QFD is the backbone of Design For Six Sigma. Real structuring tool, it gives the way to follow of a project of development of product / service.
Originally, the QFD was developed by two Japanese teachers, Yoji AKAO and Shigeru Mizuno, in the late 1960s. In a search for total quality, their goal was to make sure to deploy the “Voice of the Customer “through the entire design process.
Yoji AKAO developed this tool from 1965 at Matsushita Electric in Japan. But its first real use will be in 1973 on the Kobe Shipyard of Mitsubishi Industries.
The tool was then used at Toyota and then throughout the Japanese industry before and finally in the United States and Europe.
Let’s start with Toyota, which introduced this method in 1977 with the result1 :
Launch cost decreased by: 20% in October 1979, 38% in November 1982 and a cumulative result of 61% in April 1984.
In the same period the Time To Market was reduced by a third.
Reduction of the number of modifications during the project.
Finally, a study a larger number of companies shows2:
- 30 to 50% less change
- 30 to 50% less development time
- 20 to 60% less launch costs
- 20 to 50% less customer complaint
An empirical study (21 companies – 5 English, 11 US, 1 Netherlands, 1 Sweden, 2 Italy and 1 Hong Kong) was conducted by researchers to understand the barriers to QFD. This study shows 5 major problem families3.
Group 1 : Management 41,4%
Class 2: group work
Class 4: maintain commitment
Class 8: inappropriate for culture
Group 2: Complete the house of quality 32.9%
Class 1: get the customer’s voice
Class 6: complete HOQ
Group 4: Preparation 11.4%
Class 3: training
Class 9: planning
Group 3: Resource 11.4%
Class 5: delay
Class 7: Software
Group 5: other 2.9%
Le modèle de processus
Par une suite de “maison” QUOI-COMMENT, le QFD traduit la Voice Of Customer, en fonction, puis en solution puis en plan de fabrication, puis… Le COMMENT d’une “maison” devient le QUOI de la “maison” aval jusqu’à obtenir l’ensemble des éléments nécessaires à la production. Le processus complets se compose en 4 “maisons“4 :
- House of Quality : Transformation du besoin en fonction compréhensible par les ingénieurs.
- Parts deployment : Transformation des fonctions en éléments techniques.
- Process planning : Transformation des éléments techniques en process de fabrication.
- Production planning : Transformation du process de fabrication en besoin de fabrication.
The process model
By a suite of “home” WHAT-HOW, the QFD translates the Voice of Customer, in function, then in solution then in plan of manufacture, then … The HOW of a “house” becomes the WHAT of the “house” “downstream to get all the elements necessary for production. The complete process consists of 4 “houses” 4 :
- House of Quality: Transformation of need into understandable function by engineers.
- Parts deployment: Transformation of functions into technical elements.
- Process planning: Transformation of technical elements into manufacturing process.
- Production planning: Transformation of the manufacturing process into manufacturing needs.
Matrix I: the house of quality
The house of quality
The heart of the QFD is the quality matrix also called “house of quality”5,6 (so named by Mr. Fukuhara because of its shape). This is the first matrix of the QFD process, and the most important one. Here is the detailed diagram:
A: The needs of the customer – The Voice Of Customer
The first step is to identify the needs, their expectations … 7:
- A1: Identify the customers
- A2: Capitalize their expectations: qualitative element, called “function” in the traditional functional analysis
- A3: Identifying the importance: quantitative element
B: Strategic Analysis – The Voice Of Business
The right-hand section integrates the notions of competition and business strategy into the prioritization of the development project.
B1: competitive evaluation
This involves evaluating the products of the competition according to the customer expectations previously identified. The assessment of the quality of competitors is made on the perceived level of quality that customers can have over our competitors and ourselves8. The rating scale is from 1 to 5 where we compare against the average level 3, our competitors and ourselves. 1 being a level well below the average level and 5, a level above the competition.
B2: Target level of perceived quality
After knowing the perceived performance, we will set the target level of perceived quality that we want in comparison with competitors. The evaluation is carried out as follows:
- 1 : Well below average perceived quality.
- 2 : A little below the perceived average.
- 3 : In the average of perceived quality.
- 4 : A little above the level of perceived quality.
- 5 : Above the average level of perceived quality
B3: Strategic Polarization
The product is evaluated by a “strategic polarization” criterion to integrate the product strategy of the company. For example, if all products in the industry have a “lack” on a technology, a company may decide to get ahead and improve this product function before others. This criterion has the value:
- 1: Little strategic interest.
- 1.25: Moderate interest.
- 1.5: Strong strategic interest.
This figure reflects the “strategic will of the company given the degree of importance assigned by customers“. This figure “influences quality in design to meet the requirements that have a higher degree of importance” 9.
B4: Final importance of incoming elements
It only remains to multiply the previous coefficients in order to know that they are the strategic opportunities and that they are the expectations on which the company must focus.
Final importance = (Target B2 – Current position B1) * hierarchy input elements A3 * strategic polarization B3
The interpretation at this level is as follows:
- Value <0: This indicates that we are already at a level above the level of customer demand and the market.
- Value> 0: our current level is well below customer demand and market and our strategy is to improve on this.
Note that a model is proposed to calculate the importance of the incoming elements via the Kano matrix:
Importance = (Target / Current level)1/K * Initial Importance
Source : X.X. Shen, K. C. Tan (2000) – Integrating Kano’s model in planning matrix of quality function deployment
K being the value of the Kano coefficient, proposed with 0.5, 1 and 2 respectively for must be, one dimensional and attractive.
C : How
C1: Functional criterion
It is about transforming tertiary expectations into “measurable, global and proactive” data 10. Ultimately, it is a question of qualifying and then quantifying the functions, as recommended by the NF X50-150 functional analysis standard, by writing the corresponding criteria and levels for each function.
In order to find all the criteria, we can use the 5W2H tool. For each function, one must ask all the questions (“who uses it?”, “How many times?” …) and identify the corresponding criteria.
It is possible that there are different families of criteria. In this case, we can use the methodology used to classify expectations (KJ method).
C2: the units
We will identify the units of measurement for each of the criteria (ml, cl …).
C3: sense of performance
This line makes it possible to identify the positive direction of performance of the criteria in relation to the existing one. 3 directions can be given: increase, maintain or decrease. For example, we can put a performance threshold for a noise, and give the direction of the reduction of this noise, or also identify a certain level below which the product is no longer considered quality. This measurement is identified by arrows:
- ↑: means that the more this value increases and the more the customer is satisfied (engine performance …)
- ↓: means that the lower the value, the more satisfied the customer is (braking distance …)
- Ο: means that the value is a target. Above or below this value reduces customer satisfaction (size of a city car …)
Note that the meaning of the performance is not included in the mathematical model of the matrix. Like other elements, they are there to ask the right questions and correctly assess the levels of risk and complexity.
D: correlation matrix
The correlation matrix will make it possible to record the correlations between the QUOI (part A) and the COMMENT (part C) of the matrix in order to identify the importance of the development of such and such a criterion of the functions. The evaluation is carried out according to the following scale11:
- 9: strong correlation
- 3: moderate correlation
- 1: weak correlation
- White: no correlation
E: technical correlation
It is a question of determining which are the influences between criteria. For example, for a function, we have a weight criterion, then for another, a criterion of speed. It is very likely that these two criteria interact in a negative way. This matrix will make it possible to identify them, and to predict the compromises to be made.
For example, to lift a window faster, you need a more powerful motor. The technical correlation matrix shows that if we increase the size of the engine, we increase the weight of the assembly and we must increase the section of the hinges.
2 options are available to us in the case of negative technical correlations:
- The first is to find a solution to solve the problem by eliminating or minimizing it. For example, using a composite material to increase strength while decreasing weight (which is usually a negative technical correlation)
- Then, if no solution is satisfactory, compromises must be made by decreasing one or more targets determined in part F. These compromises can be at the Quality, Cost or Delay level.
The point is to identify the conflicting criteria for launching creative sessions, especially with the help of TRIZ, to find solutions.
The scale to evaluate this correlation is as follows:
F: technical matrix
F1: Relative importance of HOW
This measurement makes it possible to summarize the importance of needs with the correlation between what and how. This synthesis is calculated by the formula:
Relative Importance of HOW = Importance Level of Requirement (B4) * Sum of Correlations (D)
F2: Competitors’ Technical Evaluation
The first competitive analysis focused on the presence and perceived quality of functions at the beginning of the project and without giving the level of criteria finally decided for the product.
This second part of the competitive analysis is about a comparison of ourselves with the competition compared to an average level (for example, we know the number of gears for this range of road is 11). We proceed in the same way as in part B1, by measuring and comparing the level finally decided for our product vis-à-vis that of the competition.
The evaluation is done by comparing our level of performance against the average level of the market and customer demand. The scale of evaluation goes from 1 to 5, 1 being a low level, 5 being the highest level.
F3: Targeted Technical Level
In the same way as in B2, we can help to identify the target level of the different criteria with respect to F1 results. Generally, it is an intuitive data indicated by the management. The scale of evaluation goes from 1 to 5, 1 being a low level, 5 being the highest level.
F4: Strategic Polarization
In the same way as in the B3 part, we re-evaluate the different criteria in relation to the company’s wish to go as far in the improvement or not, this with regard to the results in F1. Generally, it is an intuitive data indicated by the management.
F5: Difficulty Factor
This is to identify the difficulty of achieving the target. It depends on the technical difficulty of the solution, but also on the target. Thus, if our current level is very low, and our target is high, we will have to introduce innovations with the consequences that means. This identification highlights the development challenges ahead.
It should be noted that, by convention, if an element is indicated as very complex, it becomes less of a priority because it requires an important innovation. The challenge of QFD, conceived in a philosophy of ensuring a robust design, a great innovation is not the goal.
F6: importance of the outgoing elements
In the same way as for the QUOI, we will calculate a factor of importance of the outgoing elements of the matrix, the Critical To Quality (set composed of a function and a technical level of this function).
Importance of Outbound Items = (F3 Target Level – F2 Product Level) * Relative Importance F1 * Technical Item F4 * F5 Difficulty Factor
The interpretation of this evaluation is as follows:
- <0: We are already above the technical performance target level or above the perceived quality target level.
- = 0: Indicates that we are already already at the level desired by the customer and desired in terms of performance that we set ourselves. In other words, it is not a priority.
- > 0: The highest value tells us that it is the priority action.
Matrix II – Technical characteristics of the product
Matrix 2 is representative of the component deployment, and corresponds to the “Analyze” phase of the DMADV process.
The outgoing elements of the matrix 1, are the incoming elements of the matrix 2. We will break down the product into various components and subcomponents in order to specify the product design, which we will evaluate with respect to the incoming elements ( the CTQs with their weights).
Critical To Quality
They are the outputs of the previous matrix accompanied by their respective weighting. They become the inputs of this matrix.
We will have for example, the wish of a good brightness with a bulb for a room of 30m2.
Determination of specifications
Depending on the weights, we can directly prioritize the performance levels and specifications of each feature. In proportion to each of the weights, we will determine.
- A target value: example good brightness for a room of 30m2
- A lower specification (LIS): good brightness example for a room of 20m2
- A higher specification (LSS): good brightness example for a room of 40m2
Determination of technical characteristics
In the same way, we will identify for each functional criterion specific technical characteristics. Let’s take our example of bulbs. We have in CTQ the wish of a good brightness for a room of 30m2. We will have as a technical criterion the thickness of the glass.
Matrix III – Process Selection
Matrix 3 allows us to identify the characteristics of the production processes needed to achieve the requested level of specificity. We are here in the “Design” phase of the DMADV process.
The solution identified, we must size it. Each component of the solution is detailed. Take the example of the deployment of the bulb component.
A bulb is described according to precise engineering criteria which will determine its performance and its difficulty of realization, this in direct correlation with the results of the matrix 1.
Process of fabrication
Once we have identified the technical characteristics of the product, we will look for the manufacturing processes. For this, we proceed in 2 steps:
- Look for solutions: research and develop the set of possible techniques to maximize the process capability. Test all innovations and technical improvements.
- Select the process: determine the optimal method of realization respecting the degree of precision required at a lower cost (calculate the cost factor of precision for each process).
Matrix IV – Production Specifications
The last matrix of the QFD process allows to put under control the manufacturing process of the product. We are in the “Check” phase of the DMADV process. We will compare the different characteristics of the processes chosen with the different parameters and methods to ensure the desired result.
Technical criterion of the process
We will identify the technical criteria in correlation with the processes. We will find for example the size of a tool, the type of materials … We must put all the technical criteria allowing us to ensure the capability of the machine vis-à-vis the need.
Adjustment, control points …
These criteria are then correlated with the different methods and means of production. We will find for example the update of the ranges of controls, standalone maintenance … Or the training technicians qualities vis-à-vis a new control to perform.
At the end of this step, we must have identified:
- Component production specifications, control plans and purchasing procedures. Decide on subcontracting or not, and edit the specifications
- Quality control plans for parts realization: establish the key points that are characteristics of the quality of the parts, then translate them into verification points for the process.
- The quality control chart identifying the persons in charge, assembly procedures, instructions, specifications.
- Quality control charts to subcontractors and suppliers.
1 – L. P. Sullivan (1986) – Quality Function Deployment : a system to assure that customer needs drive the product design and production process
2 – B. A. Bicknell, K. D. Bicknell (1995) – The road map to repeatable success – using Quality Function Deployment to implement change
3 – A. Martins, E. M. Aspinwall (2001) – Quality Function Deployment : an empirical study in the UK.
4 – J. R. Hauser, D. Clausing (1988) – The house of quality, Harvard Business Review
5 – L. Cohen (1995) – Quality Function Deployment: how to make Quality Function Deployment work for you
6 – L. P. Sullivan (1986) – Quality Function Deployment: a system to assure that customer needs drive the product design and production process
7 – M. L. Shillito (1994) – Advanced Quality Function Deployment: linking technology to market to customer needs
8 – C. K. Kwong, H. Bai (2001) – Determining the importance weights for customer requirements in QFD using fuzzy AHP with an extent analysis approach
9 – K. S. Chin, J. Lam, J. S. F. Chan, K. K. Poon, J. Yang CHIN (2005) – A cimosa presentation of an integrated product design review framework
10 – Y. Akao (1990) – Prendre en compte les besoins du client dans la conception de produit
11 – American Supplier Institute (1994), Quality Function Deployment
Loch, C. H., C. Terwiesch, S. Thomke. 2001. Parallel and sequential testing of design alternatives
H. E. Piedras (2003) – Optimisation multicritère des deux premières phases du déploiement de la fonction qualité
J. L. Ribeiro, F. S. Fogliatto, T. Caten (2001) – Minimizing manufacturing and quality costs in multiresponse optimization
K. Paryani, A. Masoudi, E. A. Cudney (2010) – QFD Application in the hospitality industry : a hotel case study
L.K Chan, M.L WU (2002) – Quality Function Deployment : A comprehensive review of its concepts and method