DMAIC Measure Phase: Building a Trustworthy Data Foundation in Plastics Production

Part 3 of our DMAIC series

A quality improvement project can only be as reliable as the measurements behind it. In the DMAIC Measure phase, teams establish how a process performs today, and check that the data they use to describe it can be trusted.

For plastics manufacturers, this means looking beyond a machine’s data screen or the latest inspection report. It means connecting product and process measurements to the machine, product, shift, and material lot that produced them. It also means checking that measurement tools, sensors, and collection methods can distinguish real process changes from measurement noise.

This is the foundation for a credible baseline. Without it, teams may spend time analyzing variation that is not actually in the process.

Measure what matters to quality and operations

Start with the outcome the project aims to improve. If the goal is to reduce dimensional defects, for example, identify the critical-to-quality characteristics and the process conditions that could affect them. Then define exactly what each measure means, where it comes from, and how often it is collected.

Depending on the product and improvement goal, useful measures may include:

• Machine and process parameters: temperatures, pressure, line speed, throughput, cycle time, and setpoints.
• Product characteristics: dimensions, wall thickness, weight, ovality, and other specification-critical properties.
• Quality losses: scrap, rework, failed inspections, and defect types.
• Operations performance: downtime, stops, changeovers, and production counts.
• Material context: resin type, color, supplier or material lot, and blend or recipe information.

Choose measures that help describe the problem, not every data point the plant can collect. Define units, measurement locations, inspection methods, and defect categories consistently. A “scrap count” that one shift records by rejected part and another estimates by container is not one dependable measure.

Confirm the measurement system before trusting the baseline

Measurement system analysis (MSA) is a practical way to test whether a measurement method is fit for its purpose. It considers more than the gauge itself. The full system may include the instrument or sensor, fixture, software, measurement method, part handling, environment, and the people taking readings.

A simple starting question: Can your team trust that a change in the result reflects a change in the product or process, not a change in how it was measured?

Check resolution, repeatability, and reproducibility

Resolution is the smallest increment a measurement device can show. If its display rounds away changes that matter to the specification, it may not be useful for the job. A more precise display alone is not proof of accuracy, however.

Repeatability asks whether the same person, using the same method and tool, gets similar results when measuring the same part again. Reproducibility asks whether different people can get similar results. A gauge R&R study (repeatability and reproducibility) helps quantify these sources of measurement variation.

For example, if two operators measure the same molded feature and routinely get different readings, their results may obscure a real process shift. The next step may be to clarify the method, improve the fixture, or provide training, not adjust the mold based on uncertain data.

Distinguish calibration from stability

Calibration checks an instrument’s reading against a reference standard. It supports traceability, but a current calibration does not automatically prove that the complete production measurement method is consistent in day-to-day use.

Stability looks at whether results remain consistent over time. A sensor can be calibrated and still drift in operation, or a manual method can change as operators, conditions, or work instructions vary. For critical measures, define how calibration and ongoing verification are handled, and investigate unexpected changes before using the readings to draw conclusions.

In-line measurement can help by capturing product and process information during production, reducing dependence on occasional manual checks. But it is not a substitute for MSA. Validate the sensor, its installation, its data, and its relationship to the product specification.

Capture a baseline that reflects the process

Once the measurement method is understood and sufficiently reliable, collect a baseline over a representative period. Include the normal sources of variation relevant to the project: products or recipes, machines, shifts, operators, and material lots. Record the conditions that could explain differences in results.

Sampling should reflect how the process behaves:

• Extrusion is continuous. In-line sensors can capture ongoing measurements such as dimensions or wall thickness. Teams may combine that stream with scheduled physical checks and additional checks at meaningful events, such as a startup or material change.
• Molding is cyclic. Measurements can be linked to selected cycles, parts, or cavities, with added checks after a tool, material, or process change.

There is no universal frequency that fits every line or product. Set sampling frequency in line with process dynamics, product risk, customer or regulatory requirements, and the team’s control plan. The aim is enough representative data to understand performance without confusing every repeated sensor reading with a separate, independent inspection.

Also decide how the team will handle missing, suspect, or out-of-limit data. A baseline should document the process as it ran, not silently discard inconvenient readings or combine unlike conditions.

Avoid the hidden factory of manual data collection

Clipboards, shift-end estimates, and spreadsheet reconciliation may seem manageable at first. Over time, they can introduce delays, transcription errors, inconsistent categories, and gaps between the reading and its production context. A value entered at the end of a shift may no longer be easy to connect to the machine state, product, or material lot that generated it.

That creates a “hidden factory” of data work: teams spend time collecting, checking, and reconciling records instead of using them to improve the process. It can also distort a baseline. If some shifts record every stop and others record only major ones, a comparison between shifts may reflect different logging practices rather than different performance.

A consistent data foundation helps close this gap. Each measurement should be traceable, where relevant, to its machine, timestamp, shift, product, and material lot. That context makes comparisons more useful and helps quality and operations teams ask better questions when results change.

Build a connected Measure phase with iDOO

At iDOO, we help plastics manufacturers make production data more accessible and useful. Our approach combines measurement technology expertise, industrial data collection, and practical shop-floor context.

iDOO builds on the measurement heritage of iNOEX and the plastics manufacturing experience of the CiTEX Group. Our team works alongside yours to understand the production challenge, identify relevant data sources, and support a measurement approach suited to your process.

The Industrial Data Manager (iDM) collects production data from connected machines and systems, helping teams establish a more consistent view of operations. The iDOO Portal brings relevant production information, KPIs, and analyses together so teams can review data by machine and production context. Where automated quality documentation is needed, the iDOO Product Quality Pass can generate records from production data and compare them with product standards.

These tools do not replace MSA or your quality procedures. They can help reduce manual handling, improve data availability, and make it easier to build a traceable baseline. Explore iDOO’s solutions or read more about connecting production data through open standards.

A quick Measure-phase readiness check

Before moving into Analyze, ask your team:

1. Have we defined exactly what each critical measure means and how it is taken?
2. Have we checked that the gauges, sensors, and methods are suitable and consistent?
3. Can we trace readings to the machine, shift, product, and material lot?
4. Do we have a representative baseline, collected at a frequency suited to the process?

If you cannot answer “yes” to all four, the gap is useful information. It can help prioritize where to strengthen your measurement foundation first.

Want a practical review of your production data and measurement approach? Contact the iDOO team or book a demo. Share the process you are evaluating, the quality objective, and which of the four checks needs attention. We can explore a suitable next step together.

With a trustworthy baseline in place, the DMAIC series moves to Analyze: identifying the process factors that drive variation and defects, then using evidence to focus improvement work.