DMAIC Improve Phase: Testing and Implementing Quality Improvements in Plastics

Part 5 of our DMAIC series: Turn confirmed root causes into tested, repeatable improvements on the shop floor.

Finding a root cause is an important milestone. But it does not, by itself, improve quality. In the Improve phase of DMAIC, quality and operations teams develop countermeasures, test them under controlled conditions, verify results, and prepare the changes for wider use.

For plastics manufacturers, this means improving a process without trading one problem for another. A shorter molding cycle matters only if the part still meets specifications. Less startup scrap is valuable only if the extrusion line reaches stable quality reliably. Improvement must be measurable, and it must hold in production.

Move from root cause to countermeasure

Start with the causes confirmed during Analyze. Generate potential countermeasures that address those causes directly, rather than treating visible symptoms.

For example, if data shows that a process drifts after a material change, possible actions might include standardizing material conditioning, defining a setup sequence, or tightening a critical operating range. If scrap is linked to cavity-to-cavity variation, the response may involve inspecting or balancing the mold, not simply adjusting the machine to improve the average result.

Compare candidate actions against practical criteria:

• Expected effect on critical-to-quality measures
• Impact on cycle time, scrap, material use, and energy
• Ease of testing and implementation
• Safety, maintenance, and downstream considerations
• Risk of introducing variation elsewhere

A simple solution-selection matrix can help teams make the trade-offs visible. Keep operators, process engineers, quality, and maintenance involved: each sees different constraints that can determine whether an improvement will work outside a trial.

Design the trial before changing the process

A production trial is an experiment. Decide what success means before the first setting changes, and record the conditions that could influence the result.

Choose measures that reflect both quality and operational performance. Depending on the project, these might include defect or scrap rate, critical dimensions, part weight, cycle time, startup material loss, throughput, or process capability. Use the same operational definitions and measurement approach used for the baseline. If the measurement system is unreliable or inconsistent, apparent improvement may be measurement noise.

Then select a trial approach that fits the question:

• Change one factor at a time when you need to isolate a straightforward adjustment and the process is relatively well understood.
• Use designed experiments (DOE) when several factors may affect the result or interact. DOE can help reveal those interactions with a planned set of runs, rather than an open-ended sequence of adjustments.

For example, molding results may depend on the combined effect of melt temperature, fill speed, hold pressure, and cooling time. Adjusting one setting at a time may miss an interaction between two of them. A structured experiment can help identify the factors that matter and define a more robust operating window. The U.S. EPA’s overview of DMAIC and Purdue’s Six Sigma Green Belt body of knowledge describe the role of pilot studies and improvement verification within the method.

Pilot on a manageable scale

Begin with a controlled pilot: one machine, mold, cavity, line, or product family, as appropriate. Keep other conditions as consistent as possible, or record them so you can account for their effects. Material lot, machine, shift, ambient conditions, and operator can all matter in plastics processing.

Agree on hold points and decision rules. For instance: pause if a critical dimension moves outside its agreed limit; continue only after the first-piece checks pass; compare results after a defined number of production cycles. These guardrails protect customers and help the team interpret the trial consistently.

A pilot also provides a practical check of implementation. Can operators follow the proposed settings? Are the instructions clear? Does the change require a maintenance or tooling update? A technically effective countermeasure still needs to work in the actual production routine.

Apply the method to plastics production

Optimize a molding cycle while protecting part quality.
Test a reduction in cooling time on a selected machine or mold. Track cycle time alongside relevant quality measures such as dimensions, weight, warpage, and surface defects. Confirm that the proposed setting remains acceptable through representative production, not just a short run.

Reduce startup waste during an extrusion changeover.
Establish a baseline for material discarded and time taken to reach stable quality. Pilot a standardized preparation and startup sequence, including staged materials, defined checks, and documented adjustment steps. Measure the same startup indicators before and after the change.

Stabilize a drifting process window.
Use production data to identify when the drift occurs and which factors change with it. Trial a tighter or revised operating range, then check whether quality remains consistent across relevant materials, products, and operating conditions.

Standardize machine setups across shifts.
Document the settings and checks associated with a proven result. Test whether a second shift can reproduce the outcome using the same setup routine. If results differ, investigate the information, training, or equipment conditions that may be contributing before scaling further.

The goal is not simply to find a setting that works once. It is to find a countermeasure that people can repeat and that produces a verified benefit under normal production conditions.

Verify the improvement against the baseline

After the pilot, compare like with like. Use the same definitions, measures, and sampling logic as the baseline, and include enough production data to understand ordinary variation.

Ask:

• Did the defect or scrap rate actually decrease?
• Did critical product characteristics remain within specification?
• Did process capability improve, where capability is an appropriate measure?
• Was cycle time or changeover time reduced without compromising quality?
• Did the result hold across the selected shifts, cavities, or production runs?

A favorable average is not the whole story. Look at variation and stability as well. Capability indices such as Cp and Cpk are most useful when the measurement system is suitable and the process is stable; they should not be treated as proof of a lasting improvement on their own. If the change improves one metric but harms another, revisit the countermeasure and balance the result against the project’s objectives.

Scale the winning change, with evidence

Once the pilot meets its agreed criteria, plan the rollout. A change that works on one machine or shift may behave differently on another because of equipment, mold condition, product mix, or operator practice.

Scale in steps. Confirm the change on additional cavities, machines, lines, or shifts as relevant. Compare results at each stage and address differences before moving on. Update setup instructions, recipes, quality checks, and training so the new method is clear and accessible to the people who use it.

This is also where improvement knowledge can easily be lost. Record the countermeasure, conditions tested, settings that worked, results observed, and known limitations. That record gives the next team a more reliable starting point than an undocumented “best setting.”

How iDOO supports data-led improvement

Improvement depends on timely, trustworthy feedback. iDOO’s Industrial Data Manager helps capture production data, while the iDOO Portal brings relevant production information, KPIs, and analyses into a shared view. During a trial, teams can use available process data to see how a change affects production without waiting for a long, manual reporting cycle.

Portal comparisons can help teams examine whether gains hold across shifts, products, or production locations, depending on the data available and the way the solution is configured. Digital routines can support consistent rollout, while the Experience-based Assistant helps make practical process knowledge more accessible to teams.

We bring plastics manufacturing experience to that work. Together, we can help your team connect the improvement question to the production data, test a change with clear criteria, and capture what the plant learns.

Ready to improve with confidence?

Use these questions to locate your next step:

1. Have you confirmed the root cause with production evidence?
2. Have you defined the baseline, success measures, and trial conditions?
3. Can you compare results across the machines, cavities, products, or shifts needed to validate the change?

If your root cause is still uncertain, revisit Analyze. If you have a confirmed cause but lack a clear trial plan or timely data, you may be ready to scope an Improve project. If your pilot has delivered a measurable result, prepare to standardize and monitor it.

Talk with iDOO about your improvement challenge or book a demo to explore how production data and guided routines can support your work.

Next in the series: the Control phase, how to sustain improvements, monitor process performance, and respond when results begin to drift. Explore more articles on the iDOO blog.