The industrial usability of quality function deployment: a literature review and synthesis on a meta-level

Thomas Lager
Centre for Management of Innovation and Technology in Process Industry, Lulea University of Technology, SE-97187 Lulea, Sweden. [email protected]

Quality function deployment (QFD) is a development methodology that has been in industrial use for about 30 years. For industry professionals, it is thus important to know what type of outcomes they can expect from using QFD, but also how to behave according to best practice. In this paper, the research results from nine studies of the industrial usability of the QFD methodology have been reviewed, analysed and synthesised on a meta-level. In this metaanalysis, a new framework for the assessment of methodologies has been developed, and the QFD methodology as such has also been used as a research instrument. The results show that the previously often-cited most important outcome, shorter time-to-market, has no scientic support at all. The good news is that the outcomes better products and improved information dissemination and retrieval have strong support.

1. Introduction

n times of tight company resources and a continual need to improve company performance in all areas, including research and development (R&D), the use of effective and efcient tools and development methodologies for R&D is one avenue to follow. But is the price a company pays to learn, introduce and use a methodology worth the benets it will share in various forms such as product improvements, better work processes, etc? Sometimes new methodologies have hardly been introduced in a company before new tools or methodologies are presented and advocated by academics, consultants or pioneering industry professionals. There are certainly pros and cons for the use of development methodologies, but the frequent industrial use of different types of methodologies and the associated company resources they demand make assessment of

the usability of development methodologies a growing concern for company R&D management see for example Lindemann (2002).

1.1. The industrial need for development methodologies

What is then the reason that more or less complicated development methodologies like Quality Function Deployment, abbreviated QFD, are needed in industrial product and process development? One answer could be that the size and complexity of todays industrial organisations have made the distance between production, development, marketing and the customer so great that all company functions have difculties in clearly hearing the voice of the customer and disseminating it further within the organisation. If the frequent and spontaneous contact between 409

R&D Management 35, 4, 2005. r Blackwell Publishing Ltd, 2005. Published by Blackwell Publishing Ltd, 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA.

Thomas Lager the producer and user of products one experienced long time ago has been broken, a system error in the functionality of the organisation has been created that could cause the need to develop and use methodologies like QFD in order to overcome this problem. Todays company cross-functional internal and external teamwork, sophisticated product specications and shorter product development cycles make development work often very demanding on the organisation. Nor is it enough any more to develop a product that pleases the customer; it must also be better than competing products in the global arena. Development methodologies like QFD are sometimes criticised for being too complex and time-consuming, but it may not be these methodologies as such that are too complex. Are they only an image of todays complex industrial development environment? QFD is now, after about 30 years of industrial use, a well-proven and mature development methodology. Not only is QFD a methodology that is frequently used in industrial development processes (Cristiano et al., 2000), but new development concepts like Design for Six Sigma also rest on the foundation of QFD (Tennant, 2002). It is not only of interest to develop new methodologies and concepts but also to improve and further rene existing methodologies. This type of improvement must, however, start from an assessment of the performance of the existing methodology, which of course is also the case for QFD. The assessment of the usability of methodologies is consequently an activity of major industrial interest but, because of the difculties involved, should not only be left to industrial practitioners; it should be a major concern of scholars of innovation and technology management. methodology in order to improve the methodology. The purpose of this study is to assess the industrial usability of the QFD methodology and try to provide answers to the following questions. The general research question is: How good is the industrial usability of the QFD methodology? The more specic research questions are: What is best practice and what are the success factors in the introduction and use of QFD? Assuming that a company is using best practice, what are the expected outcomes? Do the expected outcomes match the necessary input of company resources? And nally: Do differences in time, geography or other organisational characteristics inuence the use of the QFD methodology? In this paper, the research results from nine studies of the use of the QFD methodology have been brought together, analysed and synthesised on a meta-level. The QFD methodology has also been used as a tool for this qualitative assessment, and the results are presented in two QFD matrices: one for success factors and another for expected outcomes. The results are further discussed, and implications for industry and academic research are drawn. But rst of all a general framework for the assessment of development methodologies will be presented.

2. A framework for the assessment of development methodologies

How does one assess a methodology? Is userfriendliness the only relevant criterion for such an assessment? After reecting on these questions, the following more philosophical standpoint was adopted for this study: Methodologies cannot be assessed on their own merits but only in relation to the activities they are meant to facilitate or improve. In the case of methodologies that are supposed to facilitate and improve R&D, and product development in particular, the assessment of the improvement of this activity is consequently the ultimate measure of the industrial usability of the methodology. The complexity of the assessment of an R&D methodology made it necessary to develop a framework based on some sort of model of R&D. Figure 1 shows a simplied conceptual model of R&D and product and process development as a further development of a model by Lager and Horte (2002). In this model, one can distinguish input and output from individual
r Blackwell Publishing Ltd. 2005

1.2. The purpose of the study and related research questions

Finding ways to assess the industrial usability of methodologies is a difcult question that is not likely to have a simple answer. For industry professionals, it is of prime importance to know what types of outcome they can expect from using a development methodology, in order to select the most appropriate one to use for a given application. It is not only important to know the expected results, but to understand how to behave in the introduction and use of the methodology to reap its fruits. For academic scholars, it is also important to gain a deeper understanding of the 410
R&D Management 35, 4, 2005

Industrial usability of quality function deployment

Company external business environment The product & process development process Input Production function Interaction R&D function
A portfolio of product & process development projects on a programme level

Output and internal outcome from process development projects & programme

Internal output from product development

Input of total development resources

Output and external outcome from product development on a company level . Success factors & performance indicators on a company level New and improved products on the market

Proj. No1 Proj. No 1

Proj. No. .
Interaction

Output

Individual projects on a project level

Input

Input

Marketing & Sales function

Figure 1. A further developed conceptual model of research and development (R&D) at different levels of analysis (project, R&D, process and company levels) after Lager and Horte (2002). The use of a development methodology inuences performance on all levels.

projects and also on an R&D (programme) level. One can also distinguish the interaction between R&D, production and marketing function in the product and process development work process. Process development delivers its outcome within the company, while product development delivers its outcome outside the company, so it is not an easy task to measure the R&D functions share of future product revenues. The use of the methodology will inuence the efciency of the product development process, and also the effectiveness of this process. In using development methodologies, the share of the input of the necessary total resources that can be attributed to the use of methodologies can often easily be accounted for. Measuring the share of the total outcome that can be attributed to the use of a development methodology is however often very difcult. This conceptual model of R&D is referred to in the further discussions of the assessment of development methodologies.

2.1. Methodological efciency

Success in any human endeavour is never a sure thing, and this also goes for the use of methodologies. Even if a methodology has proven to be useful for other companies in a similar application, this does not necessarily imply that its use in ones own application will be successful. In pracr Blackwell Publishing Ltd. 2005

tice, it is probably advisable to test a new methodology in a company pilot project. The term implementation of a methodology is sometimes confused with use of the methodology. In this paper, implementation means the introduction of a methodology as far as carrying out a rst pilot project. Use means its continued application by a company to further projects. One should also be entitled to ask how often it is necessary for a company to use a methodology to give it a fair chance. The resources required to introduce a new methodology in a company and to implement it in a rst project will, however, vary according to its complexity and user-friendliness. The user-friendliness of a methodology is a matter of how easy it is to learn and to apply in practice. It will be inuenced by the complexity of the methodology, but also by how well it is explained and presented, and possibly also how well it is supported by associated software. Knowing about a methodology as such is not the same thing as knowing how to introduce and use it in an industrial environment. One will obtain good results from a methodology that is used properly, and poor results from misused ones. One sometimes hears unjustied complaints about a well-tried methodology, when the real culprit is the organisations own inability to introduce and use it in an industrial environment according to best practice. In studying methodologies, we are not really interested in what results a company gains from misusing them.
R&D Management 35, 4, 2005

Output

Input

Pay-back, timeliness, cost and quality of programme and projects; product and process development

411

Thomas Lager To reap the fruits of a methodology, it is necessary to use it properly. Behaving in a way that has previously proved to give good results is usually called best practice. Best practice can be formulated and condensed into a number of success factors. The general idea is that if the user of a methodology behaves according to those success factors, he can expect good outcomes from the methodology. An opposite angle to the development of success factors is to identify potential barriers to successful use of a methodology. It may be rewarding for companies to try to identify such potential barriers and how to overcome them in the advance planning of a project. In this paper, reported barriers from different studies have been translated into success factors, simply by asking the question as to how to behave to overcome the mentioned barriers, and identifying this behaviour as a success factor. It is not an easy task to estimate and budget the cost for development projects, and it is certainly not easier to estimate in advance the cost of using new methodologies. It is often a costly business to start using a methodology for the rst time in a company with a pilot project on a new product. Using the same methodology again in another product development project should be expected to cost less if the team is familiar with it and when the intellectual infrastructure is available. If the methodology is used repeatedly in the development of the next product generation, the cost of using it is possibly even less because it is well known and background information is available for the product. measured until the product has been produced and sold on the market. The nancial return must often be calculated over an extended period of time. One must also recognise that the use of a methodology may have effects at many different levels in a company; in addition to assisting the design of better products (R&D level), it may also improve total company communication between functions (company level), or give better product specications (project level). How to use a methodology successfully in an individual development project is a different matter from using it successfully in a companys product development process, or even for the company as a whole. Outputs and outcomes often occur at all these levels, depending on the objectives. If for example a company has decided to make use of the QFD methodology to improve company work processes or cross-functional communication, the measurement of the results from a single project is not relevant.

2.3. Methodological productivity

So far, only the calculation of input of resources to use a methodology and measuring the share of the output that the use of a methodology is supposed to deliver has been discussed. However, it is not until those measurables are combined that the company can start to assess its usefulness truly. Productivity is, by denition, the relation between input and output for an activity or some sort of process (Brinkerhoff and Dressler, 1990). When using nancial measures for input and output, one reaches a nal destination in productivity calculations: protability. But in this calculation of some sort of price/performance relation, a time factor must also be considered. If the time until we obtain the results from the use of a methodology is included, classic investment models are adopted. The difculty of calculating the pay-back from invested resources is related not only to the aforementioned difculty of estimating the share of the nancial return from each product development project but also of estimating how many projects the cost of introducing the methodology should be shared among. It is probably not correct to write off the whole cost of learning and introducing the methodology on the rst pilot project. Financial calculation of the protability of using a methodology may be difcult, if not impossible, but using the investment model as a mental roadmap will give the methodology a fairer chance. But the ultimate issue is not only whether
r Blackwell Publishing Ltd. 2005

2.2. Methodological effectiveness

Assuming methodologies are introduced and used in an efcient manner, what are the expected results? This is something often called performance. But it is not the effectiveness of a methodology as such that one wishes to measure, but the improvement in the activity that it is supposed to facilitate. The share of the output that can be attributed to the use of a methodology can be called methodological effectiveness. Measuring the output or outcome from product and process development is not an easy task (see Figure 1). A single product development project does not deliver its nal outcome within the R&D function. The cross-functional R&D process is an interaction with marketing and production functions, and the nal outcome of the total product development activity cannot be 412
R&D Management 35, 4, 2005

Industrial usability of quality function deployment a methodology can improve various aspects of product development, but whether the resources committed give sufcient pay-back to make the investment an attractive option and alternative for allocating limited company resources. tributed to solving their problems with rusting cars in the mid-1970s. The rst and original system, developed by Professor Akao (1990), is still the dominant system used in Japan and was introduced in a slightly modied form by GOAL/ QPC (King, 1987). This system includes a large number of matrices and is often referred to as the Matrix of Matrices. A more simplied version developed by Dr Fukuhara received wide publicity when it was presented in the Harvard Business Review under the name House of Quality (Hauser and Clausing, 1988), while the stair of four consecutive matrices is often referred to as the Four Phases of Matrices. QFD was introduced at the beginning of the 1980s in the USA and a little later in Europe, at the end of the 1980s.

2.4. Assessing the usability of development methodologies some conclusions

There are a few conclusions to be drawn from the previous discussion in the assessment of development methodologies:  Methodologies cannot be assessed on their own merits, but only in relation to the activities they are meant to facilitate and improve.  As one cannot expect to obtain good results from a methodology if it is not used properly, the identication of best practice and development of relevant success factors for its use are of the utmost importance.  It is fairly easy to account for the costs of using a methodology after a project is nished, but not in advance. Unfortunately, it is very difcult to assess the nancial share of the nal outcome from the use of a methodology. This fact may unjustly cause companies to regard the introduction and use of methodologies as a cost item, and disregard the revenue side. Although nancial outcomes are difcult to calculate, it is probably better and a more dynamic mental model to regard the introduction and use of methodologies as an investment.  Following this line of thought, in the nal company assessment of a methodology, it is not enough to state types of outcome and how important are the results they produce; one must also try to assess whether the activity was worth the effort. For this assessment, calculations of pay-back may not be the best tool, but rather the sound judgement of the people who were involved in the activity and the use of the methodology.

3.2. QFD systems, matrices (components) and rooms

QFD is unfortunately often presented in a rather supercial manner using simplied descriptions of the House of Quality and nishing with a presentation of how customer demands are progressed to production planning using four consecutive matrices. The methodological reality is, however, more complex, as there are rst of all two different systems in use today. Looking into the industrial use of the methodology, one nds that very few companies have ever attempted phase progression into deeper levels and fully used the methodology (Martins and Aspinwall, 2001). The four-stage progression system is not even applicable in process industry because assembled products are not produced. It has been decided in this presentation to call the different matrices components and the different uses of a set of matrices QFD systems. The individual parts of the different matrices are traditionally called rooms, which has been found to be a practical designation. In all systems, there is one common component (matrix) where it all starts. This component is usually called the House of Quality.

3.3. House of Quality 3. A short presentation of the QFD methodology 3.1. A brief history of QFD
QFD originated in Japan, where quality work had been carried out from the beginning of the 1950s. It had its breakthrough in the Japanese manufacturing industry, and is often attributed to Toyota Auto Body, whose use of QFD successfully conr Blackwell Publishing Ltd. 2005

The translation of Customer Demands (WHATs) into Product Properties (HOWs) takes place in the House of Quality in both the Matrix of Matrices and Four Phases of Matrices systems. The importance ratings of the hierarchically arranged customer demands, including comparison with competing products in the customer dimension (benchmarking), are usually collectively called the Voice of the Customer (Figure 2).
R&D Management 35, 4, 2005

413

Thomas Lager presented by Akao (1990), it contained 16 matrices (components) and included four areas: Quality Deployment, Technology Deployment, Cost Deployment and Reliability Deployment. The number of matrices in the system may appear intimidating, but one can take comfort from Akaos advice to use only as many matrices as are useful for the application.
Customer Benchmarking

Correlation Matrix
Directions for improvements

Product properties
Customer demands Importance ratings

Relationship Matrix

Importance ratings
Individually designed rooms

Technical Benchmarking Present target values New target values

Figure 2. House of Quality or the Product Planning matrix.

The Relationship Matrix is used to translate the Voice of the Customer into an engineering dimension, developing measurable product properties. This includes identifying the direction in which individual properties should be developed with a view to pleasing the customer, as well as calculating the importance of individual product properties. In the engineering dimension there is now an opportunity to run a technical benchmarking of product properties. After completion of those rooms and individually designed rooms that are applicable to an individual project, the target values for a new or improved product can be set after a thorough matrix analysis.

3.4.2. The Four Phases of Matrices a system designed by Dr Fukuhara in Japan, and further developed in the USA by the American Supplier Institute. Starting with the development of the House of Quality, Customer Wants are translated into Design Requirements. These Design Requirements can serve as WHATs in a consecutive matrix and express the demands on Part Characteristics (Figure 3). Further on, they can be progressed to the Process Planning and Production Planning matrices (American Supplier Institute, 1989). This is often called phase progression, and the matrices are numbered from 1 to 4. Using the four matrices, customer wants are thus progressed from product development and design into manufacturing. It is important to recognise that not only are the individual product properties progressed to the following matrix but also their associated calculated importance ratings and, of utmost importance, the selected target values.

4. A literature review and a further classication of the individual studies

After a thorough review of previously published literature on the industrial usability of the QFD methodology, only nine studies were found that complied with the selection criteria of being a scientic study and not reports from individual companies, or general opinions expressed in books or manuals. It is thus claimed that these studies include most, possibly not all, major studies previously published in scientic international journals. The studies are further classied in Table 1 in the following section.

3.4. Beyond the House of Quality

Beyond the House of Quality there are two distinct QFD systems in use today. The origins of these two systems have previously been presented; the Matrix of Matrices is the most common system in Japan and part of the USA, while the Four Phases of Matrices system is also common in the USA and predominant in Europe. 3.4.1. The Matrix of Matrices a system designed by Professor Akao in Japan, and further developed in the USA by GOAL/QPC. The Matrix of Matrices can be considered an advanced system for experienced QFD users. In its original form as 414
R&D Management 35, 4, 2005

4.1. A brief presentation of the results from nine studies of the industrial usability of QFD
In this section, a short introduction and summary of the studies are given to introduce the reader to the individual results and provide a better underr Blackwell Publishing Ltd. 2005

Industrial usability of quality function deployment

Design Requirements
Design Requirements Customer Wants

Part Characteristics
Part Characteristics

Key Process Operations

Key Process Operations

Production Requirements

House of Quality

Parts Deployment

Process Planning

Production Planning

Figure 3. Four Phases of Matrices: a simplied presentation of the system and its components (American Supplier Institute, 1992).

standing of the character of the individual studies. The studies are presented in chronological order. 4.1.1. Survey and Review on QFD in Japan (Akao et al., 1987). This rst study of QFD use in the world presents the results from a large survey of mainly large manufacturing industries in Japan. The structure and the type of questions in the enquiry provide information in a somewhat different dimension than later studies of QFD. One gains a clear impression that the Japanese companies had already come far along the road of using QFD and closely related quality tools (14 products on average had been developed with the QFD methodology in each company). One can note that the information on demanded quality came mainly from customer complaints and eld sales, and not from new market research. It can also be noted that the QFD work was carried out in teams and that department managers took an active part in this development work. The highest ranked expected outcomes from QFD were: easier setting up of design quality and initial quality troubles reduced. 4.1.2. Evaluating QFDs Use in US Firms as a Process for Developing Products (Grifn, 1992). This rst study of the use of QFD in the USA reports the results from 35 development projects in nine companies. The results show that it cannot be taken for granted that by using QFD a company will benet from lower development costs or shorter development times. Only 25% of the projects in this study were considered to be successful, while roughly the same amount of projects were considered to have given mixed results or no change. Fifteen per cent of the projects were considered failures and as a consequence the use of the methodology was abandoned. Surprisingly, nearly all projects showed
r Blackwell Publishing Ltd. 2005

other interesting company benets: more rational decisions in the product development process, improve information ow, meld team together, forge cross functional relationships and capture knowledge. Success factors for the use of the methodology were: management treats QFD as an investment in methodology and not only as a cost, selection of smaller and not too complex projects, selection of development projects of the incremental type, secure management support and building a committed team that believes in the methodology. To sum up, the study presents a picture that the rst QFD project does not necessarily shorten the time for development: on the contrary, the project may take even longer. The QFD project will be even more difcult for those companies that are not used to working in crossfunctional teams. Looking at this from another perspective: cross-functional teamwork is likely to improve. The conclusion is that the benets of the QFD methodology will not be realised in the rst project, but in the further improvement of the companys product development process. 4.1.3. Design for Assembly (DFA), Failure Mode and Effects Analysis (FMEA) and QFD in product development: experiences from six companies (Norell, 1992). In this qualitative study of four Swedish companies, the company use of the QFD methodology had been to formulate the demands on the product in the preliminary phases of the development process. Only the House of Quality had been used, the methodology was adapted to company needs and the work had been carried out in cross-functional teams. Problems with the methodology were as follows, not in rank order: difcult to distinguish between WHATs and HOWs, the methodology is sometimes subjective and not precise and very timeconsuming, and difcult to estimate the time frame.
R&D Management 35, 4, 2005

415

416
A tale of CustomerIs QFD good two QFDs driven for product product development? development Forty companies through QFD say yes in the US Customerdriven product development through QFD in Japan QFD: an empirical study in the UK How to balance QFD and the efciency of new product development process: experiences and practices from the eld
2002 Italy Case study Large org. Known users 1997 Sweden Survey Known users Known users Members of JUSE 400 USA Survey USA USA Case study Survey Japan Survey 1997 1998 2000 2000 2001 UK Survey 35 40 (92) Not stated Register of QFD training 2 417 246 5 Mainly manuf. Mainly manuf. Mainly large 90% (31) 87% (80) (2) Mainly large Manuf. Mainly manuf. Mainly manuf. Mainly manuf. Manuf. Mainly large 37% (101) 38% (47) Mainly large 28% (19) Not stated (5) Not stated Mainly large Quantitative (descr.) Quantitative Qualitative Quantitative Quantitative Quantitative (descr.) Qualitative

Thomas Lager

Table 1. Selected studies for the meta-analysis

R&D Management 35, 4, 2005

Survey and review of QFD in Japan

Evaluating QFDs use in US Firms as a process for developing products

DFA, FMEA QFD: the Swedish and QFD in experience product development: experience from six companies

1992

1992

USA Case study

Sweden Case study

Publication 1987 year Country Japan Type of Survey investigation Sampling Members of JUSE

Known users

Known users

No. of 418 9 (35) 4 companies (projects in parentheses) Type of Mainly manuf. Mainly manuf. Mainly manuf. industries Size of Mainly large Not stated Medium sized company Response rate 30% (80) (35) (4) (number of respon-dents using QFD) Type of Quantitative Quantitative Qualitative analysis (descr.) (descr.)

r Blackwell Publishing Ltd. 2005

QFD, quality function deployment; DFA, design for assembly; JUSE, Japan Union of Scientists and Engineers; FMEA, failure mode and effects analysis; descr., descriptive; manuf., manufacturing.

Industrial usability of quality function deployment Benets that are mentioned are: well retrievable information and a good group consensus. Other advantages that are reported are improved contacts with the marketing department and within the cross functional team and a well structured work process with no irrelevant chatting. One common point of view is that the QFD methodology has a strong inuence on the development of good product specications. However, the respondents found it difcult to judge whether this signicantly inuenced the design of the nal product. All companies in the study had decided to use the methodology further, but not to make it compulsory. The possible disadvantage of a conservative effect of using QFD was deemed to be related to the nature of each development project. 4.1.4. QFD: The Swedish Experience (Ekdahl and Gustafsson, 1997). The QFD experience of 31 Swedish manufacturing companies was investigated in this survey. Companies that were fairly experienced users of the QFD methodology had, in most cases stopped after nishing a House of Quality, and did not carry out further phase progressions. Problems encountered in using QFD were as follows in rank order: lack of management support, group member commitment, lack of resources and lack of QFD experience. The working practice had normally been successive half-day meetings. In the QFD matrices that were developed, 70% used less than 30 WHATs and an average of 41 HOWs. The majority of the respondents considered the use of QFD as a successful undertaking, but there was no clear evidence that the use of the methodology had had an inuence on the nal outcome of the projects. This experience is partly underlined by the ranking of benets achieved after implementing QFD, which were as follows, in rank order: better communication, knowledge was preserved, unity in the group, more rational decisions and better designs. A further in-depth study of three seasoned QFD user companies from this study stressed the time-consuming aspect of QFD, but it was felt that the structured work process and retrievable information compensated for this disadvantage to a large extent. 4.1.5. Is QFD good for product development? Forty companies say yes (Vonderembse et al., 1997). A study of the use of QFD in mainly manufacturing Fortune 500 companies assessed the results from 80 QFD projects. For many of the companies in this study this was their rst QFD project. The study shows, very convinr Blackwell Publishing Ltd. 2005

cingly, that all companies concluded that QFD had improved their products. The methodology created an information-intensive climate, where communication is improved and ideas are freely communicated. Nearly all companies were of the opinion that project documentation was improved, and so was organisational learning. The results do not, however, indicate either lower product costs or signicantly reduced time to market. How to behave for a successful use of the methodology, or factors that inuence the outcomes of QFD, are grouped into categories: technical dimensions, organisational dimensions and personal dimensions. The most important factors that correlate to success were as follows: management support, good development team and strong commitment of the project team. Regarding the methodological aspects of using QFD, good knowledge of the methodology is important and contact with the customers during the development of the voice of the customer. The sizes of the matrices varied considerably, from 24 to 62,500 cells. The most common matrix size (80% of all projects) for a House of Quality was 1004,800 cells. There was a positive correlation between matrix size and a good product!

4.1.6. A tale of two QFDs (Comstock and Dooley, 1998). These two case studies focussed on the implementation of the methodology and the selection of the right application. It is an interesting study for many reasons, of which the unconventional approach in both projects is one. In one of the projects, aimed at the development of a completely new consumer product, a market research company was used as a consultant not only for market research but also to build the House of Quality. In the other project (for an industrial customer in the defence industry), the technical specications were used as the WHATs in the House of Quality. The results from the study can be summarised as follows:  The methodology should be adapted to the needs of the project (do not get stuck in methodological rigidities, but use a pragmatic approach).  Use a well-composed cross-functional project team (QFD work cannot be contracted, but there is a need for a QFD expert).  Develop matrices of a manageable and appropriate size (do not let the matrices be too small and banal; use of sub-matrices is recommended).
R&D Management 35, 4, 2005

417

Thomas Lager  Determine the objectives well and be careful to select a project that ts the use of QFD methodology. could be strongly correlated to better product and process performance using QFD. 4.1.8. QFD: an empirical study in the UK (Martins and Aspinwall, 2001). This study presents the results from a large survey in Great Britain including 246 companies. From the 69 enquiries that were returned, it emerged that only 19 companies used or had used QFD. Half of the companies that answered that they did not use the methodology had never heard of it before. It is noteworthy that the companies in the survey, which were among the largest in the United Kingdom in terms of turnover, were also selected because of their commitment to total quality management. Most of the responding organisations were from the manufacturing industry. The results from the study showed that 39% of the companies that used QFD considered a better product (reduced costs, time, complaints and defects and better adherence to product specications) was the most important result from using QFD. The rest of the companies did not see the most important results as directly related to the product itself. Top-ranked benets were an improved and more dynamic decision process (24%), improved teamwork skills (20%) and better communication internally and with customers (15%). It is noteworthy that more or less all QFD users stopped after nishing the House of Quality, and practically nobody carried out any further phase progression as a matter of routine. This study, which focusses on difculties and barriers in the implementation of the QFD methodology, listed the following problem areas in rank order: time-consuming, lack of good methodological knowledge, lack of group commitment, lack of management support, difcult to retrieve information and data for the voice of the customer and lack of resources. Out of the 19 companies that had tried the methodology, six had not carried out more projects, mainly because of difculties in implementation. One conclusion is that lack of training and not using external facilitators could be one cause of these difculties. 4.1.9. How to balance QFD and the efciency of the new product development process: experiences and practices from the eld (Baglieri et al., 2002). This case-study reports on the use of QFD in ve large manufacturing companies. There is general but unfortunately not very specic agreement among the companies on the organisational and conceptual power of QFD. From an organisational point of view, the main
r Blackwell Publishing Ltd. 2005

4.1.7. Customer-driven product development through QFD in the US and Japan (Cristiano et al., 2000, 2001). This comparative study of the use of QFD in 101 companies in the USA and 47 in Japan gives the QFD experience from mainly large manufacturing companies in both countries. The results from the studies are presented in two separate papers, of which the rst (Cristiano et al., 2000) presents the use of QFD and expected outcomes. Of the responding companies in the USA, 69% were using QFD compared with the unexpectedly lower gure in Japan (33%). Most of the companies used only the House of Quality, but some also reported the use of phase progression to the Part Deployment Matrix. More than half of the companies in the USA had experienced good management support in the use of QFD, but signicantly less support from management was reported in Japan. The majority of the American companies reported that the work was carried out in cross-functional groups, but the number was signicantly lower in Japan. The Japanese companies relied on customer complaints to a large extent, while the American ones made extensive use of market research and interviews for their House of Quality. The QFD benets are reported as better overall project results and better product design. Companies in the USA reported better customer satisfaction, which was signicantly lower in Japanese companies. No such ndings as improved product design for manufacturing or lower production costs were reported. In the area of better group communication and better decision process, companies from the USA reported signicantly higher gures than the Japanese companies. These areas included: better cross-functional work, better knowledge building, more rational decisions, better group dynamics and better communication with market and design departments. No evidence of reduced time to market was reported from either Japanese or American companies. The second publication from the study (Cristiano et al., 2001) lists success factors for the application of QFD. Very thorough statistical analyses were carried out but more descriptive data on individual success factors were unfortunately not explicitly presented. Necessary management support and collection of new customer information were the only factors that 418
R&D Management 35, 4, 2005

Industrial usability of quality function deployment conditions for making QFD more effective were considered to be: top managements strong commitment, short-term orientation and strong investment in training. 4.2.3. Type of investigation, sampling procedure and response rate. Most of the studies are surveys, but a few case studies have been included. Most of the samples are what one could call convenient samples as old registers of known users are used or personal communications with well-known users. One could guess that the lower the response rate, the stronger the likelihood that those who did not have good experience of QFD were those who opted not to answer a questionnaire. The response rates differ across the studies from the highest (90%) to the lowest (28%). Combining the response rate with the sample, one obtains a gure on how many companies are nally represented in the results. This gure has been used in the weighting of the studies.

4.2. A classication of the studies

To start with, it can be recognised that none of the selected studies are based on simple random samples. This implies that a strict statistical generalisation of the result in each individual study is impossible, a question that will not be elaborated further. It is thus not possible to assess the frequency of the use of QFD methodology in a country, based on the studies. The characteristics of the individual studies are further structured and compiled in Table 1. 4.2.1. Year of publication. The year of publication of the research results has been selected rather than the date of the study. Referring to the short introduction on QFD methodology, the methodology was introduced at the beginning of the 1970s in Japan, the beginning of the 1980s in the USA and even later in Europe. The year of publication is of interest because of the possibility to investigate a longitudinal effect and a timedependent change in company use of the QFD methodology. Splitting the comparative study from the USA and Japan into two parts makes it possible to compare an early Japanese study with a late Japanese study and, similarly, to compare an early American study with a late one. 4.2.2. Country, industry sector and company characteristics. Most of the studies have been conducted in one single country (presumably because of ease of contact with respondents and for other practical reasons), except for one large comparative study between the USA and Japan. Most of the companies that are studied are not from Process Industry but from other Manufacturing Industry. The sizes of the companies in the studies are mostly large or very large, and the usability of QFD methodology in Small and Medium Sized Enterprises (SMEs) is thus a totally open question. The different studies use different methodological approaches and different numbers of respondents. The number of respondents varies dramatically from the largest survey in Japan (418) to the smallest case study in the USA (2). In some studies, several projects have been studied within the same company.
r Blackwell Publishing Ltd. 2005

5. Research approach and methodological considerations

What is the scientic evidence from the industrial usability of development methodologies and the QFD methodology in particular? It is often stated that good science in a mature discipline rests on a foundation of accumulated research ndings. That is to say, new research projects do not necessarily have to start from a clean sheet of paper. This is, however, not always the case in management research; in his argumentation for the use of meta-analysis, Rosenthal says, The new work of the physical sciences builds directly upon the older work of those sciences. The social sciences, on the other hand, seem almost to be starting anew with each succeeding volume of our scientic journals (Rosenthal, 1991, p. 4).

5.1. Meta-analysis methodology and selection criteria for this study

Meta-analysis has been described in several publications on research methodology (Hunter et al., 1982; Lipsey and Wilson, 2001; Paterson et al., 2001). Using meta-analysis as a research tool is of interest if the results from a number of previous studies are available and the research question involves combining this accumulated knowledge. But meta-analysis is not just a matter of comparing the results across the studies, but also of reexamining primary research ndings in the light of the total collective body of knowledge that has been gained from a number of different studies and researchers. As a rule of thumb, it has been stated that a dozen discrete studies should be
R&D Management 35, 4, 2005

419

Thomas Lager available for meaningful meta-analysis, while more than a hundred is too ambitious (Paterson et al., 2001). Meta-analysis can be carried out comparing quantitative or qualitative data or a mixture of both. The methodological approach is that each study and its primary research ndings (including case studies with only a few companies, to surveys with a large number of companies), are looked upon as an individual case in the metastudy. The individual cases are then compared on a meta-level using a qualitative across case study methodology. Research results from quantitative surveys are thus possible to combine with qualitative studies using qualitative methods in a metasynthesis. This approach is selected not only because of the impossibility of assessing the differences in the statistical information from the individual quantitative studies using classical statistical meta-analytical tools but also because of the impossibility of combining the ndings with the results from the qualitative research reports. the more the companies using QFD that are represented in the results, the more should be the weight attributed to such a study. Not to overemphasise this weighting, the following scale was decided upon: Small number of respondents represented in the study (o5) was given a weight of 1; medium number of respondents represented in the study (550) was given a weight of 2; large number of respondents represented in the study (450) was given a weight of 3. (A sensitivity analysis, giving each study the same weight, did not give results that substantially differed from the presented results.) 5.2.2. Relating success factors and expected outcomes with individual studies in two matrices. Success factors and expected outcomes were gathered from the review of all studies and were arranged in hierarchical structures; see Figures 4 and 5. The following standard symbols and weightings in QFD were used in the matrices: a double ring represents a weight of nine, a single ring three and a triangle one. Going through each study, the strength of the evidence was estimated for each success factor and expected outcome, and support for different outcomes was based on the frequency of respondents who gave support to each item: Double ring strong support for this success factor or outcome (more than 70% of respondents); single ring medium support for this success factor or outcome (6940% of respondents); triangle weak support for this success factor or outcome (3910% of respondents); and empty no support for this success factor or outcome (less than 10% of respondents). Using the data in the matrices, a classical cumulative addition was then carried out for each success factor and expected outcome, and for each item the weight of each study was multiplied by the strength of support for each study. The total sum is given in the column to the right (the relative weights are given in the associated horizontal bar charts).

5.2. Using QFD methodology as a crosscase comparative tool for qualitative meta-analysis
QFD methodology is not rst a tool for product and process development. There are reports of it being used in many different areas like developing company strategy, developing products in service industry, etc. The use of QFD as a research tool is another possible area that is explored in this study. The QFD methodology in this study has thus not only been used to relate the individual studies to different success factors and expected outcomes, but in the development of their hierarchical structures referred to in Figures 4 and 5. 5.2.1. Estimating the relative weight of individual studies. In comparing results from the selected studies, one important question is whether they should carry the same weight. This could then for example be related to issues like the overall rigour of the investigation, response rate, the relevance or inclusion of individual questions in the enquiry, etc. Because of the difculty of assessing the quality of these parameters, it was decided to let only the number of respondents using QFD represented in the nal results inuence the weighting of the individual studies. After the decision to give individual studies different weights, the next question was to decide which scale to use. It seemed reasonable to assume that 420
R&D Management 35, 4, 2005

6. Analysis and synthesis of individual research ndings on a meta-level

It has previously been described how success factors and expected outcomes were identied during the review of the selected studies. The hierarchic structure of these factors and outcomes is presented in Figures 4 and 5.
r Blackwell Publishing Ltd. 2005

Industrial usability of quality function deployment

Calculated weight of individual success factors

Evaluating QFDs Use in US Firms as a Process for Developing Products DFA, FMEA and QFD in Product development: experience from six companies

Customer-Driven Product Development Through Quality Function Deployment in the U.S. and Japan (The Japanese results)

Customer-Driven Product Development Through Quality Function Deployment in the U.S. and Japan (The U.S. results)

Is Quality Function Deployment good for product development? Forty companies says yes

How to balance QFD and the efficiency of New Product Development process: experiences and practices from the field

Survey and review of Quality Function Deployment in Japan

Calculating the importance of individual success factors

QFD: The Swedish Experience

Weight of individual QFD studies

Well defined objective Well defined objectives for QFD

A tale of two QFDs

Quality function deployment: an empirical study in the UK

2 3 0

Selection of a good application for QFD

The selection of an existing product The selection of a product that needs to be improved Givethe first QFD project ample time for completion (QFD is time consuming)

6 0 2 0

Success factors for using QFD methodology

Secure necessary management support Strong Management Support Secure necessary resources (financial & personal) Management regards QFD as investment in methodology

119 6 11 0

A QFD team of excellence

Appoint a strong project manager (champion)

0 18 53 0

Createa well-balanced cross funtional team Select committed team members

A proper use of the QFD methodology

Provide good training of the team in QFD methodology Use an experienced facilitator through out the project Use the QFD methodology in a flexible and creative manner Find clever solutions to handle large matrices Collect new information from customers

35 8 9 10 94 0

Figure 4. Success factors for using the quality function development (QFD) methodology.

6.1. Developing a list of success factors, relating them to the individual studies and calculation of their importance
In Figure 4, the calculated relative importance of individual success factors is presented in
r Blackwell Publishing Ltd. 2005

the bar chart to the right. The results show that on the highest hierarchic level, no single area stands out, but only discrete success factors on the bottom level can be recognised. The success factor that is given the highest weight, management support, seems to be
R&D Management 35, 4, 2005

Calculated relative weight of individual success factors

Companys own importance rating of individual success factors

QFD Studies

421

Thomas Lager

QFD Studies
Is Quality Function Deployment good for product development? Forty companies says yes Evaluating QFD's Use in US Firms as a Process for Developing Products How to balance QFD and the efficiency of the New Product Development process: experiences and practices from the field

Calculated relative weight of expected outcomes

DFA, FMEA and QFD in Product development: experience from six companies

Customer-Driven Product Development Through Quality Function Deployment in the U.S. and Japan (The Japanese results)

Customer-Driven Product Development Through Quality Function Deployment in the U.S. and Japan (The U.S. results)

Survey and review of Quality Function Deployment in Japan

Expected and desired QFD outcomes

QFD: The Swedish Experience

Weight of individual QFD studies Product related

Better overall customer satisfaction Better Product (product design) Lower Manufacturing Cost Lower Design Cost

A tale of two QFDs

Quality function deployment: an empirical study in the UK

2 66 97 7 0 0

Potential outcomes from using QFD methodology

Work Process Related

A Generally Improved Product Development Process Improved Cross Functional Team work Shorter Time-to-market Shorter Design Time Better Product Specifications Better Decision Support

2 25 9 2 43 47 0

Company related

Improved Company Communication Retrievable and Accumulated Company Knowledge

59 69 0

Improved Manufacturing Process

11 0 0 12 22 45

Longitudinal comparison (accumulated sum 16 of relationship strenghts)

Figure 5. Expected outcomes from using the quality function development (QFD) methodology. A QFD programme from IDEACore has been used (IDEACore, 2003).

Others

crucial to the successful use of QFD, and seven of the 10 studies support this conclusion. Collecting new information from customers is the second highest weighted success factor, followed closely by select committed team members. 422
R&D Management 35, 4, 2005

6.2. Developing a list of expected outcomes, relating them to the individual studies and calculating their weight
Calculated relative weights are presented in a column to the right in the matrix in Figure 5.
r Blackwell Publishing Ltd. 2005

Calculated weight of expected outcomes

Companys own importance rating of individual desired outcomes

Industrial usability of quality function deployment Looking at the outcomes on the highest hierarchic level, the product- and company-related outcomes stand out. The outcome from QFD that gets the highest weighting is a better product (design). Very close to this, we also nd the related outcome, better overall customer satisfaction. It is noteworthy that the outcome better product is recognised in all studies but one. Better overall customer satisfaction has been recognised in six studies. Company-related outcomes, retrievable and accumulated company knowledge and improved company communication, are also highly weighted outcomes to expect from using the QFD methodology. Compared with the previously presented areas, the highest weighted outcomes in the work process-related area are better decision support and better product specications, but they receive only about half the weight compared with the other. The low weight of all outcomes in this area can be interpreted to mean that one cannot expect any outcomes in this area at all. There is practically no support at all for outcomes that are time-related shorter design time or shorter time-to-market. Nor is there any support for an improved manufacturing process, which may be caused by the sparse use of the deeper levels of the QFD systems. nised at all in the former, but receive a high weighting in the latter. It is difcult to draw any conclusions about differences in individual outcomes, but the overall pattern is somewhat similar. On the other hand, the overall total weighting of all expected outcomes is nearly four times higher in the later study, which could indicate a growing general support for the methodology. Do geographical or cultural differences inuence success factors and expected outcomes for a methodology, something one could designate a space aspect of methodologies? There is one study (Cristiano et al., 2000) that makes a direct comparison between two countries. Referring to the previous presentation, there is a signicant difference between the USA and Japan, which may be related to differences in company behaviour or culture.

6.4. Industry sector and company size

Does the industrial usability of QFD methodology differ between different industrial sectors? One could expect that different industrial environments call for different QFD approaches that the Automotive Industry, for example, uses or should use QFD in a different way than the Pulp & Paper Industry. Are success factors and expected outcomes then different in different industrial sectors? Companies from several industry sectors are included in the sample in most of the studies cited here, but the results are not discussed from this angle. Looking at all samples in Table 1, one can only say that the results show some sort of average results from Manufacturing Industry. There are possibly many different company characteristics that could affect the use of the QFD methodology; one of them may be the size of the company. The size of the companies in the studies is often large and in many cases very large. It must, however, be recognised that even in very large companies, individual QFD projects are often studied and the organisational unit is then much smaller. The results presented generally represent the use of QFD in large companies.

6.3. Time and space

The inuence of time on the use of QFD could be called a longitudinal aspect of the methodology. Does regular use of a methodology in a company inuence its usability? On a country level, does the perception of the usability of a methodology change over time? There are two studies of the use of QFD in Japan; see Table 1. The rst study was carried out in 1987 by Akao (Akao et al., 1987.), and the second one was presented in 2000 (Cristiano et al., 2000). The time span between the studies is thus about 10 years. For individual success factors, the results have not changed over time but are fairly similar, with the exception that the importance weightings are stronger in the last survey. One difference is that the latter study shows stronger relations in the area of product-related expected outcomes. The rst study in the USA was carried out by Grifn (1992), and the second by Cristiano et al. (2000). The two most important success factors for all studies, management support and collect new information from customers, are not recogr Blackwell Publishing Ltd. 2005

7. Discussion of research ndings

In the study of success factors for different kind of company activities, management support is often identied as a prerequisite for success. It is thus no surprise that the success factor secure necessary management support receives the highR&D Management 35, 4, 2005

423

Thomas Lager est weighting in this study too. It is often recognised that the introduction of new working methods and methodologies is a top down affair and not bottom up. Collection of new customer information is always preached in QFD textbooks, and the creation of the Voice of the Customer is thus a methodological cornerstone, that in this study, has been veried to be the second most important success factor and thus of vital importance to successful QFD projects. The third highest ranked success factor, select committed team members is something that is easily understood to be highly important, as all QFD projects should be carried out in cross-functional groups. QFD projects may thus serve as a facilitating organisational tool to overcome company functional barriers. It is also likely that the cross-functional teams will collect not only new customer information but also information that is not so market oriented and more product-development oriented. The results from the study of expected outcomes are somewhat surprising. One of the most frequently cited arguments for why companies should introduce and use QFD is an expected reduction of time to market. The title of one of the rst published books on QFD in America was Better design in half the time (King, 1987). This slogan about the Japanese experience seems to have developed into some sort of myth of huge proportions. There is, however, no scientic evidence for this argument. Not even in the report by Akao (Akao et al., 1987) that is referred to by King in his book is there any evidence of such an expected outcome. One could possibly expect that after developing a comprehensive QFD matrix on a product, the development of the next product generation ought to be faster, but there is only weak support even for this result in a few studies. One can only guess that companies that have embarked on the QFD road in order to speed up product development have been greatly disappointed. So let us leave this misconception behind once and for all. There is also practically no support at all for the likelihood of obtaining lower manufacturing costs or lower design costs with QFD surprisingly not even in Japan, where QFD systems often include cost deployment. The very positive message to companies is however that they can expect better products and product design and better overall customer satisfaction. In the knowledge era (Nonaka and Takeuchi, 1995) the message is also that QFD will improve company communication. The strength of the communication tool 424
R&D Management 35, 4, 2005

did get a high weighting, but the weighting was even higher for retrievable and accumulated company knowledge. It is noteworthy that the above-mentioned expected outcomes were supported by most of the studies in this analysis. The inuence of time and space on research results is seldom discussed. The results from the comparison of America and Japan show, however, that we clearly have such an inuence. The longitudinal effect is not clear and probably needs to be studied with more qualitative tools, probably starting at a company level. Do the expected outcomes, as reported from this study, justify the use of the QFD methodology? One nally comes back to the framework in previous chapters and the issue of methodological productivity. This question has not been addressed or answered in any of the studies, possibly because of the lack of such a framework for the assessment of methodologies.

8. Implications for industry and academic research

The hierarchically structured lists of success factors and expected outcomes of the use of QFD methodology are research ndings that may be useful for industry professionals in their own process of understanding and targeting what types of results they would like to achieve using the QFD methodology and the possible necessary types of behaviour that might be associated with such a success. Using the list of expected outcomes, an internal importance rating for the specic QFD project could be made. Using an importance rating of individual success factors may give more focus for the project team in their QFD endeavour. The structure of expected outcomes and success factors could be used in further empirical studies of the usability of the QFD methodology not necessarily only as they are presented here. but to stimulate even better total structures, including new possible important outcomes and integrating new ways to behave successfully in the use of the methodology. The results from all studies represent, to a large extent, the usability of QFD in large manufacturing industry. The use and adaptation of the methodology to different sectors of industry and to smaller organisations is probably an interesting avenue to follow in further research.
r Blackwell Publishing Ltd. 2005

Industrial usability of quality function deployment

9. Conclusions
The results from this meta-analysis denitely dismiss the old myth that QFD will give shorter time-to-market in product development. The good news is that one can expect to develop a better product that will give customers better overall satisfaction! Not only that, but it shows that the methodology confers several intangible benets especially in the area of better communication and retrievable information. One can only exact that the promotion and introduction of QFD in the future will be based on those wellproven benets and not give companies false expectations. The answer to the rst research question, i.e. whether one can expect improved outcome of product development, is yes. Expected outcomes have been identied, but the question still remains whether these outcomes match the necessary use of company resources. The excessive cost of products that fail in the market (Cooper, 1993) suggests that this actually could be the case. The answer to the second question, whether it is protable to invest in QFD methodology, is probably yes. A framework for the assessment of development methodologies has been created that has proven useful for this study and expectedly also for the assessment of other methodologies for company R&D. A meta-analysis research methodology has been used, and the QFD methodology as such has been used as a tool to facilitate the cross-study analysis and comparison. This has proved to be an interesting research approach that may be helpful in other research projects in the further effort to accumulate research ndings across studies.

References
Akao, Y. ed. (1990) Quality Function Deployment: Integrating Customer Requirements into Product Design. Cambridge: Productivity Press. Akao, Y., Ohfuji, T. and Naoi, T. (1987) Survey and Reviews on Quality Function Deployment in Japan. In Proceedings of ICQC 1987, Tokyo. American Supplier Institute (1989) Quality Function Deployment Three Day Workshop. Dearborn: American Supplier Institute Inc. American Supplier Institute (1992) QFD Kundorienterad produktutveckling hardvara & process. Dearborn: American Supplier Institute Inc. Baglieri, E., Stabilini, G. and Zamboni, S. (2002) How to balance QFD and the efciency of the new product development process; experiences and practices from the eld. In: Herzwurm, G. (ed.), Proceedings of the Eighth International Symposium on Quality Function Deployment. Munich: QFD Institute. Brinkerhoff, R.O. and Dressler, D.E. (1990) Productivity Measurement, 19. London: SAGE Publications. Comstock, T.N. and Dooley, K. (1998) A tale of two QFDs. Quality Management Journal, 5, 3245. Cooper, R.G. (1993) Winning at New Products. New York: Addison-Wesley Publishing Company. Cristiano, J.J., Liker, J.K. and White, C.C. (2000) Customer-driven product development through quality function deployment in the U.S. and Japan. Journal of Product Innovation Management, 17, 286308. Cristiano, J.J., Liker, J.K. and White, C.C. (2001) Key factors in the successful application of quality function deployment (QFD). IEEE Transaction on Engineering Management, 48, 8195. Ekdahl, F. and Gustafsson, A. (1997) QFD: The Swedish experience. In: Gustafsson, A., Bergman, B. and Ekdahl, F. (eds.), Transactions from The Ninth Symposium on Quality Function Deployment. Novi: QFD Institute, pp. 1529. Grifn, A. (1992) Evaluating QFDs use in US rms as a process. Journal of Product Innovation Management, 9, 171187. Hauser, J.R. and Clausing, D. (1988) The house of quality. Harvard Business Review, MayJune, 6373. Hunter, J.E., Schmidt, F.L. and Jackson, G.B. (1982) Meta-Analysis Cumulating Research Findings Across Studies. Beverly Hills: Sage Publications. IDEACore (2003) QFD Designer. Birmingham: IDEACore. King, B. (1987) Better Design in Half the Time Implementing QFD Quality Function Deployment in America. Methuen: GOAL/QPC. Lager, T. and Horte, S.-A. (2002) Success factors for improvement and innovation of process technology in Process Industry. Integrated Manufacturing Systems, 13, 158164. Lindemann, U. (2002) Efciency and effectiveness of working methods. In: Herzwurm, G. (ed.), ProR&D Management 35, 4, 2005

Acknowledgements
Many thanks are due to Kurt Falk (Volvo Car Corporation), for reading and reviewing the manuscript and suggesting useful improvements. My sincere thanks are due to Professor Takao Fujiwara, who provided a creative research environment during my stay as a visiting researcher at the Toyohashi University of Technology in Japan. Many thanks are also due to Professor Yoji Akao and Professor Kozo Koura, for supplying material on QFD and arranging visits to industry. Financial support from Jan Wallanders and Tom Hedelius research foundation is gratefully acknowledged.
r Blackwell Publishing Ltd. 2005

425

Thomas Lager
ceedings of the Eighth International Symposium on Quality Function Deployment. Munich: QFD Institute. Lipsey, M.W. and Wilson, D.B. (2001) Practical MetaAnalysis, Vol. 49. Thousand Oaks: Sage Publications Inc. Martins, A. and Aspinwall, E.M. (2001) Quality function deployment: an empirical study in the UK. Total Quality Management, 12, 575588. Nonaka, I. and Takeuchi, H. (1995) The Knowledge-Creating Company. Oxford: Oxford University Press. Norell, M. (1992) Advisory tools and co-operation in product development. In Department of Machine Elements. Stockholm: Royal Institute of Technology. ISSN 0282-0048 Paterson, B.L., Thorne, S.E., Canam, C. and Jillings, C. (2001) Meta-Study of Qualitative Health Research. Thousand Oaks: Sage Publication Inc. Rosenthal, R. (1991) Meta-Analytic Procedures for Social Research, Vol. 6. Newbury Park: Sage Publications. Tennant, G. (2002) Design for Six Sigma. Aldershot: Gower Publishing Limited. Vonderembse, M., Van Fossen, T. and Raghunathan, T.S. (1997) Is quality function deployment? Good for product development forty companies say yes. Quality Management Journal, 4, 6579.

426

R&D Management 35, 4, 2005

r Blackwell Publishing Ltd. 2005