Practical Case Study of the Seven QC Tools: A Complete Path from Workshop Data to Improvement Breakthrough

1. Introduction: A Real Improvement Story

In March 2025, a certain automotive component company, Huacheng Precision Engineering (a pseudonym), faced a severe quality crisis on one of its chassis structural welding production lines.

This production line supplied the rear subframe assembly for a joint venture brand SUV, involving 36 welding points and 4 weld seams, with a daily capacity of 320 units. Since the beginning of the year, the first pass yield (FPY) in post-weld inspection had been continuously declining, from an initial 94% to 82%. The customer's PPM (parts per million) rate soared from 800 to 3200, directly triggering a Supplier Corrective Action (SCA) warning.

The Quality Manager convened an improvement team, which included a welding engineer, a production team leader, a quality inspection team leader, and a newly certified Green Belt improvement facilitator, and required them to identify the root cause and propose an improvement plan within three weeks.

"We don't have expensive inspection equipment or Six Sigma Black Belts, but we have one thing—the Seven QC Tools," said Mr. Zhang, the improvement facilitator, at the kick-off meeting.

Over the next three weeks, this frontline team, armed with the check sheet, stratification, Pareto chart, cause-and-effect diagram, scatter diagram, histogram, and control chart, completed a full improvement loop from data collection to root cause identification, from solution validation to effect solidification.

This article will use this case as the main thread to analyze the practical application scenarios, operational points, and output results of the Seven QC Tools. It is not just a user manual for the tools but also a replicable improvement path.

2. Tool One: Check Sheet—Transforming "Feelings" into "Data"

2.1 Problem Background

At the start of the improvement initiative, the team faced its first challenge: everyone said, "There are many welding defects," but no one could clearly specify the types of defects, their proportions, the workstations where they occurred, or the time periods.

The quality inspector conducted a visual and tool inspection of 20 units daily at the final inspection station, marking "defective" when a defect was found. By the end of the week, the only information left was "the defect rate is about 18%." Such data was insufficient to guide the improvement direction.

2.2 Design and Implementation of the Check Sheet

Mr. Zhang guided the team to design a stratified check sheet, categorizing defects by location and type (Table 1).

Defect Record Check Sheet (Excerpt)

Implementation Points: Full inspection replaced sampling, with each welded piece inspected immediately upon completion. Each shift used one sheet, and the quality inspector recorded "√" marks in real-time. Data was collected continuously for 5 working days, totaling 1,532 full inspections.

2.3 Practical Value of the Check Sheet

The value of the check sheet lies not in "recording" but in converting vague quality issues into structured data. When the team obtained the complete records for 5 days, the vague perception of "many welding defects" became clear numbers: "30 instances of excessive spatter, 20 instances of lack of fusion, 12 instances of porosity, etc." This provided the first cornerstone for subsequent analysis.

Key Points Summary:

• The design of the check sheet must first define the "stratification dimensions"—by defect location, type, shift, equipment, etc.
• Full inspection is preferred over sampling; the larger the data volume, the higher the reliability of subsequent analysis.
• Frontline operators directly participate in recording to avoid the chain of "data entry → transmission → distortion."

3. Tool Two and Three: Stratification and Pareto Chart—Identifying the "Vital Few"

3.1 Stratification: Discovering Hidden Patterns in Data

After collecting the data, the team did not rush to create charts but first performed stratification analysis. Mr. Zhang stratified the data by three dimensions:

Stratification One: By Shift

• Day Shift: Total defects 38, defect rate 9.1%
• Night Shift: Total defects 76, defect rate 18.3%

The defect rate for the night shift was twice that of the day shift—this discovery gave the team its first direction.

Stratification Two: By Workstation

• Workstation 1 (Robot Welding): 24 defects
• Workstation 2 (Manual Spot Welding): 62 defects
• Workstation 3 (Manual Positioning Spot Welding): 28 defects

Stratification Three: By Operator The defect data for 12 operators during the week, including both day and night shifts, was separately analyzed. It was found that the defect rates for two night shift operators were 2.1 times and 2.7 times higher than those of their counterparts on the day shift.

3.2 Pareto Chart: Focusing on the Most Critical Defect Types

Based on the aggregated data from the check sheet, the team created a Pareto chart. The defects were ranked by frequency:

The top three defects (excessive spatter, lack of fusion, porosity) accounted for 55.9% of the total defects, forming the "vital few"—solving these three issues would eliminate more than half of the defects.

3.3 Combined Effect of the Two Tools

The combination of stratification and the Pareto chart is one of the most practical pairings in the Seven QC Tools. Stratification helps the team identify "where the differences lie," while the Pareto chart helps the team lock onto "what to prioritize."

In this case, the team concluded that the excessive spatter and lack of fusion at the night shift manual spot welding workstation (Workstation 2) were the top priorities for improvement. This conclusion narrowed the improvement scope from "36 welding points and 4 weld seams across the entire line" to "one workstation, two operators, and two defect types"—reducing the scope by over 80%.

4. Tool Four: Cause-and-Effect Diagram—Systematically Digging for Root Causes

4.1 Organizing Cause-and-Effect Diagram Analysis

After identifying the target of "excessive spatter and lack of fusion at the night shift Workstation 2," the team held a field cause-and-effect diagram discussion meeting. Participants included the welding engineer, equipment maintenance personnel, night shift team leader, operator representatives, and quality inspectors.

Mr. Zhang drew a large "fishbone" on the whiteboard, with the fish head pointing to "excessive spatter and lack of fusion at the night shift Workstation 2." The fishbone was divided into five major categories: Man, Machine, Material, Method, and Environment.

After a 2-hour brainstorming session and on-site verification, the team listed approximately 30 potential causes. Through on-site confirmation and rapid validation, they ultimately identified 6 key causes:

4.2 Practical Points of the Cause-and-Effect Diagram

The value of the cause-and-effect diagram lies not in "drawing the diagram" but in:

1. Promoting Cross-Functional Collaboration—people from different roles contribute causes from their perspectives
2. Preventing Omissions—using a structured framework (Man, Machine, Material, Method, Environment) to reduce blind spots
3. Establishing a Cause-and-Effect Logical Chain—from the phenomenon to the direct cause and then to the root cause
4. Directly Outputting Improvement Topics—each confirmed cause can be converted into an improvement action item

5. Tool Five: Scatter Diagram—Verifying Cause-and-Effect Relationships

5.1 Speaking with Data

After identifying the 6 key causes, the team faced a critical question: Is there a statistically significant correlation between these causes and "excessive spatter" and "lack of fusion"?

Taking "insufficient protective gas flow" as an example, the welding engineer proposed that when the gas flow rate is below 12L/min, the protection effect of the molten pool decreases, leading to an increase in porosity and spatter. However, this judgment was based on experience and needed data validation.

The team collected gas flow records and corresponding defect rates for each shift over the past week and created a scatter diagram. The x-axis represented the protective gas flow rate (L/min), and the y-axis represented the defect rate for excessive spatter in that shift.

The scatter diagram showed: when the gas flow rate was between 12-15L/min, the defect rate for excessive spatter remained low at 3-5%; when the flow rate dropped to 9-11L/min, the defect rate for excessive spatter sharply increased to 8-15%; when the flow rate exceeded 16L/min, the defect rate for excessive spatter also slightly increased (due to turbulence affecting the molten pool).

The data clearly presented a "U-shaped" relationship—optimal flow rates were between 12-15L/min, with both too low and too high flow rates leading to increased defect rates.

5.2 Judging the Scatter Diagram

The practical judgment of the scatter diagram does not rely on complex correlation coefficient calculations. The team used the most intuitive "five-point judgment method":

• Positive Correlation: As X increases, Y increases → e.g., the degree of wear on the wire feeding wheel and the frequency of spatter
• Negative Correlation: As X increases, Y decreases → e.g., the number of months of operator experience and the defect rate
• Non-linear Correlation: e.g., the U-shaped relationship between gas flow rate and defect rate
• No Correlation: Points are randomly distributed on the chart → eliminate non-correlated factors
• Stratification Anomaly: Data naturally forms two clusters, indicating the presence of hidden stratification variables

In this case, the team confirmed through the scatter diagram that "protective gas flow rate" and "wire feeding mechanism condition" had a significant correlation with welding defects, while the variability in the thickness of the galvanized layer between batches, although present, had a weaker correlation with the current line defects (it might impact downstream processes) and thus was not a priority for this improvement.

6. Tool Six and Seven: Histogram and Control Chart—Evaluating Process Capability and Stability

6.1 Histogram: Understanding the Distribution

Before implementing improvements, the team needed to answer a fundamental question with data: How much variability does the current process have?

The team randomly selected 100 welded pieces from the production line and measured the critical dimension—the welding positioning dimension X (standard value 50±0.5mm)—and created a histogram.

The histogram showed: the data exhibited a "bimodal" distribution, with one peak centered at 49.8mm and another at 50.3mm, and a clear "valley" between the two peaks. This distribution pattern indicated that what appeared to be a single process actually had two different process states.

Combining stratification, the team found that the welding positioning dimension for the day shift was concentrated around 49.8mm, while the night shift was concentrated around 50.3mm. This suggested a systematic deviation in the adjustment of the welding positioning fixture between the two shifts. This finding further validated the cause of "operators lacking experience in adjusting welding parameters" identified in the cause-and-effect diagram.

6.2 Control Chart: Assessing Stability

Before implementing improvements, the team conducted a 5-day continuous control chart monitoring of the key quality characteristic—weld penetration depth. Each day, 5 units were sampled from each shift, and an Xbar-R chart was used.

The control chart for the first 3 days showed: the R chart (range chart) was within control limits, but the Xbar chart (mean chart) showed that the night shift data points consistently fell above the mean line, and on the 4th and 5th days, the night shift data points exceeded the upper control limit (UCL).

According to the criteria for identifying out-of-control conditions: the presence of 7 consecutive points on one side (above the mean line) constitutes a "run" out-of-control condition, indicating a systematic shift in the process mean. Exceeding the control limit further indicates that the shift has become unacceptable.

The combined use of the histogram and control chart led the team to two key conclusions:

1. The Process is Unstable—the night shift has a systematic deviation in welding parameter control.
2. The Process Capability is Insufficient—even the day shift data has a Cpk of only 0.87, below the industry benchmark of 1.33.

These conclusions provided a quantitative baseline for the subsequent improvement plan—improvement was not just about "reducing defect rates" but also about "controlling the process and meeting capability standards."

7. Implementation and Effect Verification of Improvements

7.1 List of Improvement Measures

Based on the 6 key causes identified using the Seven QC Tools, the team formulated and implemented the following improvement measures:

7.2 Effect Verification

After implementing the improvements, the team continuously tracked data for 4 weeks:

Week 1 (Implementation Period): The defect rate decreased from 18% to 9.2%, primarily due to the repair of the wire feeding mechanism and the improvement in lighting.

Week 2 (After Parameter Standardization): The defect rate further decreased to 4.5%, with the effects of standardized welding parameters beginning to show.

Week 3 (Stabilization Period): The defect rate stabilized at 2.1%, and the difference between the night shift and the day shift was reduced to less than 0.5 percentage points.

Week 4 (Consolidation Period): The defect rate remained around 1.5%, and the process Cpk improved from 0.87 to 1.42.

The control chart showed: the night shift deviation on the Xbar chart disappeared, and all data points were randomly distributed around the center line, indicating that the process had entered a statistically controlled state.

7.3 Consolidation and Promotion of Improvements

Three months later, a review showed that the monthly average defect rate for this production line stabilized between 1.3% and 1.8%, a decrease of 87% compared to before the improvements. Annual quality losses were reduced from 460,000 yuan to 86,000 yuan, a reduction of about 81%.

The more significant gain was that the improvement team solidified the use of the Seven QC Tools into a "four-step standardized improvement process":

1. Data Collection (Check Sheet) → 2. Analysis and Focus (Stratification + Pareto Chart) → 3. Root Cause Exploration (Cause-and-Effect Diagram + Scatter Diagram Verification) → 4. Capability Monitoring (Histogram + Control Chart)

This process was subsequently promoted to the factory's other three welding lines and one painting line, achieving significant results.

8. Revisiting the Practical Logic of the Seven QC Tools

8.1 The Essence of the Tools is a "Thinking Framework"

Many companies training on the Seven QC Tools only teach "how to draw the charts" but overlook "why to use this tool at this stage." The true value of the Seven QC Tools lies in their ability to form a complete improvement thinking chain:

What happened? → Check Sheet (factual data) Where are the differences? → Stratification (layered exposure) What should be solved? → Pareto Chart (focus on key issues) Why did it happen? → Cause-and-Effect Diagram (systematic thinking) Is it true? → Scatter Diagram (data verification) Can the process meet the standards? → Histogram (capability assessment) Is the process stable? → Control Chart (continuous monitoring)

8.2 Three Common Misconceptions

Misconception One: More tools are better. In practice, solving a specific problem often requires only a combination of 2-3 tools. In this example, the tools truly used for analysis were Check Sheet → Stratification → Pareto Chart → Cause-and-Effect Diagram → Scatter Diagram, while the Histogram and Control Chart were mainly used for baseline assessment and effect confirmation.

Misconception Two: More data is better. Data quality is far more important than data quantity. A well-designed check sheet collecting data for 3 days can be more valuable than a chaotic monthly data set.

Misconception Three: Analysis is the end. The Seven QC Tools are not just "analysis tools"; their ultimate goal is to derive improvement actions. Each identified cause must correspond to an executable improvement measure; otherwise, the analysis is merely theoretical.

8.3 Integration with Six Sigma DMAIC

Notably, the Seven QC Tools are naturally compatible with the Six Sigma DMAIC methodology. The improvement path in this example corresponds to each stage of DMAIC:

This means that even without systematic Six Sigma training, frontline teams can use the Seven QC Tools to follow an improvement path that closely mirrors the complete DMAIC process. This is the fundamental reason why the Seven QC Tools have remained relevant for decades—they lower the threshold for solving problems without compromising the quality of the solution.

9. Conclusion

The story of Huacheng Precision Engineering is not unique. In the context of Chinese manufacturing transitioning from "scale expansion" to "quality-driven," many small and medium-sized manufacturing enterprises face not the question of "whether they have advanced tools" but "whether they can fully utilize basic tools."

The Seven QC Tools—these seemingly simple and even somewhat outdated "old methods"—are precisely the key to solving this problem. They do not require expensive software investments or highly educated statistical experts; they only require that frontline teams are willing to take the time to record data, analyze it in layers, draw a fishbone diagram, and verify a hypothesis.

True quality improvement is never a miracle that falls from the sky but is the inevitable result of accumulating data, focusing on key issues, and systematically exploring root causes.

Seven old tools, one improvement path—transforming data into action, from problems to a closed loop.

Knowledge Number: 5.2.4

Version: v20260722

Author: Excellence Quality Think Tank Excellence Quality Think Tank is dedicated to providing systematic professional knowledge, methodologies, and practical tools to quality management practitioners, helping enterprises continuously enhance their quality capabilities.