DMAIC Define Phase: How to Scope a Quality Improvement Project in Plastics

Part 2 of our DMAIC series: turn a broad quality concern into a focused, measurable project.

“Quality is inconsistent” is a useful signal, but not yet a project definition. Which product? Which line or press? What does inconsistent mean, how often does it happen, and how large is the gap between current performance and the quality objective?

In the Define phase of DMAIC: Define, Measure, Analyze, Improve, Control. The team answers those questions before it searches for causes or chooses solutions. That groundwork matters. A project that starts with an unclear problem can spend time improving the wrong process or measuring success in different ways.

For plastics manufacturers, a strong Define phase connects a customer or business need to specific process boundaries, measurable quality characteristics, and a baseline the team can trust.

Turn a complaint into a measurable problem

A useful problem statement describes what is happening, where and when it is happening, how large the gap is, and why it matters. It should report the problem, not guess at its cause.

Start with “quality is inconsistent.” Then ask:

• Which product, product family, line, press, or mold is affected?
• Which defect or characteristic is involved?
• Over what time period?
• How often does it occur, or how far does performance vary?
• What is the impact on scrap, rework, delivery, or the customer?

Example: extrusion scrap

From January through March, Extrusion Line 3 recorded an average scrap rate of 8.0% for Product A, compared with the plant’s 5.0% target. Scrap is measured as rejected material weight divided by total material started. The difference contributes to avoidable material loss and additional handling.

The numbers here are illustrative. Your project should use a consistent, agreed time window and verified plant data.

Example: injection-molding variation

During the last 12 weeks, the cavity-to-cavity range for a critical part dimension on Mold 12 exceeded the approved internal limit in 18% of recorded production runs. This has increased inspection and sorting requirements.

This statement identifies a gap without claiming that the mold, material, machine settings, or operator practice caused it. Those are possible hypotheses for later phases, not conclusions to build into Define.

Write a goal that makes progress testable

A goal statement describes the improvement the team intends to achieve, by when, and within what guardrails. It should be specific enough to measure but should not prescribe the solution.

For the extrusion example:

Reduce the scrap rate on Extrusion Line 3 for Product A from an 8.0% baseline to 5.0% or lower by September 30, while maintaining agreed product specifications and delivery requirements.

For molding:

Reduce the cavity-to-cavity range for the selected critical dimension on Mold 12 by 20% from the verified baseline by September 30, while keeping every cavity’s output within the approved specification.

These are project-goal examples, not promised outcomes. A team should confirm that the target is meaningful and achievable with the available time, resources, and data.

The ASQ overview of DMAIC describes Define as the stage for establishing the project’s purpose, scope, customer requirements, and goals. In practice, a well-written goal gives quality, operations, and leadership a shared reference point throughout the project.

Define the CTQs that matter to customers

CTQ means critical to quality: a measurable product or process characteristic that reflects what the customer needs. Translate broad expectations such as “consistent product” or “good appearance” into clear measures and acceptance criteria.

In plastics manufacturing, CTQs may include:

• Dimensions: diameter, wall thickness, width, length, or flatness, with defined specification limits.
• Weight: part weight or weight variation, measured at an agreed frequency.
• Appearance: a consistent classification for issues such as gels, black specks, flash, short shots, sink marks, scratches, or discoloration.
• Functional properties: performance requirements such as strength, flexibility, fit, or a passed functional test.
• Process-related quality outcomes: scrap, first-pass yield, or rework, when they are directly relevant to the project goal.

For each CTQ, document how it is measured, the unit of measure, the applicable limits, and where the data comes from. “Visual defects” is not yet an operational definition. Teams need a shared classification, ideally supported by examples, so an operator, inspector, and engineer count the same condition in the same way.

Set the scope: machine, product, and time period

A project can be important and still be too broad. Set boundaries that make the work actionable.

• Machine or process: one extrusion line, a defined group of presses, or a specific molding cell.
• Product: one SKU, product family, resin, mold, or tool where the issue has been observed.
• Time period: a stated baseline window long enough to reflect normal production conditions, including relevant shifts or changeovers.

Also record what is out of scope. For example, a project might address one line and product family, while excluding downstream assembly and other sites for now. Clear boundaries prevent the team from unintentionally expanding the project every time a related issue appears.

Scope should be narrow enough to investigate, but not so narrow that it hides the real problem. Compare production across shifts, products, or machines during Define where the available data supports it. A concern that appears to be local may be present across several assets, or may be concentrated in one place.

Choose baseline measures and agree on the data

A baseline makes the size of the problem visible and gives the team a fair way to evaluate change. Choose metrics that match the problem statement and define their calculation before collecting or comparing results.

For example:

• Scrap rate: scrap weight divided by total material started, multiplied by 100.
• Reject rate: rejected parts divided by total parts produced, multiplied by 100.
• First-pass yield: units that pass without rework divided by units entering the process, multiplied by 100.
• CTQ performance: measurements compared with the approved specification, reported with a consistent sampling method.
• Business impact: material loss, rework, inspection effort, or other agreed cost measures.

Identify where each measure comes from: machine or production records, quality checks, inspection systems, or a defined manual log. Confirm that the time periods, denominators, defect categories, and product identifiers are consistent. If the team cannot explain how a number was calculated, it is not ready to serve as a project baseline.

The Define phase is not the time to prove root cause. It is the time to establish a reliable starting point and a practical plan for measurement. If you find that data is incomplete or inconsistent, record that as a gap to address in the Measure phase.

Bring operators and leaders into the definition

Operators see production conditions that may not be visible in summary reports. Quality teams understand customer requirements and inspection methods. Maintenance, process engineering, and leadership each bring a different view of constraints, risk, and value.

Involve these stakeholders while defining the project, not only when presenting results. Ask operators how defects are identified and recorded. Confirm acceptance criteria with quality. Align with leadership on the business objective, project owner, resources, and decision-making cadence.

This collaboration helps the team agree on the problem before the project moves forward. It also makes the scope and objective more practical for the people closest to production.

How iDOO supports a stronger Define phase

Teams need a clear view of production to size a problem and choose a useful scope. The iDOO Portal brings production services, key KPIs, machine status, and analyses together. Depending on the data available and connected at your site, teams can use that transparency to examine where a quality concern is occurring, how performance differs across lines or locations, and which loss categories deserve attention.

A Pareto-style view of loss categories can help narrow a broad concern to the largest contributors. KPI baselines give a quality objective a measurable starting point. Cross-site comparisons can help you assess whether a pattern is concentrated at one machine or appears across multiple operations.

Our Industrial Data Manager supports production data capture and analysis. The Portal can also integrate existing applications and data, helping teams build on their current environment. We bring plastics manufacturing experience through the CiTEX Group, working alongside your quality and operations teams to connect the project question to the realities of production.

Define-phase readiness check

Before moving on, ask your team:

1. Can we describe the quality problem with a product, process, location, and time period?
2. Do we have a measurable baseline and a clear definition of how it is calculated?
3. Have we identified the relevant CTQs and agreed how to assess them?
4. Have operators, quality, operations, and leadership aligned on scope and goal?

Give one point for each “yes.” A score of 0–1 suggests the project needs more definition; 2–3 means there are open questions to resolve; 4 means the team has a solid starting charter. It is a discussion aid, not a formal audit.

If you are planning a quality improvement project, talk with the iDOO team about your project scope, baseline data, or production visibility needs. When you reach out, share your score, the question you are still working through, and the process or product involved. That helps us understand where you are in your improvement journey.

Next in the series: the Measure phase, how to confirm the baseline, check data quality, and understand process performance before analyzing causes.