7 QC Tools in Japanese Quality Reports: How to Read Them
When a Japanese supplier answers a quality complaint, the reply often arrives as a fuguai hokokusho (不具合報告書, defect report): a few pages of charts with Japanese labels, a fishbone diagram, and a before-and-after comparison. QC circle teams at Japanese plants present their improvement work the same way. Many of the charts in those documents come from one toolkit, the 7 QC tools, called QC nanatsu dogu (QC七つ道具) in Japanese.
You probably know the tools under their English names. What slows readers down is the Japanese name on each chart, the order in which the charts appear, and what a Japanese quality team expects you to conclude from them. This guide covers the tools as Japanese plants use them, the Japanese labels you will see, and what to check before you accept a report.
Contents
- What Are the 7 QC Tools (QC Nanatsu Dogu)?
- The QC Story: Why Report Charts Follow a Sequence
- How to Read Each Tool in a Japanese Report
- Cp and Cpk in Before-and-After Comparisons
- The New Seven QC Tools (Shin QC Nanatsu Dogu)
- Strengths and Limits of the 7 QC Tools
- Checklist: Reviewing a Japanese Supplier's Defect Report
- Where This Guide Does Not Apply
- FAQ
- Sources
What Are the 7 QC Tools (QC Nanatsu Dogu)?
The 7 QC tools are basic methods for organizing and analyzing mainly numerical data from a process: dimensions, weights, and counts of nonconforming parts. The Union of Japanese Scientists and Engineers (JUSE, Nikkagiren) describes them as methods that anyone on the shop floor can use, not only statisticians. In Japan they work as a shared language for improvement. Production operators, line leaders, and quality engineers draw the same charts and work from the same data. Japanese quality professionals call this data de mono wo iu (データでものを言う, "let the data speak"): decide on facts and data, not on gut feel or experience alone.
JUSE's own list actually contains eight methods: stratification, Pareto chart, cause-and-effect diagram, graph, check sheet, histogram, scatter diagram, and control chart. Graphs and control charts count as one tool, which brings the total to seven. ASQ's list of the "seven basic quality tools" is close but not the same. It names stratification as the classic seventh tool, notes that more recent lists often put the flowchart in that slot, and sometimes includes the run chart. If a Japanese document lists seven tools that do not match your training material, the counting convention is the likely reason, not a different method.
Basic methods for organizing numerical data, from finding a problem to confirming the fix.
JUSE counts graphs and control charts as one tool and includes stratification to reach seven.
The Japanese Names You Will See on Reports
Japanese reports often label the tools and chart axes in Japanese only. The table below gives the terms that appear most often, with the reading in romaji.
| Japanese | Reading | English |
|---|---|---|
| パレート図 | pareto-zu | Pareto chart |
| 特性要因図 | tokusei yoin zu | Cause-and-effect (fishbone, Ishikawa) diagram |
| 管理図 | kanri-zu | Control chart |
| チェックシート | chekku shito | Check sheet |
| ヒストグラム | hisutoguramu | Histogram |
| 散布図 | sanpu-zu | Scatter diagram |
| 層別 | sobetsu | Stratification |
| 件数／累積比率 | kensu / ruiseki hiritsu | Count / cumulative percentage (Pareto axes) |
| 不良率 | furyo-ritsu | Defect rate |
| 特性／要因 | tokusei / yoin | Effect (the quality characteristic) / factor |
| 規格上限／規格下限 | kikaku jogen / kikaku kagen | Upper / lower specification limit (USL / LSL) |
| 中心線 | chushin-sen | Center line (CL) |
| 上方管理限界／下方管理限界 | joho kanri genkai / kaho kanri genkai | Upper / lower control limit (UCL / LCL) |
| 対策前／対策後 | taisaku-mae / taisaku-go | Before / after countermeasure |
| 真因 | shin'in | True cause (root cause) |
| 歯止め | hadome | Lock-in step that keeps the fix in place (literally "a brake" or "stopper") |
The QC Story: Why Report Charts Follow a Sequence
Japanese teams rarely use these tools on their own. They place them in a problem-solving sequence called the QC story (QCストーリー), which turns the PDCA cycle (Plan, Do, Check, Act) into concrete steps. A typical sequence runs: select the theme, understand the current situation and set a target, analyze the causes, plan and carry out countermeasures, confirm the effect, and standardize the fix so it stays in place.
JUSE treats each QC method as a means to an end and matches tools to the problem-solving stage: finding the problem, pursuing the true cause, and confirming the effect. That is why a defect report or a QC circle presentation tends to show a trend graph first, then a Pareto chart, then a fishbone diagram, then before-and-after data. Once you know the sequence, you can tell which stage each chart is meant to prove, and which stage is missing. If your customer requires an 8D report, many of these charts fit into its disciplines; see how to write an 8D report and how it maps to Toyota's 8-step method. For how QC circles fit into daily kaizen, see kaizen in Japanese companies.
Japanese reports often follow this sequence, using the tool that fits each step.
How the steps are divided varies slightly by source and company.
How to Read Each Tool in a Japanese Report
The basic idea behind all seven tools is that every process varies. Part of that variation comes from chance: small differences in material and environment that add up and can never reach zero. The rest has an assignable cause, such as tool wear, a setup mistake, or a change of material lot, and that part can be found and removed. Most of the tools exist to separate the two and track down the assignable cause.
Check Sheet (Chekku Shito)
A check sheet is a form for collecting data quickly and without gaps. Japanese plants use two kinds: a recording type, where operators tally defects by type, and an inspection type, where they tick off check items. Handwritten tallies often use the character 正 (sei) instead of four strokes and a slash. The character has five strokes, so each complete 正 means five.
What to check: whether the sheet also recorded the date, machine, and operator. Without those conditions, the data cannot be stratified later. Many plants now collect the same data on tablets or directly from gauges and equipment, which cuts down on missed entries and copying errors. The logic stays the same: decide first what data you need and why.
Graph
Bar charts, line charts, and pie charts make the data visible. In a defect report, the first chart is usually a line graph of the defect rate over time. It shows when the problem started, or whether it clusters on certain days of the week. Check that the time axis reaches back far enough to show the normal level before the problem.
Pareto Chart (Pareto-zu)
A Pareto chart totals defects or failures by category, sorts the bars from largest to smallest, and overlays a line for the cumulative percentage. JUSE notes that the really large problems and causes are usually concentrated in two or three items. Even with ten defect types, the top two or three often account for most of the count, so teams fix those first.
What to check: the unit on the vertical axis. Ranking by count and ranking by loss amount (in yen) can give a different order. Also check how large the "other" (その他, sonota) bar is. If it is big, the categories may be too coarse.
A machined part, one month of defects: 90 in total. Bars are sorted largest first; the right column is the cumulative percentage (累積比率).
Fictional data for illustration. The top two categories account for 74% of the count, so the team would start there. Ranking by loss amount (yen) could change the order.
Cause-and-Effect Diagram (Tokusei Yoin Zu)
The cause-and-effect diagram, widely known as the fishbone or Ishikawa diagram after Kaoru Ishikawa, is called tokusei yoin zu in Japan, literally "characteristic-factor diagram." The tokusei (特性) at the head of the fish is the result being studied, such as an out-of-tolerance dimension. The branches list the yoin (要因, factors) that may affect it. Teams often organize the major branches by the 4M: man (人, hito), machine (機械, kikai), material (材料, zairyo), and method (方法, hoho), to avoid missing whole areas.
What to check: every factor on a fishbone diagram is a hypothesis. A good report shows which factors the team verified with data or by looking at the actual parts, and which it ruled out. If the countermeasure jumps straight from the diagram without that verification, ask for it. Teams often follow the fishbone with a why-why analysis on the narrowed-down factors; see why-why analysis (5 Whys) for Japanese customers.
Histogram (Hisutoguramu)
A histogram divides the data into intervals and shows how many values fall into each one as bar heights. It shows the spread, the shape of the distribution, and where the data sits relative to the upper and lower specification limits.
What to check: a histogram with two peaks suggests that data from two different conditions (two machines, two shifts, two material lots) has been mixed. That is a signal to stratify. Before-and-after histograms are also a common way to show that a countermeasure worked.
Scatter Diagram (Sanpu-zu)
A scatter diagram plots two variables against each other to see whether they are related, for example process temperature against a dimension, or tightening torque against defect rate. What to check: correlation does not prove causation. A report that uses a scatter diagram as its only proof of the true cause deserves a follow-up question.
Control Chart (Kanri-zu)
A control chart shows whether a process is in a stable state. Walter Shewhart developed it in the United States in the 1920s. In Japan, its use is standardized in JIS Z 9020-2 (Control charts, Part 2: Shewhart control charts), which corresponds to ISO 7870-2. The chart plots data over time with a center line and upper and lower control limits. Points outside the limits, or patterns unlikely to occur by chance, such as a run of points on one side of the center line, indicate an assignable cause.
The control limits are not the drawing tolerance. A control chart asks whether the process is behaving as it usually does, not whether parts are in spec, so it can reveal a change before any defects appear. Choose the chart type by data type: for measured values such as dimensions or weight, an X̄-R chart (Xbar-R管理図) pairs the average with the range; for counts such as the number or rate of defective parts, np charts and p charts are used.
Stratification (Sobetsu)
Stratification means splitting data into groups by condition and comparing them. JUSE sums it up with a common Japanese saying, wakeru to wakaru (分けると分かる, "split it and you'll see it"). Typical splits are by person or shift, by machine or line, by material or supplier, by work method or measurement method, by time (morning or afternoon, day of the week), and by environment such as temperature and humidity. Stratification is rarely a chart of its own. It shows up as a Pareto chart or histogram drawn separately for each machine or shift.
Cp and Cpk in Before-and-After Comparisons
Once a process is stable, Japanese reports often state its process capability index to show how reliably it meets the specification. Customers sometimes set a minimum Cpk for their suppliers.
- Cp = (upper specification limit − lower specification limit) ÷ 6σ, where σ is the standard deviation. It compares the tolerance width with the process spread.
- Cpk = distance from the mean to the nearer specification limit ÷ 3σ. Because it also accounts for an off-center mean, Cpk is never larger than Cp.
Example: a part specified at 10.0 ± 0.2 mm (tolerance width 0.4 mm) with a standard deviation of 0.05 mm has a Cp of 0.4 ÷ (6 × 0.05) ≈ 1.33. If the mean drifts to 10.1 mm, only 0.1 mm remains to the upper limit of 10.2 mm, and Cpk drops to 0.1 ÷ (3 × 0.05) ≈ 0.67. The spread is the same, but the risk of defects is very different.
Cp = 1.00 means the tolerance width equals 6σ. If the process is centered and the data is normally distributed, about 0.27% falls outside the limits on both sides combined. In practice, 1.33 or higher is a common benchmark for adequate capability. When a report quotes Cpk before and after a countermeasure, check that the process was stable when the data was taken, because the index assumes a stable process.
Cp looks at spread only; Cpk looks at spread and an off-center mean.
σ = standard deviation. Both indexes assume a stable process.
The New Seven QC Tools (Shin QC Nanatsu Dogu)
The 7 QC tools handle mainly numerical data. The new seven QC tools (新QC七つ道具, shin QC nanatsu dogu, also called N7) handle mainly verbal data: facts, opinions, and ideas written as short statements. JUSE lists them as the affinity diagram, relations diagram, tree diagram, matrix diagram, arrow diagram, PDPC (process decision program chart), and matrix data analysis. In JUSE's shorthand, the original set is Q7.
N7 tools turn vague, complex problems into diagrams that a whole team can agree on. Japanese companies often use them at the planning stage, when numerical data is not yet available in sufficient quantity, and for problems that span several departments. For example, an affinity diagram groups opinions collected from customers or the shop floor to show the overall picture. A tree diagram breaks a goal down into concrete actions, and an arrow diagram lays out the schedule. A common pattern is to plan with N7 and verify results with Q7.
They handle different data and stages, and work best in combination.
Based on JUSE explanatory material, organized by MONOSCOPE.
Strengths and Limits of the 7 QC Tools
The main strength is accessibility. The tools need no advanced statistics, only paper and pen or a spreadsheet, so operators, managers, and engineers can discuss the same data. Discussions depend less on intuition or on whoever speaks loudest. The same flow (Pareto to narrow the focus, fishbone to list causes, control chart to confirm the effect) works in any industry, and well-organized charts make quality explanations to customers more convincing.
The limits matter when you review a report:
- Data quality. Results are only as good as the data. Missing records, or data without the conditions needed for stratification, cannot support a sound conclusion.
- Complex problems. The tools suit fairly simple relationships. When many factors interact, teams need design of experiments (DOE), multivariate analysis, or AI-based data analysis.
- Charts for their own sake. Drawing the chart can become the goal, and the report ends at documentation. A chart matters only if the team states what it learned and what it will do next.
- After-the-fact focus. The tools trace defects that have already happened. To find potential failures and defects at the design and process stage, before they occur, teams use FMEA (failure mode and effects analysis).
Related: AIAG-VDA FMEA for Japanese Customers: DFMEA vs PFMEA
Checklist: Reviewing a Japanese Supplier's Defect Report
Use this list when a defect report or QC circle presentation arrives. Each item ties a chart to the question it should answer.
- Trend graph covers the period before the problem, so the normal level is visible
- Check sheet data includes date, machine, operator, and material lot
- Pareto axis unit stated (count or yen) and "other" (その他) bar small
- Stratified charts (by machine, shift, lot) shown, not only the plant total
- Fishbone factors marked as verified or ruled out, with the evidence for each
- True cause (真因) confirmed with data or actual parts, not only a scatter diagram
- Before-and-after data (対策前／対策後) taken under comparable conditions
- Cp/Cpk values based on a stable process, with the sample size and period stated
- Lock-in step (歯止め): work standard revised and a control chart or check in place
- Other products, lines, or plants checked for the same cause (yokoten, horizontal deployment)
Questions to Send Back to the Japanese Side
If parts of the report are missing or unclear, these questions are short and direct enough to send as written:
- "Could you send the Pareto chart broken down by machine (or shift) as well as the total?"
- "Which factors on the fishbone diagram were confirmed by data or by checking actual parts, and which were ruled out?"
- "What is the sample size and data period for the Cpk values before and after the countermeasure?"
- "How did you confirm the process was stable when you took the 'after' data? Could you share the control chart?"
- "What lock-in step (hadome) did you put in place, and which work standard was revised?"
- "Have you checked other lines or products that could have the same cause?"
Where This Guide Does Not Apply
The number of tools, the names of QC story steps, and the report format differ by company, industry, and customer. Some companies count stratification as a separate tool and merge graphs with control charts; others use their own report templates or require 8D. Benchmarks such as Cpk 1.33 are common practice, not a universal rule, so the value that applies to you is the one in your customer's or supplier's quality agreement. The control chart patterns that count as "abnormal" also vary. Confirm which rules your counterpart uses rather than assuming the Western Electric or Nelson rules.
FAQ
What are the 7 QC tools?
They are the check sheet, Pareto chart, cause-and-effect (fishbone) diagram, histogram, scatter diagram, graph and control chart (counted as one), and stratification. This is JUSE's grouping. ASQ's list is similar but often replaces stratification with a flowchart.
What is tokusei yoin zu?
It is the Japanese name for the cause-and-effect (fishbone or Ishikawa) diagram. Tokusei is the effect or quality characteristic at the head of the fish, and yoin are the factors on the branches.
What is the difference between the 7 QC tools and the new 7 QC tools?
The 7 QC tools analyze mainly numerical data on the shop floor. The new 7 QC tools (N7) organize mainly verbal data and are used for planning and for complex, cross-department problems.
What is a QC story?
It is the Japanese problem-solving sequence built on PDCA: select a theme, understand the situation, analyze causes, carry out countermeasures, confirm the effect, and standardize. Japanese defect reports and QC circle presentations usually follow it, which is why their charts appear in a predictable order.
Are control limits the same as specification limits?
No. Specification limits come from the drawing and define acceptable parts. Control limits come from the process data and show whether the process is behaving as usual.
Sources
- Composition and purpose of the 7 QC tools, stratification, problem-solving stages: JUSE, "Key points for using the QC seven tools, Part 1" (Japanese)
- The seven methods of the new QC tools (N7) and their features: JUSE, "Key points for using the new QC seven tools, Part 1" (Japanese)
- Quality control, TQM, and PDCA: JUSE, "Knowledge: Quality Control" (Japanese)
- Seven basic quality tools (Western list): ASQ, "Seven Basic Quality Tools"
- Shewhart control charts standard: Japanese Standards Association, JIS Z 9020-2 preview, 2023 edition, which states its correspondence to ISO 7870-2:2013 (Japanese)
- Definition of Cp and Cpk: Cybernet Systems, "Process capability index (Cp, Cpk)" (Japanese)
Original article in JapaneseThis guide is adapted from a MONOSCOPE article written for readers in Japan: QC七つ道具とは？品質管理手法の仕組みと使い方を図解でわかりやすく解説
About the editor
MONOSCOPE EditorFOUNDER & EDITORSpent about ten years at a plant equipment maker and a plant engineering company, proposing, designing, and commissioning equipment, then moved into web marketing. Now a web consultant for B2B manufacturers, helping them explain their products and technology. Licensed in Japan as an Energy Manager and a Third-Class Chief Electrical Engineer.
Has run an engineering blog since 2017 (500+ articles) and a Japanese YouTube channel with about 20,000 subscribers. Each English guide is written from the Japanese shop-floor point of view and checked against the cited standards.
