How to reduce injection molding scrap without major capital investment.
What appears to be a scrap problem is often a variation problem somewhere else in the molding system.
Scrap rarely has a single origin. It can come from process variation, tooling condition, startup and restart practices, material condition and handling, machine capability, part or mold design, uncontrolled process changes, differences in technician and operator practice, or quality requirements that the current process cannot reliably meet. Several of these usually interact on the same part.
Some recurring scrap can be reduced without major capital spending once its source is correctly identified. That makes the first question not “what equipment should we buy?” but “where is the scrap actually being created, and what condition is creating it?”
A method that works on the floor.
Do not attack scrap as one plant-wide percentage. Break the loss down until the operation can see where it originates and which conditions correlate with it.
- 01MeasureWhere is the loss occurring?
- 02StratifyPart, tool, machine, defect, shift, condition
- 03ParetoWhat is driving the largest loss?
- 04InvestigateWhat changed, or what differs?
- 05CorrectAddress the verified cause
- 06StandardizeEstablish the expected condition
- 07VerifyDid the loss decline and stay controlled?
Measure scrap where it is created.
Plant-level scrap percentage is a reasonable business KPI, but it is not sufficient for root-cause work. Where practical, stratify the loss by part number, mold, machine, defect or reason, shift, material and material lot where relevant, and startup versus steady-state production — plus any known process condition or change. Not every plant needs every field; the right level of detail depends on the operation and the problem in front of it.
Keep reason codes short enough that they are used consistently. A code set with excessive complexity, or one applied inconsistently across shifts, produces data that looks precise but cannot support a decision.
Find the few combinations driving the loss.
Once the data is stratified, the operation can ask useful questions: which parts generate the most scrap dollars or units; which molds keep appearing; whether one machine behaves differently from others running comparable work; which defect categories dominate; whether the loss concentrates on a shift; whether it is startup scrap or steady-state scrap; and whether it began after a tooling, material, maintenance, or process change.
The objective is not another chart. It is to move from “we have a scrap problem” to “this defect is concentrated on this part-and-tool combination under these conditions,” which is something an engineer or technician can actually investigate.
Separate startup scrap from steady-state scrap.
Mold changes, material changes, restarts after downtime, and process stabilization can all generate scrap, and they should be measured separately when they are material to the total. What matters is determining whether the loss occurs during startup, after a changeover, after a restart, intermittently during the run, or continuously in steady state. That distinction changes the investigation entirely.
Where startup loss is significant, useful controls include documented startup procedures, verified setpoints, material readiness, mold temperature stabilization, first-piece or first-article verification where appropriate, and defined criteria for releasing the process to production.
Verify the process before adjusting it.
If three technicians run the same tool three different ways, the plant is running three processes. Uncontrolled adjustment also hides the original source of variation: by the time the part is acceptable again, nobody can say what actually moved.
The discipline is straightforward — established and documented process settings, process sheets that reflect what is actually run, defined process windows where they are appropriate, controlled changes, and a record of significant adjustments. Before touching a setpoint, confirm the process has in fact moved and identify what moved it. Scientific molding fundamentals help here, but the objective is repeatability, not adherence to any one methodology.
Treat tooling condition as a process input.
Worn vents, damaged shutoffs, inconsistent cooling, deteriorating mold condition, or other tooling issues can reduce process capability and create variation that processing adjustments alone cannot reliably correct. Relevant conditions include vents, shutoffs, cooling circuits and water flow, wear surfaces, slides and lifters where applicable, hot runner performance where applicable, mold temperature consistency, and known repair history.
A tooling history connects defects to interventions, repairs, and recurring conditions. That record lets maintenance be planned against known condition and history rather than against failure. Cycle counts can be part of that picture, but they are not by themselves the right trigger for every tool.
Control material to the resin's actual requirements.
Material requirements depend on the resin and grade. For hygroscopic materials, the relevant controls typically include drying temperature, dew point, drying and residence time, airflow, and exposure after drying — set to the supplier's requirements for that material and verified against the actual process, not assumed from another job.
Independent of drying, most operations benefit from material identification, contamination prevention, defined and consistent regrind percentage, color and additive handling, lot traceability where required, and a controlled procedure for material changes at the press. Defects from these sources routinely get diagnosed as processing problems.
Do not rule out the machine — and do not start there.
Not every scrap issue is process, tooling, or material related. Where evidence points to the press, shot-to-shot repeatability, temperature control, hydraulic or electric drive performance, check-ring behavior, clamp performance, and controller or sensor condition are all worth verifying. The point is not to buy a machine because scrap exists, and not to assume equipment condition is never the cause. Diagnosis decides that.
Make the improvement repeatable.
An improvement is more likely to be sustained when the operation can measure the loss, identify its source, correct the underlying condition, establish the operating standard, assign ownership for it, and detect when performance starts drifting away from that standard.
A one-time correction without monitoring leaves the plant blind to a return of the same condition — or to a different cause producing the same defect.
Scrap is rarely isolated.
Scrap consumes machine time and material. Rework consumes labor and capacity. Containment can consume inspection resources. Repeated process intervention can destabilize a running job, and schedule disruption can create additional changeovers and startup losses. Depending on the operation, the same defect can touch cycle performance, uptime, tooling, labor, scheduling, quality, delivery, and material usage.
That system-level view is the subject of where profit hides in a manufacturing operation, and of why profits hide on the production floor.