A capability report and a control plan answer different procurement questions. The report can show how a defined process performed against a defined specification at a stated time. The control plan states how the supplier intends to monitor and react to the characteristics and process conditions that matter during production. A buyer should approve neither as a generic quality attachment. First identify the exact part, drawing revision, resin grade, mold and cavity configuration, machine, measurement method, production condition, and acceptance decision to which each document applies.
Requirements are the controlled obligations in the purchase order, quality agreement, drawing, applicable market rules, or customer-specific requirements. Recommendations are practical requests that make a supplier comparison more meaningful when the contract is silent. Capability targets, tolerance coverage, study design, records, and submission level should be agreed for the part and risk; they cannot be inferred reliably from a single molded sample, a company presentation, or a general claim about a tool. The suitable evidence depends on resin behavior, part geometry, tool architecture, machine and automation, market expectations, and the parties’ contract terms.
Approve a production condition, not a number in isolation: a Cpk result is useful only when its specification, measurement system, data basis, process stability, and change status are clear.
01
1. Define the capability question before asking for Cpk
Start with a controlled characteristic list. Mark the dimensions, mass, appearance, assembly interfaces, functional features, material attributes, cavity-specific concerns, and traceability items that actually require evidence. For each item, identify the drawing revision, nominal and specification limits or an unambiguous visual or functional criterion, datum and inspection location, method, unit, sampling condition, and acceptance authority. A capability study cannot resolve an unclear tolerance, an undefined cosmetic standard, or a function that has no agreed test method.
Ask the supplier to state whether the reported index is Cp, Cpk, Pp, Ppk, or another defined measure, and to provide the calculation method rather than only a pass/fail conclusion. Cp describes potential spread relative to bilateral specification width, while Cpk also reflects how centered the process mean is within the limits. NIST frames capability as comparison of an in-control process output with specification limits; data from an unstable process should therefore be investigated before it is used as proof of routine capability. Distribution assumptions, independent observations, outliers, transformations, and one-sided limits also affect whether an index is meaningful.
Do not impose a universal index, sample count, or tolerance for every feature. A buyer may make a project-specific threshold contractual for designated critical-to-quality characteristics, but should document the rationale, the measurement risk, the intended production state, and the consequence of a shortfall. A noncritical nominal dimension, a sealing interface, a safety-related feature, and a high-cosmetic surface may need different evidence. Capability values do not by themselves demonstrate material conformance, product safety, regulatory compliance, functional performance, or future mold life.
02
2. Check whether the data sets are comparable and measurable
When comparing suppliers, reject superficial comparisons such as one Cpk figure against another without context. Request the raw or suitably traceable study summary by characteristic, including individual results or subgroup logic, date and run duration, part and tool revision, cavity identification where relevant, machine and mold identification, resin manufacturer, grade and lot, color or additive condition, resin preparation, operator or automation condition, and any start-up, adjustment, hold, or discard rules. A short supervised run may not represent a longer production condition, especially where thermal equilibrium, cavity variation, material handling, or machine recovery matter.
Verify the measurement system before relying on tight-tolerance capability evidence. The supplier should identify the gauge, fixture, program, resolution, calibration status, part positioning or datum method, environmental controls where relevant, and the result of an appropriate measurement-system assessment. Gauge repeatability and reproducibility studies address errors from instruments and production measurement conditions; a capability value calculated from an unsuitable measurement method can give false confidence or conceal real variation. Align the supplier’s method with the buyer’s incoming-inspection or laboratory method before using small differences as a commercial decision point.
Review variation at the right level. Pooled results can hide an unfavorable cavity, insert, shift, machine, or resin-lot pattern. Request cavity-coded data when a multi-cavity tool, family tool, or cavity-specific risk makes that practical, and ask how data are stratified and how the supplier reacts to a cavity trend. Conversely, do not require a complex statistical study where a characteristic is destructive, functional, categorical, or not suited to normal-distribution assumptions; agree a proportionate functional test, attribute-control method, or alternative analysis instead.
03
3. Request a control plan that can operate on the production floor
A usable control plan links the approved part requirements to the actual molding and downstream route: incoming resin and colorant, drying or conditioning where applicable, mold setup, molding, in-process checks, automated detection, trimming or assembly, final inspection, packing, identification, and shipment. It should name the part and process revision, operation, characteristic or parameter, specification or operating window, measurement or monitoring method, sampling or monitoring frequency, record, responsible role, reaction plan, and escalation or approval route. The plan should be consistent with the supplier’s process-risk analysis; it should not be a generic list of temperatures and visual checks copied between unrelated tools.
Separate requirements from recommendations in the approval package. The buyer should require controls for contractual and product-critical characteristics, required traceability, defined release checks, and any customer-specific submission obligations. The buyer may recommend additional trend charts, safe-launch checks, cavity-level controls, process photographs, or a temporary heightened inspection period when risk, launch history, or supply-chain distance justifies them. If a proposed control is burdensome, ask the supplier to explain the failure mode, detection capability, and alternative evidence rather than deleting it without a technical basis.
The following comparison table can be used in an RFQ quality annex, supplier review, or production-readiness meeting. It is a decision aid, not a universal template; the controlled plan should be tailored to the actual resin, geometry, tool architecture, press, market, and commercial agreement.
| Control-plan element | What the buyer should request or compare | Approval point |
|---|---|---|
| Controlled input | Approved resin grade, permitted colorant or additive condition, material identification and preparation rules; tool and part revision; approved machine or qualified machine family where applicable. | Approve the documented production basis and any limits on substitutions before the capability run or release. |
| Critical characteristics | Link each designated feature to a drawing, functional, appearance, or traceability requirement, inspection method, datum or viewing condition, frequency, and record. | Approve that every contractual or risk-critical requirement has a defined verification route; do not rely on a broad final-inspection statement. |
| Process monitoring | Relevant parameter windows, monitoring source, alarm or trend rule, cavity identification, and the record used to show the run stayed within the approved condition. | Approve only after the supplier explains how a drift, alarm, or out-of-window condition is contained and evaluated. |
| Reaction plan | Stop or hold boundary, material and product segregation, recheck method, disposition authority, corrective-action record, and restart criteria. | Confirm who may release held product and when buyer notification or written deviation approval is required. |
| Change control | Triggers for resin-source, tool, cavity, hot-runner, machine, gauge, automation, parameter-window, subcontractor, or packaging changes; required evidence and notification timing. | Approve the change route in the quality agreement before routine production, including any re-study or resubmission decision. |
04
4. Ask for evidence of control, not an idealised trial
Request a concise production-readiness package that makes the study reconstructable. Useful evidence can include the controlled drawing and specification list; bill of material or resin identification; DFM and tool-revision status; process flow; process-risk analysis or equivalent risk record; control plan; measurement-method evidence; capability or other agreed study; inspection and functional results; cavity map or coding approach; defect and containment history; master samples or approved visual reference where needed; and deviation status. For automotive programmes, a PPAP-type submission may be contractually appropriate. For other sectors, use the documents needed for the product and agreement rather than labeling every package PPAP.
Review whether the capability study was run under conditions intended for production. The supplier should disclose material substitutions, temporary fixtures, atypical manual inspection, reduced cavity count, start-up tuning, rework, excluded observations, or process adjustments made during the run. These facts are not automatically disqualifying, but they can change the decision. Where one correction was made, require a revision-controlled record of the change, its reason, what output was affected, and what evidence is required to re-establish the approved condition.
Compare suppliers against the same input package and decision rule. A lower price or a favorable sample result should not override missing measurement evidence, an undefined reaction plan, or a capability calculation for a different resin or tool revision. Equally, do not treat a detailed document set as proof of manufacturing performance without samples and records tied to the actual production configuration. Ask each supplier to identify open risks, assumptions, exclusions, and actions with owners and due dates.
05
5. Set explicit acceptance gates for launch, routine production, and change
Use separate approval gates. Before the capability or validation run, approve the controlled input package, critical-characteristic list, measurement and sampling approach, and planned evidence. Before initial production release, approve the actual run record, results against the agreed acceptance criteria, open-deviation disposition, control plan, traceability and reaction plan, and any required functional or customer submission evidence. For routine shipments, release should follow the agreed lot controls and records; it should not reopen every development question unless an exception or change occurs.
Write the escalation path before there is a problem. Define which events require immediate containment, buyer notification, written concession, investigation, resampling, capability reassessment, process revalidation, or a new submission. Examples may include an out-of-specification condition, a control-plan reaction, a significant cavity-specific trend, material or tool change, mold repair that could affect a controlled feature, machine transfer, measurement-method change, or a recurring defect. Whether an event is significant must be decided against the product risk and contract terms, not assumed from the fact that the part is injection molded.
Keep approvals revision-specific and time-stamped. The buyer’s signed approval should identify what was approved, the part and tool revisions, applicable material and production basis, evidence reviewed, deviations accepted, expiry or review triggers if any, and the authority granting the approval. This protects both parties from treating an early trial, a prior resin grade, or a verbal meeting note as an unlimited production authorization. It also provides a clear starting point when engineering, supplier management, or quality teams compare later performance with the originally accepted process.
Practical takeaway
Before you move
to the next review.
- Classify each required capability or control-plan item by product risk, state the controlled specification and measurement method, and name the accepting authority.
- Request study context with every capability result: part and tool revision, cavity coverage, machine, resin and preparation condition, run history, data treatment, and measurement-system basis.
- Compare suppliers using the same characteristic list, study conditions, calculation definitions, sampling logic, and release decision rather than isolated Cpk values or sample photographs.
- Approve a control plan only when it defines floor-level monitoring, records, containment, reaction, traceability, and change-control responsibilities for the actual production route.
- Document separate written gates for study readiness, initial production release, routine shipment, deviations, and changes that may require resubmission or renewed evidence.
FAQ
Common questions
What Cpk should a buyer require from an injection molding supplier?
There is no single defensible Cpk requirement for all molded features. Set a project-specific threshold only after defining the characteristic, specification, risk, measurement method, study basis, and production condition. A contractual target may be appropriate for designated critical-to-quality features, but it should not be copied automatically to cosmetic, functional, attribute, or non-normal data. Ask the supplier to show the calculation and context, then record the agreed acceptance rule in the quality agreement or submission plan.
Is a capability report enough to approve molded-part production?
No. A report may support a defined dimensional decision, but approval also needs confirmation of the controlled part and tool revisions, approved resin and production condition, valid measurement method, applicable visual or functional results, traceability, control plan, reaction plan, and disposition of open deviations. The depth of evidence should match the part’s risk, intended market, and contractual requirements.
Should a control plan lock every molding-machine setting?
Not necessarily. A plan should control the inputs and conditions that are material to product conformity, but the right format may be approved windows, validated recipes, monitored trends, alarms, or other defined controls rather than a single value for every setting. The supplier should explain the parameter source, permitted adjustment authority, record, and reaction if the process leaves the approved condition. Resin instructions and equipment limits still apply.
When should a supplier repeat capability evidence or seek buyer approval after a change?
Decide this in advance through change control. A change to resin source or grade, tool geometry or repair, cavity condition, hot-runner setup, machine, automation, measurement method, process window, or subcontracted operation can warrant review if it could affect a controlled requirement. The supplier should notify the buyer according to the agreement, assess the affected characteristics, and propose proportionate evidence. Written buyer approval should be required where the contract, customer requirement, or risk assessment says so.
References
Source material
- [1] National Institute of Standards and Technology (NIST), “What is Process Capability?”
- [2] National Institute of Standards and Technology (NIST), “Gauge R & R Studies”
- [3] Automotive Industry Action Group (AIAG), “Control Plan”
- [4] Automotive Industry Action Group (AIAG), “Production Part Approval Process”
This guide is general information for early procurement and engineering discussion. It does not replace product-specific DFM, qualification, validation, regulatory assessment or legal advice.