QFD Quality Function Deployment in Practice —— A Systematic Method for Converting Customer Voice to Special Characteristics

1. Why QFD is the "Translator" of Design Quality

Many manufacturing companies face this dilemma: the R&D department produces products according to the specification book, and all functional tests are passed, but customers are still dissatisfied. Some customers complain that "the operation is too complex," others that "the noise is too loud," and still others that "the dimensions are unstable." The R&D personnel feel wronged: the specification book is provided by the marketing department, and they just follow it. Where does the problem lie? It lies in the chain of demand transmission—customers express their needs in "everyday language," while engineers define products in "technical language," and there is a lack of a bridge to translate between the two. QFD (Quality Function Deployment) is that bridge.

QFD was born in the 1960s in Japan, proposed by scholars such as Yoji Akao and Shigeru Mizuno. It was initially applied at Mitsubishi Heavy Industries' Kobe Shipyard and later popularized by automotive companies like Toyota, becoming one of the core tools of APQP (Advanced Product Quality Planning). Its essence is a systematic method: converting customer voice (VOC) into design requirements, part characteristics, process requirements, and production control requirements, ensuring that every step from concept to mass production aligns with the customer's true needs.

Many companies treat QFD as a "diagramming tool," thinking it's just about drawing a few quality house tables to pass audits. This is the biggest misunderstanding of QFD. QFD is not just a diagram but a mechanism for demand transmission. It addresses the most fundamental issue in R&D quality: whether the direction is correct. DFMEA (Design Failure Modes and Effects Analysis) focuses on "whether there are design flaws," and special characteristic management focuses on "which characteristics must be strictly controlled," while QFD focuses on "what we should design." If the direction is wrong, all subsequent efforts are like building a house on a faulty foundation.

2. Four Stages of QFD: How Demands Gradually Become Control Requirements

The most classic form of QFD is a four-stage deployment, also known as the "four quality houses." The four stages are interlinked, with the output of one stage serving as the input for the next.

The first stage is product planning, where customer demands are converted into product characteristics (engineering characteristics). For example, if a customer says "the car door should be easy to close," the engineer translates this into measurable characteristics such as "closing force less than 45N" and "closing sound pressure level less than 65dB." This stage's quality house, also called the "product planning quality house," is the most central and commonly used in QFD.

The second stage is part deployment, where product characteristics are converted into key part characteristics. For instance, the product characteristic "closing force less than 45N" requires multiple part characteristics to support it, such as the stiffness of the door hinge, the compression of the sealing strip, and the engagement depth of the door lock.

The third stage is process planning, where part characteristics are converted into process requirements. For example, the characteristic of the sealing strip's compression corresponds to process parameters such as temperature, pressure, and time in the foaming process.

The fourth stage is production control, where process requirements are converted into specific production control plans, including control items, control methods, and inspection frequencies, ultimately linking to the control plan (Control Plan).

It is important to note that not all projects need to go through all four stages. Small and medium-sized enterprises or mature products often achieve most of their goals by completing the first and second stages. The depth of QFD should match the project's importance and resource investment, with the key being to solidify the most critical product planning quality house.

3. House of Quality (HOQ): The Core Carrier of QFD in One Diagram

The House of Quality (HOQ) is the signature tool of QFD, named for its house-like shape. A complete HOQ consists of six parts.

The left wall is the customer demands and their importance. This is the input for the HOQ, listing the customer demands that have been organized and stratified, and assigning importance weights to each demand (usually scored from 1 to 5 or 1 to 10).

The ceiling is the product characteristics (engineering characteristics). Engineers translate customer demands into measurable and controllable product characteristics, typically using "＋" to indicate that a higher value is better, "－" to indicate that a lower value is better, and "×" to indicate that a value closer to the target is better.

The room is the relationship matrix. This matrix uses symbols or numbers to indicate the strength of the relationship between each customer demand and each product characteristic: ◎ for strong correlation (9 points), ○ for moderate correlation (3 points), △ for weak correlation (1 point), and blank for no correlation. This step is the soul of the entire HOQ, requiring joint discussions by R&D, quality, process, and marketing personnel, rather than a single person's arbitrary decision.

The roof is the correlation matrix of product characteristics, indicating the mutual influence between characteristics: positive correlation means both characteristics change in the same direction, negative correlation means one increases while the other decreases. The roof's role is to expose design conflicts, such as the contradiction between "reducing weight" and "increasing stiffness," or "increasing speed" and "reducing noise." Once negative correlations are identified, either the design approach should be adjusted, or the priorities for trade-offs should be clarified, providing a basis for subsequent scheme evaluations.

The floor is the importance ranking of product characteristics, calculated by weighting the relationship matrix with the importance of customer demands. This ranking directly identifies key characteristics.

The basement is the competitive assessment and target setting. The left side compares the company's products, major competitors, and industry benchmarks against customer demands (e.g., on a 5-point scale), while the right side compares the technical values of product characteristics. Finally, a target value is set for each product characteristic.

Combining these six parts, the HOQ is complete. Its greatest value lies in centralizing information scattered across various departments into one diagram, making the relationships between demands, characteristics, competition, and targets clear and transparent, providing a basis for discussions and decisions.

4. Practical Step One: Collection and Organization of Customer Voice

The left wall of the HOQ determines the quality of the entire house. The accuracy of the collected customer voice directly affects the success or failure of QFD. This step answers two questions: What do customers really want? How important is each item?

The sources of customer voice should be diverse. Common channels include market research and customer interviews, customer complaints and after-sales data, dealer and sales feedback, competitive benchmarking analysis, industry standards and regulatory requirements, and cross-departmental expert experience within the company. Relying on a single channel can lead to bias, such as focusing on price and delivery time if only sales feedback is considered, or focusing on existing faults if only customer complaints are considered. Only by cross-verifying multiple channels can a complete demand picture be formed.

After collecting the raw demands, use the affinity diagram method (KJ method) to organize them. Raw demands are often scattered and colloquial, such as "no water leakage on rainy days," "pleasant closing sound," and "easy to open." Write these raw descriptions on cards and categorize them layer by layer according to their inherent connections, forming a "customer demand hierarchy": the top layer is strategic demands, the middle layer is tactical demands, and the bottom layer is specific, actionable demands. The categorization process should involve repeated discussions to ensure that each layer's name accurately reflects the customer's original intent, not the engineer's interpretation.

Next, evaluate the importance. A 5-point or 10-point scale can be used, with scores given by customer representatives or marketing personnel, and adjusted based on factors such as complaint frequency and purchase decision weight. When necessary, the Analytic Hierarchy Process (AHP) can be used for pairwise comparisons to enhance the scientific nature of the importance calculation, but the direct scoring method is usually sufficient for general projects. It is important to note that importance reflects the customer's perspective, not the engineer's. Characteristics that engineers consider "high-tech" may be worthless to customers.

5. Practical Step Two: Building the Relationship Matrix and Importance Calculation

After the left wall is built, the next step is the most critical part of the HOQ: constructing the relationship matrix to link customer demands with product characteristics.

First, determine the product characteristics. Product characteristics must be measurable technical indicators, such as dimensions, weight, power, noise, and lifespan, and not vague descriptions like "aesthetics" or "comfort." Each characteristic should correspond to specific testing methods or measurement techniques, otherwise, subsequent verification will be impossible. The number of product characteristics is generally controlled to 15 to 25, as too many can dilute the focus, and too few can fail to cover the demands.

Then, evaluate the relationship strength for each cell. Each cell in the HOQ must answer the question: How much does this product characteristic contribute to meeting this customer demand? Strong correlation is 9 points, moderate correlation is 3 points, and weak correlation is 1 point. The most common mistake is to "score all 9s"—if all cells are strongly related, there is no differentiation, and the importance calculation loses its meaning. Generally, it is recommended that each customer demand row have no more than 2 to 3 strongly related characteristics.

After completing the relationship matrix, calculate the importance of product characteristics: multiply the importance of each customer demand by its relationship strength with the characteristic, and then sum the values by column to get the weighted importance of the characteristic. For example, if "no water leakage" has an importance of 5 and is strongly related to "sealing strip compression" (9 points) and moderately related to "door stiffness" (3 points), the sealing strip compression will contribute 45 points to the importance. Summing all demand rows will rank the product characteristics by importance.

The correlation matrix on the roof should also be taken seriously. It reveals conflicts between characteristics, such as the contradiction between "weight reduction" and "stiffness enhancement," or "speed increase" and "noise reduction." Once negative correlations are identified, either the design approach should be adjusted, or the priorities for trade-offs should be clarified, providing a basis for subsequent scheme evaluations.

6. Practical Step Three: Competitive Assessment and Target Setting

The floor of the HOQ calculates the importance of characteristics, while the basement answers "to what extent." This step includes three actions: customer competitive assessment, technical competitive assessment, and target value setting.

Customer competitive assessment is from the customer's perspective, scoring each customer demand (e.g., on a 5-point scale) to compare the satisfaction levels of the company's products, major competitors, and industry benchmarks. The assessment results can clearly show: which demands we lead in, which demands have significant gaps, and which demands are "opportunities that customers care about but neither we nor our competitors have done well." For example, if customers score "fuel consumption" as 3 for our product, 4 for the competitor, and 5 for the benchmark, this is a direction that must be prioritized.

Technical competitive assessment is from the engineering perspective, comparing the actual or design values of product characteristics. Customer scores are subjective, while technical data is objective, and combining both avoids misjudgment. Sometimes customers feel "it's about the same," but technical data shows a significant gap, indicating that competitors are cutting corners in areas customers don't notice, which can be an opportunity for us. Other times, customers feel a significant gap, but technical data shows little difference, indicating that the issue lies in perception, and may require improvements in human-machine interaction, appearance, and other soft design aspects.

Finally, set the target values. Based on the importance of customer demands, competitive gaps, and implementation difficulty, determine the target values for each product characteristic. Target values should follow the SMART principle: specific, measurable, achievable, relevant, and time-bound. When setting target values, consider cost constraints and technical feasibility, and if necessary, set them in stages: short-term targets, mid-term targets, and challenge targets.

Completing this step, the product planning quality house is basically formed. From the importance ranking, select the few characteristics at the top, which are the key quality characteristics (CTQ) of the product. These characteristics will be directly input into subsequent special characteristic identification and DFMEA, becoming the source of design quality control.

7. From QFD to Special Characteristics: How Demands Become Control Focus

The value of QFD extends beyond the product planning stage; its output should provide the source input for the entire quality chain. The two most important connection points are special characteristic identification and DFMEA.

Special characteristics (Special Characteristics) are those that affect safety, regulations, assembly, function, or customer experience, and any variation can lead to serious consequences. The importance calculation in QFD provides a quantitative basis for identifying special characteristics: product characteristics that rank high in importance and are strongly related to key customer demands are candidates for special characteristics. For example, in the HOQ, "sealing strip compression" is strongly related to the high-weight demands "no water leakage" and "low noise," ranking first in importance, and should be defined as a special characteristic, marked on the drawings, and strictly controlled in the control plan.

The connection between QFD and DFMEA is equally critical. DFMEA analyzes "design failure modes and their consequences," and the severity (S) of the consequences is evaluated based on customer demands: if a failure directly impacts the customer's most important demand, the severity is 9 to 10 points. The HOQ in QFD provides this correspondence, making the severity scoring in DFMEA more objective and linked to the importance of customer demands. Additionally, the key characteristics identified by QFD are also the focus of DFMEA analysis, ensuring consistency in the analysis scope and avoiding the disconnect where "QFD says it's important, but DFMEA doesn't analyze it."

8. Five Common Misunderstandings in Implementing QFD

Misunderstanding One: Treating QFD as a Documentation Task. Many companies create HOQs to pass customer audits or system certifications, only to file them away after completion. The true value of QFD lies in the discussions and consensus-building during the process: R&D, quality, process, and marketing personnel sit together to clarify demands, characteristics, and conflicts. Without this discussion process, the HOQ is just a piece of paper.

Misunderstanding Two: HOQ Once Complete, Never Updated. Customer demands change, competitors change, and technology evolves, so the HOQ should also be dynamically updated as the project progresses. It is recommended to complete the initial version during the product concept phase, revise it during design reviews, and recalibrate it based on actual test data after prototype verification.

Misunderstanding Three: Pursuing Comprehensive and All-Encompassing. Some companies cram dozens of demands and characteristics into a single HOQ, diluting the focus. Remember, the principle of QFD is "focus": capture the most critical 5 to 8 customer demands and control the number of product characteristics to 15 to 25, concentrating limited analytical efforts on the elements that truly determine success or failure.

Misunderstanding Four: Inaccurate Importance Scoring. Letting one or two people score importance based on their feelings, or inflating all scores to make them look good, will distort subsequent calculations. Importance evaluation should be completed by a cross-functional team, and when necessary, cross-verified using complaint data and market share data.

Misunderstanding Five: QFD Disconnects from Subsequent Tools. After completing the HOQ, special characteristics are still identified by "gut feeling," and DFMEA is still based on experience, making QFD an isolated island. Ensure that the output of QFD truly flows into drawings, DFMEA, control plans, and work instructions, forming a complete demand transmission chain.

9. Conclusion

QFD is the most underestimated tool in the design quality system. It does not address the issue of "whether there are design defects," which is the responsibility of DFMEA; instead, it addresses the fundamental issue of "whether the design direction is correct." If this direction is wrong, all subsequent quality tools are merely fine-tuning on a faulty course. Translating customer voice into product characteristics, focusing product characteristics into special characteristics, and transmitting special characteristics to drawings, DFMEA, and control plans, QFD uses four quality houses to complete a full-chain translation of demands. When companies truly treat QFD as the starting point of the R&D process rather than just an audit document, design quality is rooted from the source.

Customers express demands in everyday language, and engineers design in technical language—QFD is the bridge that cannot be skipped between the two.

Knowledge code: 8.2.2

Version: v20260804

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