engineer operates it, you have a science experiment, not a manufacturing process.
Material variation is equally predictable and routinely ignored. A plastics molder might run initial samples using virgin material, only to process regrind blends during serial production. A 0.3 percent shift in shrinkage rate is barely measurable in a lab, but it will push every critical dimension out of tolerance over a four-hour production shift.
Run@Rate mandates the use of production-representative material received through standard channels. It also dictates that production operators run the machinery. Engineers must observe and monitor, but the hands on the controls must belong to the personnel who will run the line when the auditors leave.
Sample Run vs. Run@Rate Conditions
Sample Run Conditions
- Hand-picked prime material
- Process run by engineers
- CMM in a temperature-controlled lab
- Short cycle times with adjustment pauses
Run@Rate Conditions
- Standard supply chain material lots
- Process run by trained operators
- Shop-floor gauges at production frequency
- Sustained production cycle times
Phase 1: Preparation and Documentation Discipline
A structured Run@Rate begins with rigid documentation review. Quality teams must verify the control plan, process flow diagram, PFMEA, work instructions, and MSA results. If a single mandatory document is missing or incomplete, the Run@Rate does not proceed. This is not administrative bureaucracy; it is launch discipline.
Equipment verification is equally non-negotiable. Maintenance records are checked and calibration status is confirmed. Tool life expectations must be documented and respected. If a mould has a planned life of 50,000 shots and the current count is 48,000, you replace the tool before the trial. Launching on exhausted tooling invalidates the data.
Finally, personnel readiness must be validated. Training records are reviewed, but attendance at a classroom session does not equal competence. Operators must demonstrate the ability to run the process and produce conforming parts under supervision before the Run@Rate execution begins.
Phase 2: Execution at Actual Cycle Times
The execution phase demands sustained production rhythm. In IATF 16949 environments, this typically means running for a minimum of one full shift, producing at least 300 pieces, or completing a defined multiple of the production batch size. The run must be long enough to force the process to reveal its true behaviour, not its optimised best behaviour.
The process must run at the declared cycle time. There can be no pauses for measurement or adjustment. If the declared cycle time is 45 seconds, the machine produces one part every 45 seconds. If the system cannot sustain this rate while yielding conforming parts, the Run@Rate fails immediately. No exceptions are permitted.
Every event during the run must be logged chronologically. A Run@Rate log that states the process ran 300 conforming parts tells you very little. A log that records 300 parts, four clamp pressure adjustments, a tool change at part 180, and two machine restarts tells you the process is entirely unstable, even if the final parts passed inspection.
Run@Rate Execution Sequence
- 01Documentation LockControl plan, PFMEA, and work instructions verified and frozen.
- 02Equipment & Tooling CheckCalibration confirmed and remaining tool life verified.
- 03Production RunOperators run the process at declared cycle time for a full shift.
- 04Data & Log ReviewCpk calculated and chronological event log analysed for hidden instability.
Phase 3: Statistical Evaluation and Corrective Action
Data from the run is analysed for process capability and statistical stability. Control charts are reviewed for trends, shifts, or runs that indicate the process is out of control. A process can produce 100 percent conforming parts during a trial and still be statistically unstable. Stability is the prerequisite for capability.
If a process only passes when measured in a lab, but fails with a shop-floor gauge, you have a measurement gap that will swallow your quality system.
Any nonconformance triggers a structured 8D corrective action. The root cause is identified, a correction is implemented, and effectiveness is verified. If the corrective action requires a parameter adjustment or a fixture modification, the Run@Rate must be repeated in whole or in part. This is the definition of launch rigour.
Measurement during this phase must strictly follow the control plan using production systems. Parts are measured by operators using shop-floor gauges at the specified frequency. If your process only passes when measured on a coordinate measuring machine in a temperature-controlled lab, the measurement gap will swallow your quality system once production starts.
Extending Run@Rate Discipline to the Supply Chain
The principle of Run@Rate must extend to your suppliers. Too many organisations accept supplier PPAP packages at face value because the paperwork is complete and the sample parts conform. Nobody visited the supplier to watch the process run at rate, and nobody verified whether the operators were actually trained.
Supplier Run@Rate witnessing is a critical quality tool. It does not require an engineering team for every new component. It requires a risk-based approach: critical components get on-site witnessing, standard components get remote review with video evidence, and commodity components get self-certification with audit-based verification.
Digital Industry 4.0 systems can enhance this process. Real-time monitoring captures machine parameters at millisecond intervals, creating a permanent digital fingerprint of the launch. However, technology amplifies bad practices just as effectively as good ones. A digital Run@Rate with untrained operators is simply a faster way to generate the wrong answer.
Treating Run@Rate Failures as Risk Reduction
A Run@Rate failure is not a setback; it is a risk eliminated. When a process fails during a trial, it reveals a problem that would have caused production scrap, customer complaints, and warranty claims. The trial contains the failure in a controlled environment where it can be studied and fixed before it reaches the customer.
Schedule pressure is real, but launching an unproven process is mathematically disastrous. A late launch costs weeks. A failed launch costs months—containing the crisis, fixing the unverified process, and attempting to recover lost customer trust. Run@Rate discipline always pays for itself in prevented firefighting.
Excellent organisations protect the integrity of this process. When plant management demands a launch regardless of readiness, the quality leader must have the authority to stop it. If the culture punishes operators for surfacing issues during the Run@Rate, those issues will surface later in production, where the financial and reputational consequences are exponentially higher.