Lesson 1: Data Collection for Problem Solving: Checksheets, Stratification, and Trend Charts

Setting the Scene: When Gut Feeling Isn’t Enough

It’s Monday morning at a mid-sized automotive components plant. The line supervisor reports that scrap rates have been climbing for the past three weeks, but nobody can pinpoint exactly when defects occur, on which shift, or at which workstation. The maintenance team blames the material; the operators point to the tooling; the quality department suspects process drift. Everyone has an opinion, but nobody has data. This is exactly the situation that Kobetsu Kaizen is designed to resolve — and it starts not with solutions, but with systematic, disciplined data collection. Before any root cause analysis or countermeasure planning can begin, problem solvers must learn to speak with data. The three foundational instruments for this are checksheets (tally charts), stratification, and trend charts.

Learning Objectives

• Explain the purpose and structure of checksheets as the first step in structured data collection for problem solving.
• Apply stratification to separate data by relevant factors and reveal hidden patterns in production problems.
• Construct and interpret trend charts to monitor process behavior over time and identify meaningful changes.
• Connect data collection activities to the first two steps of the Kobetsu Kaizen structured problem-solving approach: Problem Selection and Problem Representation.
• Select the right data collection tool based on the nature of the problem and the type of information needed.

Checksheets: Structured Tally for Structured Thinking

A checksheet — sometimes called a tally chart — is one of the classic Seven Quality Tools (7QT) and serves as the most direct method to record how often specific events occur. Its power lies in its simplicity: a well-designed checksheet turns raw observation into countable, comparable data with minimal effort and minimal risk of misinterpretation. In the context of Kobetsu Kaizen, the checksheet directly supports Step 1 (Problem Selection) and Step 2 (Problem Representation), helping the team understand the current situation before drawing any conclusions.

Designing an effective checksheet requires clear thinking upfront. You must define exactly what phenomenon you are counting, the time intervals for recording, the location or machine involved, and the operator or shift performing the task. A poorly designed checksheet will produce data that looks precise but actually misleads. Consider what categories of defects, stoppages, or quality deviations are relevant to your problem theme, and resist the temptation to record everything — focus matters.

In the Kobetsu Kaizen framework, the tally chart is explicitly listed as a tool applicable to both problem concern and problem cause phases. This dual utility is important: early in the process, you use the checksheet to quantify the magnitude of the problem; later, you use it to verify whether a suspected cause is actually linked to the observed defect frequency. The data collected becomes the factual foundation for the Pareto diagram that typically follows.

Key Design Principles for Checksheets

• Define the data categories before you start collecting — categories should be mutually exclusive and collectively exhaustive.
• Align recording intervals with your process cycle (per shift, per batch, per hour) to ensure data is meaningful at the right granularity.
• Include contextual fields: date, time, machine ID, operator, material lot — these enable stratification later.
• Keep the format visual and quick to complete; operators on the shop floor should be able to record data without interrupting workflow.
• Validate the checksheet with a pilot run before committing to a full data collection period.

Stratification: Separating Signal from Noise

Once you have collected data, stratification is the analytical step that reveals where and under what conditions problems concentrate. Stratification means breaking down aggregated data into meaningful subgroups — by shift, by machine, by material supplier, by operator, by product family — to expose patterns that would otherwise remain hidden in total figures. This technique is at the heart of the “Speak with Data” principle embedded in the Kobetsu Kaizen Board methodology.

Imagine your checksheet shows 480 defects recorded over four weeks. The total number is alarming, but it tells you nothing actionable. When you stratify by shift, you discover that 74% of defects occur on the night shift. When you further stratify by machine within that shift, you find that a single press accounts for 60% of night-shift rejects. Suddenly, a diffuse problem becomes a focused, solvable issue. Stratification does not explain why the problem exists — that is the role of root cause analysis tools like the 5 Why or fishbone diagram — but it dramatically narrows the search space and prevents wasted effort on irrelevant variables.

Stratification is also a prerequisite for meaningful Pareto analysis. Without stratification, your Pareto diagram may show an apparently dominant defect type that is actually the aggregate of several unrelated smaller problems occurring under different conditions. Stratify first, then visualize with Pareto to confirm priorities and guide the team toward the dominant loss category.

Trend Charts: Monitoring Change Over Time

A trend chart (also called a run chart or time-series plot) displays data points in the sequence they were collected, plotted against time. While checksheets capture frequency and stratification identifies where problems concentrate, trend charts answer a different and equally critical question: Is the situation getting better, worse, or staying the same? This is particularly important in Kobetsu Kaizen when tracking progress toward the SMART goals set in Step 3 — goals that are Specific, Measurable, Attractive, Realistic, and Time-limited — with an orientation toward zero defects, zero accidents, and zero errors.

A trend chart makes process drift visible before it becomes a crisis. When plotted consistently, it allows the team to distinguish between random variation (noise) and genuine directional change (signal). Plant managers and team leaders can use trend charts during the daily stand-up at the Kobetsu Kaizen Board to assess whether countermeasures implemented in Step 6 are producing measurable improvements — or whether further analysis and action are required, as part of the Check phase of the PDCA cycle.

Practical Tips for Effective Trend Charts

• Use consistent time intervals on the horizontal axis to ensure the trend is visually accurate and not distorted.
• Add reference lines for your target value and current baseline — this makes progress immediately visible to the whole team.
• Annotate the chart with key events: maintenance interventions, material changes, operator training, process adjustments. This context is invaluable for interpretation.
• Avoid connecting data points with smooth curves that imply precision you do not have; simple line connections between measured points are sufficient and honest.

Practical Example: Precision Parts GmbH

Precision Parts GmbH, a fictional manufacturer of hydraulic valve components, was experiencing a persistent increase in dimensional out-of-tolerance rejections on their CNC turning line. The Kobetsu Kaizen team leader, Maria Schulz, launched a structured data collection initiative. In the first week, the team deployed a checksheet at each of the four CNC machines, recording every rejection by defect type, shift, and machine ID across all three shifts. By the end of week one, the checksheet had captured 312 rejection events with full contextual data.

Stratification revealed that Machine 3 on the afternoon shift accounted for 58% of total rejections, and that the dominant defect type on that machine was bore diameter oversize — representing 71% of Machine 3’s rejections alone. This finding immediately focused the team’s attention. A trend chart built from daily rejection counts on Machine 3 showed a clear upward trend starting six weeks earlier, coinciding — after annotation review — with a tooling batch change from a new supplier. The combination of checksheet data, stratification, and trend chart analysis transformed a vague “quality problem” into a precise, time-bounded, location-specific issue that the team could now address with targeted root cause analysis using the 5 Why method. The data collection phase had