designing new products and services

International Journal of Market Research Vol. 51 Issue 6

819© 2009 The Market Research Society

DOI: 10.2501/S1470785309200980

A framework for designing new products and services

Rubén Huertas-GarcíaBarcelona UniversityCarolina Consolación-SeguraTechnical University of Catalonia

Customer satisfaction is an important objective in all areas of business and services . A key issue in today’s design activities is to achieve customer satisfaction in an economical way by finding the attributes that are most valuable to customers . In this paper we propose a formal and efficient methodology to design a new service, which is an improvement on a platform service . We propose a methodology to link two tools – the statistical design of experiments (SDE), for data collection, and quality function deployment (QFD), for the development of conceptual alternatives . The focus is only on functional dimensions, but it can be used in symbolic and aesthetic dimensions . The study uses a recent survey on the development of an operations management course curriculum to illustrate the conjoint methodology .

Introduction

A company designs its products and services to distinguish them from competitors’ offerings as well as to conform to changing customer tastes and to meet legal requirements . Teams of people from diverse backgrounds (such as marketing, operations, consumer groups and supplier goods) are brought together to design new services . Marriott, for example, successfully used the project team approach to design its line of economy hotels (Murdick et al. 1990) . Studies in the management of technology suggest that cooperation and communication among marketing, manufacturing, engineering and R&D leads to greater new product success and higher product profitability (Griffin & Hauser 1992) . This is because specialists in each of these fields provide knowledge about the dimensions of attributes that configure a product: engineers have knowledge about usability,

Received (in revised form): 3 July 2009


820

designers about style and marketing professionals about symbolism (Rafaelli & Vilnai-Yavez 2004) .

The term ‘design’ expresses the different aspects of the structure of an object, as well as the choice of various parameters by means of which the object is created (Rindova & Petkova 2007) . Design serves an array of purposes: it is used in existing or new markets to develop new products and improve the quality of existing products or services . Service and product design literature explains this broad range of possibilities (see, e .g ., Ulrich & Eppinger 2004) . A new product design process is usually managed by project . Wheelwright and Clark (1992) divide new product projects into two types: fundamental research or commercial development projects . The latter projects are divided into three categories: breakthrough, platform and derivative projects . This spectrum can be separated into two categories of project, where projects in each category have broadly similar development challenges . On the one hand, there are novel products, breakthrough and platform projects, and, on the other hand, derivative products (Tatikonda 1999) . Other authors also suggest a similar classification into two groups: radical and incremental innovations (Adler et al . 1995; Rindova & Petkova 2007) .

According to Wheelwright and Clark (1992), a project to develop a novelty product involves significant changes to existing products and processes, and has higher levels of technology development and greater market uncertainty . This product serves as a potential basis for the development of other variations, the so-called derivative product, where projects have smaller and limited changes, and, therefore, require fewer resources for their development (Tatikonda 1999) . Marketing uncertainty arises from the degree of technological change affecting the extent to which an innovation is likely to be perceived as incongruous by customers (Rindova & Petkova 2007) . In a series of case studies, Dougherty (2001) observes that successful product innovation is a creative process involving successive cycles of learning by customers and producers . She notes that, in markets for novel products, customers may not be able to articulate their needs, and that these needs may change over time as they learn to use the products .

From the product/service life cycle theory perspective, new platform projects are more likely to occur early in the first stage of the product/process life cycle, whereas derivative projects are more likely to occur later in the life cycle (Bloch 1995) .

The objective of service designs is to satisfy the customer’s needs and expectations . The marketing department is usually responsible for


821

obtaining data from customers, using informal or formal market research, with the aim of understanding and identifying customers’ needs and expectations (Zirger & Hartley 1996) . The descriptions of these needs are the performance specifications . Performance specifications describe what types of things (and to what degree) the product must provide to the customer; design specifications describe how the product does this . Following Rafaelli and Vilnai-Yavetz (2004), these performance specifications can be functional, symbolic and aesthetic dimensions, and all these dimensions interact to affect the customer’s perception of value (Rindova & Petkova 2007) . Design specifications determine what type of service will meet these performance specifications, bearing in mind that even physical artefacts can generate emotional responses in customers (Bitner 1992) . These specifications are later translated into operations such as input to provide a product or service for customers .

This study proposes a formal and efficient methodology for improving an existing service: a derivative service from a platform design . We propose a methodology to link the use of statistical design of experiments (SDE), for data collection, and quality function deployment (QFD), for the development of conceptual alternatives . Pullman, Moore and Wardell (2002) propose the theoretical relationship between conjoint analysis (a technique derived from SDE) and QFD as a new product designer tool, but they do not develop this . In this work, we focus our study only on functional dimensions, without taking into account the other dimensions that are equally important in designing a new service . For instance, symbolic dimensions are in close relationship with brand and perceived image of the service; aesthetic dimension refers to the choice of colour, material and proportions of a product, and is also known as industrial or aesthetic design (Bloch 1995) . In services, the aesthetic dimension usually relates to servicescapes . Servicescapes are the architectural space and surroundings in which customers receive the service (Bitner 1992) .

In order to illustrate this procedure, we took a project for the improvement of a master’s programme . The Department of Business Administration of the UPC (Technical University of Catalonia) has been trying, for the past nine years, to teach a master’s programme that could be considered to offer a complete picture of operations management (OM) . The master’s programme, ENGIPLAN, is at present being run in the four biggest cities in Spain – Barcelona, Valencia, Bilbao and Madrid – with an average of 100–125 students per programme . Every year, at the end of the course, students evaluate the ENGIPLAN programme characteristics in a questionnaire that is supplied to them, and make suggestions about


822

areas and content they consider should be improved . The design for the improvement of the service was set out on the basis of these suggestions .

The current study is structured in six parts: the next section describes the design process for new services . The section after that deals with research, and the one that follows it defines the statistical design of the experiments . After that, we look at the empirical application and the results, following which the QFD application is described . The work ends with our conclusions .

The design process

The product design process has been studied over the years and can be described more precisely than that of services; the design of services has received less attention than that of products (Jeong & Oh 1998) due to the misconception that services are more simplistic than manufactured products (Murdick et al . 1990) . Certainly there is no reason why the design of services cannot be approached in the same highly formalised, structured manner as the design of manufactured systems and products . The design process is often controlled by a procedure known as project management (Dougherty 2001) .

Differences exist between the characteristics of the project and the team that undertakes it . Specifically, platform or novelty projects require more contingency planning effort and greater project management involvement in setting objectives, so they have higher levels of team overlap and integration, higher degrees of project-based performance evaluation of personnel, new engineering tools and less formality of the product development process than in the derivative product . Thus, projects for derived products require a greater degree of adaptation to customer needs, market segmentation, a greater formalisation and use of engineering instruments, as well as the training of the design team in their use (Tatikonda 1999) .

Nevertheless, the literature documents the existence of a formal process for new product design as a key factor in its success (Cooper & Kleinschmindt 1991) . Many service companies often employ design teams when developing a new service, although they do not usually follow a formal process (Murdick et al . 1990) . Few research programmes provide company-specific guidelines for how services must be designed to meet the quality standards demanded by customers (Jeong & Oh 1998) .

Customers usually buy products or services integrated in packages . However, customers not only buy a pack of services – their purchase and


823

consumption behaviour is a consequence of the interaction of three factors: decision unit features, product attributes and situational characteristics . Furthermore, product and service attributes are defined in three types: aesthetic, symbolic (marketing characteristics) and design features . The design features offer a combination of functional benefits or service to the consumer . In mature markets, consumers have acquired some experience and knowledge about the product; as a consequence, they have a well-established perception with regard to the functionality offered by products and their marketing attributes (Romero & Yagüe 2005) .

The process for designing a service involves the following steps: (1) accumulating information; (2) developing conceptual alternatives; (3) screening the concepts; (4) preliminary design; (5) evaluation and improvement; (6) prototyping and final design .

The concept generation stage starts with an idea for a service, which can come from sources outside the organisation, such as customers or competitors, or within the organisation: sales and front-office staff, or the R&D department . In services, the main source of ideas is usually competitors (Murdick et al . 1990) . There are many marketing research tools for gathering data from customers, but for a new service these tools have a limited use (Slack et al . 1998; Tatikonda 1999) .

There are two focuses of building concepts: the parameter-based design approach and the needs-based design approach . Although users have the most information concerning their utility function, they normally have only a partial understanding of the technical domain underlying the design problem . On the other hand, while service or product designers typically understand the technical domain well, they have only partial information about users’ preferences (Randall et al. 2007) .

In parameter-based systems, designers build a concept first and then users evaluate the design parameters of the product . With needs-based systems, users specify the relative importance of their needs first and then designers propose the parameter that is likely to maximise the user’s utility . Randall et al . (2007) show that parameter-based systems are better for expert users, while needs-based systems are better for novice users .

In this study, we follow a needs-based system, and for customised products a design problem can be considered as the search for a set of values for the product design parameters that maximise users’ utility (Ulrich & Eppinger 2004) . Gathered by marketing research, these ideas, or so-called performance specifications (the vector what), should be transformed into design specifications (the vector how) so the organisation can then evaluate and operationalise them . Concepts are different from


824

ideas because they are clear statements that both encapsulate the idea and also indicate the overall form, function, purpose and benefits generated . For example, the broad idea of an adventure holiday is further defined to include its duration, purpose, and residential nature and target market .

Concepts may have to pass through many different screens from several departments in order to ensure that, in general terms, they will be an excellent addition to the product portfolio (Bloch 1995) . The designers must select the concepts that will pass the following phases until they form a preliminary design of package and process . Time and cost limit the number of concepts that a firm can develop in detail (Srinivasan et al. 1997) .

The result of these two first steps is an acceptable concept that can be transformed into a preliminary design . The objective at this stage is to make a first attempt at:

• specifying the component services in the package• defining the processes needed to create the package – that is, the order

in which the component parts of the package must be put together .

This preliminary design passes through a stage of evaluation and improvement to determine if the concept can be developed in a better, cheaper and easier way .

When there is a consensus design it can become a prototype and a final design . Service prototypes can be demonstrated as computer simulations or as an implementation of the service on a pilot basis (Slack et al . 1998) .

These stages describe the process of designing a new product for the market; that is to say, for platform product projects . However, if the intention is to design a derivative product, the point of departure is from a preliminary design and, therefore, the design process is shorter .

The research process

In this work we depart from a preliminary design that must be evaluated and improved . This design is focused only on functional features, and not on aesthetic and symbolic (marketing) characteristics . Marketing characteristics are derived from marketing management: prices, promotions, brands, sizes, packaging, and so on (Romero & Yagüe 2005) . In order to carry out a survey, we should first determine an appropriate research population – the target market – and a proper methodology for the sampling procedure . There are many customers for an OM course: society, the industries that employ our alumni, former students and, finally, current students . In work that precedes


825

ours, by Gustafsson et al. (1999), the authors used former students, at least two years after finishing their master’s course, as customers because they considered that, during this period of time, they were able to evaluate the education they had received and compare it with their professional needs . Current students are of lesser importance since they have only just come into contact with the OM master’s programme and are thus not able to evaluate its utility . In Spain, however – particularly in the main cities – current students have become the main customers, because public and private universities are in competition to attract students . As a target market, we consider the current students following the ENGIPLAN master’s in Barcelona during the first quarter of the course .

In order to determine service attributes, Griffin and Hauser (1993) propose focus group and personal interview methods as a means of deriving the main customer service attributes . In our research, the service attributes are derived in three steps: first, we identify a list of customers’ needs from a revision of previous surveys; the second step consists of an experimental design to establish customer need priorities in terms of importance weights; while, in the third step, we confirm the customer need priorities, and break the confusion pattern for all the main effects from the two-factor interactions by means of a second experiment .

The research process began with a revision of the questionnaires filled out by students on previous courses, with the aim of identifying attributes that customers consider important for improving service quality . The OM master’s evaluation questionnaire contained an open question inviting suggestions . These suggestions were codified using the inductive method for the creation of categories, which consists of labelling repeated factors found in the text (Spiggle 1994) . Similar processes are used in the analysis of content in the service literature (Tax et al. 1998), and the method is used extensively in research into consumer behaviour to identify and relate subjects in different passages of text (Spiggle 1994) .

In selecting the attributes we follow the recommendations of Gustafsson et al . (1999), who state that the attributes can be important for interviewees in the purchase situation, that they can be modified, and that they allow comparisons with competitors . Furthermore, in the case of mature categories of products, where choice is heavily influenced by functionality, usage contexts are a key factor in new product development (Shocker & Srinivasan 1979) .

Bearing in mind the previous considerations, the numbers of relevant attributes, from the customer’s perspective, are seven primary needs, each with two secondary needs . These secondary needs are presented as a semantic differential, and Table 1 shows its transformation into variable dummies .


826

Statistical design of the experiment (SDE)

A statistical design of the experiment (SDE) is commonly used in engineering studies and medical analysis, although its origin is in agricultural research (Box et al. 1999) . However, SDE can also be used outside the scope of engineering and manufacturing production – for example, in service operations (Starkey et al. 1997) or marketing and sales (Chevalier 1975; Starkey et al . 1997; Huertas-García & Consolación-Segura 2009) .

Assuming that a product or service is definable as a vector in a multidimensional attribute space, and that the evaluation of the product or the service is based on its attribute levels, it becomes theoretically possible to relate preference to attributes .

Typically, an SDE is carried out using hypothetical descriptions of the service, or so-called stimuli . In this context, a stimulus is defined as the presentation of the attribute’s levels to the respondent . Data for the statistical design of experiments may be collected by three types of

Table 1 Scheme of the transformation of variables into dummies

(1) Characteristics of the OP master’s programme(–) Complete overview of operations management, with some insertions of other business management subjects (marketing, finance, etc.)(+) Greater repercussion of other business management subjects (marketing, finance, etc.)

(2) Characteristics of the operation management programme focus(–) General view of industrial operations management, with some insertions in other areas: different types of industrial process, services, etc.(+) Greater repercussion in the other specific areas: different types of industrial processes, services, etc.

(3) Approach of the master’s(–) An instrumental approach, greater attention to knowledge of tools and quantitative techniques(+) A wider approach to knowledge of business operations management and leadership issues

(4) Characteristics of the teaching staff(–) Educational teaching staff with professional experience(+) Professional teaching staff with educational experience

(5) Characteristics of the type of work(–) Predominantly teamwork(+) Predominantly individual work

(6) Structure of classes(–) More emphasis on lectures, with some discussion of cases(+) More emphasis on discussion of cases, with some lecture classes

(7) Knowledge of computer tools(–) A general view of computer software related to operations management(+) A deeper knowledge of some specific software packages


827

stimulus presentation: (1) verbal, (2) paragraph (descriptive cards), and (3) pictorial or in-kind presentation . Several researchers have criticised the conjoint approach to studying design reactions because it usually employs verbal descriptions of a product or service, whereas visual cues for examining design reactions are used less often (Holbrook & More 1981) . However, this problem is especially acute in situations involving aesthetic elements of form, but not in functional attributes, which are more rational (Hirschman 1983) . These stimuli describe distinct concepts and will be assessed by respondents . This research study follows the recommendation of Gustafsson et al . (1999) for presentation of MBA services, which constitutes a scenario . A scenario is a verbal description of situations where attributes are presented in an implicit form within the context . An example appears in the Appendix .

The SDE involves asking consumers to rank and rate in order of preference different product or service alternatives . Consumers make their consumption decisions based on a joint comparison of different attributes . A purchaser chooses an alternative by following a compensatory process similar to that proposed by Miller and Ginter (1979) . The SDE assumes that a consumer assigns a utility value to each level of each attribute, and makes his or her final decision based on the total utility values across attributes for a given product .

The seven outstanding attributes in the OM master’s had two different values, therefore we will need the attribute combination that maximises the utility of the target market . In order to indicate which of these seven attributes determines the choice of the best master’s programme, we propose an experiment using a fractional factorial design on two levels . Table 1 shows their transformation from variables into dummies . This represents factorial experiment 27 – that is to say, 128 elementary experiments . In order to avoid saturating the interviewees we have used only a fraction (in this case, eight experiments) . This design is known as ‘three resolutions’ (R = III) and is a fraction 8

1281

16= of full factorial 27 . It is a 2III

7–4 design . Although this design does not mutually confuse the main effects, it does confuse them with the interactions of two factors (Box et al . 1999) .

After designing the experiment, we asked a sample of clients or potential customers to evaluate each of the scenarios on an attitude scale of 1 to 10 .


828

The experimental design

The two-level design is the most widely used where the number of scenarios is a multiple of 2 – for example, 8, 16, 32, and so on (also known as geometric designs) . In this work we use a fractional factorial design of eight scenarios .

Whatever the sample size, for the case of designs with three factors at two levels 23, Schmidt and Launsby (1994) propose eight tests with five repetitions for each one . If averages alone need to be studied, the number of repetitions is half this amount, and if the experiments are in line (in field) the number of repetitions must be multiplied by 10 . In our case we consider eight tests with 20 repetitions for each one . According to Griffin and Hauser (1993), fewer than 30 respondents are typically needed to elicit a majority of relevant product needs . In order to prevent the first scenarios from being considered more important than the following ones, each one of them appears randomly in order not to affect the answer . For the construction of the fractional factorial 2III

7–4, we use the design generator: 4 = 12, 5 = 13, 6 = 23 and 7 = 123, where 1, 2 and 3 are column vectors of signs of the considered variables . When we design a factorial fractional, it is important to obtain the confusion pattern . In our case, the confusion pattern was derived from the design generator and appears in Table 4 .

Several methods are available for calculating the effects – for example, the Yates algorithm, an example of the use of which can be found in Huertas-García and Consolación-Segura (2009) . However, in this work we use the contrasting table of coefficients in which only the eight vector columns are considered – one for the average and the other seven for attributes . This algorithm is very simple and can be carried out using any spreadsheet program . The factorial design is first placed in the standard order for the three first variables . The standard order is presented in Table 2 . The first column (1) shows the alternation of one negative and one positive sign; column (2) shows the alternation consisting of two negative and two positive signs; finally, in column (3), the alternation consists of four negative and four positive signs . In general, column k consists of 2k–1 negative signs followed by 2k–1 positive signs . For the other, fourth, variable column we use the design generator 12 = 4; that is, we multiply vector 1 with vector 2 and we obtain vector 4 . The remaining columns are composed in the same way, as the result of the product of signs that their interaction represents .

As the divisor, we use the number of times that the maximum value of the variable appears, which is determined by the positive sign; in the case of qualitative variables they determine presence or absence . For example,


829

for the average variable the number of positive signs is 8, and therefore this is the divisor; for variable A the number of positive signs is 4, and this is the divisor for the remaining variables .

Once the matrix variable is in order, we arrive at the calculation . Each of the vector columns of signs is multiplied by the vector columns of values

Table 2 fractional factorial design and the averages

Service scenarios (1

) Cha

ract

eris

tics

of th

e O

M m

aste

r’s

prog

ram

me

(2) C

hara

cter

istic

s of

the

OM

pr

ogra

mm

e fo

cus

(3) A

ppro

ach

of

the

mas

ter’s

(4) (

12)

Char

acte

rist

ics

of

the

teac

hing

sta

ff

(5) (

13)

Char

acte

rist

ics

of

the

type

of w

ork

(6) (

23) S

truc

ture

of

cla

sses

(7) (

123)

Kn

owle

dge

of

com

pute

r too

ls

Average

1 –1 –1 –1 1 1 1 –1 5.252 1 –1 –1 –1 –1 1 1 4.503 –1 1 –1 –1 1 –1 1 5.354 1 1 –1 1 –1 –1 –1 6.255 –1 –1 1 1 –1 –1 1 6.356 1 –1 1 –1 1 –1 –1 4.857 –1 1 1 –1 –1 1 –1 7.758 1 1 1 1 1 1 1 6.85

Table 3 Table of calculus of the variables

Average (1) C

hara

cter

istic

s of

OP

mas

ter’s

pr

ogra

mm

e

(2) C

hara

cter

istic

s of

the

OM

pr

ogra

mm

e fo

cus

(3) A

ppro

ach

of

the

mas

ter’s

(4) (

12)

Char

acte

rist

ics

of

the

teac

hing

sta

ff

(5) (

13)

Char

acte

rist

ics

of

the

type

of w

ork

(6) (

23) S

truc

ture

of

cla

sses

(7) (

123)

Kn

owle

dge

on

com

pute

r too

ls

5.25 –5.25 –5.25 –5.25 5.25 5.25 5.25 –5.254.5 4.5 –4.5 –4.5 –4.5 –4.5 4.5 4.55.35 –5.35 5.35 –5.35 –5.35 5.35 –5.35 5.356.25 6.25 6.25 –6.25 6.25 –6.25 –6.25 –6.256.35 –6.35 –6.35 6.35 6.35 –6.35 –6.35 6.354.85 4.85 –4.85 4.85 –4.85 4.85 –4.85 –4.857.75 –7.75 7.75 7.75 –7.75 –7.75 7.75 –7.756.85 6.85 6.85 6.85 6.85 6.85 6.85 6.85

Addition 47.15 –2.25 5.25 4.45 2.25 –2.55 1.55 –1.05Divisor 8 4 4 4 4 4 4 4Estimation 5.893 –0.562 1.312 1.112 0.562 –0.637 0.387 –0.262


830

of the experimental estimated averages . The values of the resulting vector are added together and then divided by the corresponding divisor (Table 3 shows the calculations) .

Table 4 shows the estimated effects of the design, as well as the simplified confusion pattern, in which the interactions between two or more factors are not taken into account . One may also observe that the main effects are confused with the second-order interactions .

In order to break the confusion pattern for all the main effects from the two-factor interactions, the study adds a second fraction in which the signs for all the factors are reversed . This procedure is called full fold-over, and is used in resolution III designs to break the link between main effects and two-factor interactions (Myers & Montgomery 2002) . However, the two-factor interactions are confused in groups of three (Box et al . 1999) . In this second experiment we consider eight tests with 28 repetitions for

Table 4 Table of estimation and the confusion pattern

Estimation Confusion pattern

Average 5.89

(1) Characteristics of the OP master’s programme(–) Complete overview of operations management, with some insertions of other business management subjects (marketing, finance, etc.) –0.56 1→24→35→67

(2) Characteristics of the operations management programme focus(+) Greater repercussions in the other specific areas: different types of industrial processes, services, etc. 1.31 2→14→36→57

(3) Approach of the master’s(+) A wider approach to knowledge of business operation management and leadership issues 1.11 3→15→26→47

(4) (12) Characteristics of the teaching staff(+) Professional teaching staff, with educational experience 0.56 4→12→56→37

(5) (13) Characteristics of the type of work(–) Predominantly teamwork –0.63 5→13→46→27

(6) (23) Structure of classes(+) More emphasis on discussion of cases, with some lecture classes 0.38 6→23→45→17

(7) (123) Knowledge of computer tools(–) A general view of computer software related to operations management –0.26 7→34→25→16


831

each one . Table 5 shows the design of these experiments and the average scores for each one .

With the data obtained in this second experiment, we employ the contrasting table of coefficients for estimating the average and the seven attribute variables . This second fractional factorial evidence of internal consistency exists in the experimental methodology . The results of this second experiment confirm the results of the first; six of the seven variables have the same sign, and only one – ‘knowledge of computer tools’ – has a changed sign . However, this last factor is the smallest statistical representative .

Last, we calculate the standard deviations for determining the degree of significant influence on the factors considered . In order to calculate the variance of the effects, we use the average variance of the experiments following the formula proposed by Box et al . (1999):

Table 5 factorial fractional symmetric experiment

Estimation Confusion pattern

Average 4.70

(1) Characteristics of the OP master’s programme(–) Complete overview of operations management, with some insertions of other business management subjects (marketing, finance, etc.) –1.10 1→24→35→67

(2) Characteristics of the operations management programme focus(+) Greater repercussions in the other specific areas: different types of industrial processes, services, etc. 0.43 2→14→36→57

(3) Approach of the master’s(+) A wider approach to knowledge of business operations management and leadership issues 0.38 3→15→26→47

(4) (12) Characteristics of the teaching staff(+) Professional teaching staff, with educational experience 1.25 4→12→56→37

(5) (13) Characteristics of the type of work(–) Predominantly teamwork –1.61 5→13→46→27

(6) (23) Structure of classes(+) More emphasis on discussion of cases, with some lecture classes 0.59 6→23→45→17

(7) (123) Knowledge of computer tools(–) A general view of computer software related to operations management 0.16 7→34→25→16


832

Variance of the effects = 4N average variance of experiments

where N = number of experiments × number of repetitions, and the average variance of the experiments is the sum of the variances divided by the number of experiments . Table 6 shows the combined results of

the estimated values and the standard deviation of both experiments, considering that the standard deviation of the average is half the standard deviation of the effect .

Once the main performance specifications are set, the next step is to transform them into design specifications using QFD .

The methodology of quality function deployment (QFD)

Quality function deployment (QFD) is a product or service development process based on inter-functional teams (marketing, manufacturing, engineering and R&D) who use a series of matrices, which look like

Table 6 Estimation and standard deviation of the effects of both experiments

Standard Estimation deviation

Average 5.19 0.16

(1) Characteristics of the OP master’s programme(–) Complete overview of operations management, with some insertions of other business management subjects (marketing, finance, etc.) –0.88 0.33

(2) Characteristics of the operations management programme focus(+) Greater repercussions in the other specific areas: different types of industrial processes, services, etc. 0.79 0.33

(3) Approach of the master’s(+) A wider approach to knowledge of business operations management and leadership issues 0.68 0.33

(4) (12) Characteristics of the teaching staff(+) Professional teaching staff, with educational experience 0.96 0.33

(5) (13) Characteristics of the type of work(–) Predominantly teamwork –0.94 0.33

(6) (23) Structure of classes(+) More emphasis on discussion of cases, with some lecture classes 0.50 0.33

(7) (123) Knowledge of computer tools(–) A general view of computer software related to operations management –0.01 0.33


833

houses, to deploy customer inputs throughout design, manufacturing and service delivery . Mitsubishi Heavy Industries Ltd developed QFD at the Kobe shipyards in the 1970s . A wide range of Japanese companies subsequently adopted it, including the Toyota Motor Company (Akao 1993) . Don Clausing of the Xerox Company, and later MIT, and Bob King of GOAL/QPC introduced this technique to the West in the mid-1980s (Cohen 1995) . QFD has basically been used as an instrument for improving the design of products, providing savings of up to 60% in design costs and 40% in design time (Hauser & Clausing 1988) .

QFD also has its origins in industrial activity, most of the applications being in manufacturing companies, for the design of new products, using information from outside the organisation, such as from consumers . Service companies were quick to adopt it (Akao 1993), not only to satisfy external customer needs but also for internal consumers (Natarajan et al. 1999) .

From the perspective of marketing, QFD is an interesting tool because it encourages other functions besides marketing, and in some cases incorporates market research . Engineers or designers require more detail about customer needs than is provided by the typical marketing study . Such detail is necessary to make specific trade-offs in translating performance specifications to design specifications . Developing products based on the voice of the customer becomes a key criterion in total quality management (Griffin & Hauser 1993)

In what follows, we present and highlight only some of the indicators of the QFD matrix, with the aim of improving the ENGIPLAN master’s programme run at UPC .

The point of departure is the voice of the customer, which is represented in the what vector, and in this case is extracted from the statistical design of experiments (SDE) (represented in the first two text columns in Figure 1) . The market assigns each of the requirements a position of importance on a scale of 1 to 5 (where 1 = least important and 5 = most important; see Figure 1, column 13: ‘Current satisfaction performance’) . In our case, the most important

Table 7 Technical characteristics of the proposed service

• Operations management and operations strategy

• Job design and work organisation• The design of product, processes and

layout• Quality planning, control and operations

improvement• Capacity and inventory planning and

control, and logistics• Maintenance and failure prevention• Complementary subjects: cost

accounting, economy, marketing• Software lessons• Visits to companies (4 visits)• Lectures (6)• Group work


834

features are the ‘Characteristics of the teaching staff ’ and ‘Characteristics of the type of work’ .

Next, we must gather, arrange and classify the technical characteristics of the service, ENGIPLAN master’s programme, collected in the vector process, or the how vector (Figure 1, first row or Table 7) . A technical difficulty is assigned to each of the technical characteristics, which is expressed with a numerical value indicating the existing complexity involved in attaining an objective . The scale on which it is based is also from 1 to 5, with 5 signifying the greatest difficulty . This indicator is important as it means planning or improving the service at the lowest possible cost (Figure 1, row 8) .

The ratio between customer demands and characteristics of service quality is reflected in the matrix of ratios, where we measure the compatibility between what the customer wants and what the service offers . In order to establish the relationship, we use a geometric scale of values 1, 3 and 9, where 9 indicates a close relationship between what and how, 3 an average relationship, and 1 a weak relationship (Figure 1: 7 rows × 11 columns) . We focus on the close or strong relations later, when we perform the diagnosis of the matrix .

A further important indicator is competitive market evaluation (competitive benchmarks), which shows how the service is seen by the market and what our position is compared with that of our competitors . The information used to carry out the competitive evaluation is drawn from different sources of market research . We can establish a scale from 1 to 5 in which, for each of the attributes, customer perception regarding our product or service and that provided by the competition is evaluated, each company being differentiated one from another by a symbol . In our case, the competition considered is a master’s course in management operations run by a prestigious business school in Barcelona . The competitive evaluation shows what in marketing is known as positioning . This refers to the position a service occupies according to consumer perception in relation to other products or services belonging to the competition . A good knowledge of the position that a product or service occupies in the market is especially useful for orientating marketing strategy and for determining the actions necessary to maintain or correct one’s present position . The competitive evaluation, along with the assignment of priorities, reflects an authentic map of positioning . This analysis enables us to identify the strong points (sales points), the weak points and the areas where opportunities for improvement exist; in the first we occupy a favourable position as regards the competition; in the second we occupy an unfavourable position; while


835

Figu

re 1

The

Hou

se o

f Qua

lity

Stro

ng im

pact

Mod

erat

e im

pact

Operations management and operations strategy

Job design and work organisation

The design of product, processes and layout improvement

Quality planning, control and operations

Capacity and inventory planning and control, and logistic

Maintenance and failure prevention

Complementary subjects: cost accountant, economy, marketing

Software lessons

Visits to companies (4 visits)

Lecture ( 6 )

Work in group

Importance to customer

Current satisfaction performance

Competitive satisfaction performance

Goal

Improvement ratio

Sales point

Raw weight

Normalised raw weight

31

99

99

93

53

23

11

50.116144018583043

11

33

19

93

42

331.5

1.2

7.2

0.167247386759582

93

31

34

25

31.5

1.5

90.209059233449477

11

13

35

45

411.2

60.139372822299652

39

52

431.5

1.5

11.25

0.261324041811847

11

33

34

311.2

3.6

0.0836236933797909

33

91

14

34

11

10.0232288037166086

11

11

11

24

55

3

38.5

3344

6627.5

17.5

3322

2238.5

37.5

118.05

59.45

81.45

92.25

59.25

83.85

72.45

38.25

91.05

104.85

69.85

0.18

0.09

0.12

0.14

0.091

0.13

0.11

0.06

0.14

0.16

0.11

Com

petit

ive

benc

hmar

ks

Our

com

pany

Com

petit

or

(–) C

ompl

ete

over

view

of o

pera

tions

man

age-

m

ent,

with

som

e in

sert

ions

of o

ther

bus

ines

s m

anag

emen

t sub

ject

s (m

arke

ting,

�na

nce,

etc

.)

(1) C

hara

cter

istic

s of

the

OP

mas

ter’s

pro

gram

me

(2) C

hara

cter

istic

s of

the

oper

atio

ns m

anag

emen

t pr

ogra

mm

e fo

cus

(+) G

reat

er re

perc

ussi

on o

f oth

er b

usin

ess

man

agem

ent s

ubje

cts

(mar

ketin

g, �

nanc

e,

etc.

)(3

) App

roac

h of

the

mas

ter’s

(+) A

wid

er a

ppro

ach

to k

now

ledg

e of

bus

ines

s op

erat

ions

man

agem

ent a

nd le

ader

ship

issu

es

(4) (

12) C

hara

cter

istic

s of

th

e te

achi

ng s

ta�

(+) P

rofe

ssio

nal t

each

ing

sta�

, with

ed

ucat

iona

l exp

erie

nce

(5) (

13) C

hara

cter

istic

s of

th

e ty

pe o

f wor

k(–

) Pre

dom

inan

tly te

amw

ork

(6) (

23) S

truc

ture

of c

lass

es(+

) Mor

e em

phas

is o

n di

scus

sion

of c

ases

, with

so

me

lect

ure

clas

ses

(7) (

123)

Kno

wle

dge

of

com

pute

r too

ls(–

) A g

ener

al v

iew

of c

ompu

ter s

oftw

are

rela

ted

to o

pera

tions

man

agem

ent

Tech

nica

l di�

culti

esD

esig

n ta

rget

sRa

wW

eigh

tN

orm

alis

ed

0

1

2

3

4

5

6

Figu

re 1

Th

e H

ouse

of Q

ualit

y


836

the third case indicates the market niches or segments where a business opportunity exists due to unsatisfied or insufficiently satisfied needs .

The strong points can be converted into sales points, which serve to establish the attributes we require to draw up our policy of differentiation, in so far as whether we are in a favourable situation as regards the competition; these attributes are also requirements valued by the market .

A further indicator refers to technical difficulty, which is the difficulty we may encounter in improving each one of the hows, and to this end we employ a scale from 1 to 5, where 1 indicates little and 5 great difficulty .

Finally, the matrix of correlations for the technical characteristics, also known as the ‘ceiling of the house of the quality’, enables us to identify which hows can be found within other hows, and which ones are in conflict with each other . In our case, we have identified only positive correlations .

QFD transforms the customer’s requirements into features of quality . The QFD methodology facilitates the evolution of traditional quality systems based on control and orientated only towards results for modern systems, based on planning and focused on the client .

Conclusions

The existence of a formal process for designing new products or services is documented in the literature as a key factor for success (Cooper & Kleinschmindt 1991) . Few research paradigms suggest specific guidelines for translating market demand into a service company’s production process, and this is particularly crucial for service organisations with limited resources (Jeong & Oh 1998) . Moreover, empirical studies of design issues are rare in marketing journals (Bloch 1995) .

In this study, we propose a formal and efficient methodology for improving a service that is already on the market, and which is therefore a project for designing a derivative service from a platform design . For data collection, we used the statistical design of experiments (SDE) and, for the development of conceptual alternatives, we used quality function deployment (QFD), with the aim of integrating technological and market knowledge (Dougherty 2001) . In order to illustrate this procedure, we took the project of improving a master’s programme . This example serves as a demonstration of the necessity for a systematic and structured procedure in service design processes, and the use of these instruments for developing strategies that stimulate the innovation of services, as well as the functional improvement of such services .


837

In addition, we introduce the use of SDE and QFD, and illustrate how these instruments can be used together in a complementary manner to achieve a better understanding of the voice of the consumer and its transformation into attributes that add value to services . The joint application of SDE and QFD has great potential for application, both in product and service designs, because they enable improvement of products or services based on degree of satisfaction perceived by consumers – the consequence of which can be to improve customer satisfaction, retain customers and therefore increase market share . In the case of a service, it is vitally important to determine the added value, since customers value services according to the added value provided . It is also important to detect the requirements that contribute greater value, as well as to determine what implications arise in the allocation of resources, particularly in terms of costs . An important literature base suggests that customers’ perceptions of value are based on cognitive and emotional responses to product or service characteristics created through design choices via the functional, symbolic and aesthetic dimensions . One of the limitations of our study is the consideration of solely functional attributes, without taking into account aesthetic and symbolist attributes, which can be inaccurate and misleading (Strati 1992) . Bruce and Whitehead (1988) propose the use of a ‘design mix’ as a means of satisfying perceived needs in terms of ‘emotional’ responses and ‘rational’ evaluations .

The main limitation, however, concerns the number of factors and levels that can be used, which are generally fewer than eight . Therefore, the procedure should be used in later phases of the service design, when a series of defined attributes are available . Although the QFD process may appear complicated if used in a complete package service, it can be used at the beginning in a core service or in a supporting service . Furthermore, the designer may expand QFD by including new attributes or a new service, and if QFD is built with a spreadsheet program, as in this work, revisions and extensions are easy to accommodate .

Finally, our ideas also contribute to innovative research focusing on the importance of producer–customer interactions in the creation of new services . A natural extension would be the use of response surface methodology to determine the optimal combinations of attributes that will maximise customer utility .


838

Appendix: Example of scenario

Scenario 1 Evaluation

(1)CharacteristicsoftheOMmaster’sprogramme

(–) Complete overview of operations management, with some insertions of other business management subjects (marketing, finance, etc.)

(2) Characteristics of the OM programme focus

(–) General view of industrial operations management, with some insertions in other areas: different types of industrial processes, services, etc.

(3)Approachofthemaster’s (–) An instrumental approach, greater attention to knowledge of tools and quantitative techniques

(4) Characteristics of the teaching staff

(+) Professional teaching staff, with educational experience

(5) Characteristics of the type of work

(+) Predominantly individual work

(6) Structure of classes (+) More emphasis on discussion of cases, with some lecture classes

(7) Knowledge of computer tools (–) A general view of computer software related to operations management

Name: .................................................................................................................................

Age: ......................................................... Sex: ..................................................

ReferencesAdler, P .S ., Mandelbaum, A ., Nguyen, V . & Schwerer, E . (1995) From project to process

management: an empirically-based framework for analyzing product development time . Management Science, 41(3), pp . 458–484 .

Akao, Y . (1993) Despliegue de funciones de calidad . Integración de las necesidades de los clientes en el diseño del producto . Editorial TGP-Hoshin Sc . cop . Madrid .

Bitner, M .J . (1992) Servicescapes: the impact of physical surrounding on customers and employees . Journal of Marketing, 56 (April), pp . 57–71 .

Bloch, P .H . (1995) Seeking the ideal form: product design and customer responses . Journal of Marketing, 59, pp . 16–29 .

Box, G .E .P ., Hunter, W .G . & Hunter, J .S . (1999) Statistics for Experimenters: An introduction to Design, Data Analysis, and Model Building . New York: John Wiley & Sons, Inc .

Bruce, M . & Whitehead, M . (1988) Putting design into the picture: the role of product design in consumer purchase behaviour . Journal of the Market Research Society, 30(2), pp . 147–162 .

Chevalier, M . (1975) Increase in sales due to in-store display . Journal of Marketing Research, 12 (November), pp . 426–431 .

Cohen, L . (1995) Quality Function Deployment: How to Make QFD Work for You . Reading, MA: Addison-Wesley Publishing Company .

Cooper, R .G . & Kleinschmindt, E .J . (1991) New product processes at leading industrial firms . Industrial Marketing Management, 20(2), pp . 137–147 .


839

Dougherty, D . (2001) Reimagining the differentiation and integration of work for sustained product innovation . Organization Science, 12, pp . 612–631 .

Griffin, A . & Hauser, J .R . (1992) Patterns of communication among marketing, engineering and manufacturing – a comparison between two new product teams . Management Science, 38(3) (March), pp . 360–373 .

Griffin, A . & Hauser, J .R . (1993) The voice of the customer . Marketing Science, 12(1) (Winter), pp . 1–27 .

Gustafsson, A ., Ekdahl, F . & Bergman, B . (1999) Conjoint analysis: a useful tool in the design process . Total Quality Management, 10(3), pp . 327–343 .

Hauser, J .R . & Clausing, D . (1988) The house of quality . Harvard Business Review, 65(3), pp . 63–73 .

Hirschman, E .C . (1983) Aesthetic, ideologies and the limits of the marketing concept . Journal of Marketing, 45 (Summer), pp . 45–55 .

Holbrook, M .B . & Moore, W .L . (1981) Feature interactions in consumer judgments of verbal versus pictorial presentation . Journal of Consumer Research, 8, pp . 103–113 .

Huertas-García R . & Consolación-Segura, C . (2009) Using Statistical Design of Experiments (SDE) Methodologies? International Journal of Market Research, 51, 1, pp . 113-134 .

Jeong, M . & Oh, D . (1998) Quality function deployment: an extended framework for service quality and customer satisfaction in the hospitality industry . International Journal of Hospitality Management, 17, pp . 375–390 .

Miller, K .E . & Ginter, J .L . (1979) An investigation of situational variation in brand choice behaviour and attitude . Journal of Marketing Research, 16 (February), pp . 111–123 .

Murdick, R .G ., Render, B . & Russell, R .S . (1990) Service Operations Management . Newton, MA: Allyn & Bacon .

Myers, R .H . & Montgomery, D .C . (2002) Response Surface Methodology . New York: John Wiley & Sons, Inc .

Natarajan, R .N ., Martz, R .E . & Kurosaka, K . (1999) Applying QFD to internal service system design . Quality Progress, 32(2), pp . 65–71 .

Pullman, M .E ., Moore, W .L . & Wardell, D .G . (2002) A comparison of quality function deployment and conjoint analysis in new product design . Journal of Product Innovation Management, 19, pp . 354–364 .

Rafaelli, A . & Vilnai-Yavetz, I . (2004) Emotion as a connection of physical artifacts and organizations . Organization Science, 15, pp . 671–686 .

Randall, T ., Terwiesch, C . & Ulrich, K .T . (2007) User design of customized products . Marketing Science, 26(2), pp . 268–280 .

Rindova, V .P . & Petkova, A .P . (2007) When is a new thing a good thing? Technological change, product form design, and perceptions of value for product innovations . Organization Science, 18(2), pp . 217–232 .

Romero, J . & Yagüe, M .J . (2005) Identifying the influence of product design and usage situation on consumer choice . International Journal of Market Research, 47(6), pp . 667–686 .

Schmidt, S . & Launsby, R . (1994) Understanding Industrial Designed Experiments (4th edn) . Air Academy Press . Cited by Escalante, E .J . (2003) Seis sigma . Metodología y técnicas . Ed . Limusa . México .

Shocker, A .D . & Srinivasan, V . (1979) Multiattribute approaches for product concept evaluation and generation: a critical review . Journal of Marketing Research, 16 (May), pp . 159–180 .

Slack, N ., Chambers, S ., Harland, C ., Harrison, A . & Johnston, R . (1998) Operations Management . London: Pitman Publishing .

Spiggle, S . (1994) Analysis and interpretation of qualitative data in consumer research . Journal of Consumer Research, 21 (December), pp . 491–503 .


840

Srinivasan, V ., Lovejoy, W .S . & Beach, D . (1997) Integrated product design for marketability and manufacturing . Journal of Marketing Research, 34, pp . 154–163 .

Starkey, M ., Aughton, J . & Brewin, R . (1997) Extending process thinking: design of experiments in sales and marketing . TQM Magazine, 9(6), pp . 434–439 .

Strati, A . (1992) Aesthetic understanding of organizational life . Academic Management Review, 17(3), pp . 568–581 .

Tatikonda, M .V . (1999) An empirical study of platform and derivative product development projects . Journal of Product Innovation Management, 16, pp . 3–26 .

Tax, S .S ., Brown, S .W . & Chandrashekaran, M . (1998) Customer evaluation of service complaint experiences: implications for relationship marketing . Journal of Marketing, 62 (April), pp . 60–76 .

Ulrich, K .T . & Eppinger, S .D . (2004) Product Design and Development (3rd edn) . New York: McGraw-Hill .

Wansink, B . (1994) Advertising’s impact on category substitution . Journal of Marketing Research, 31 (November), pp . 505–515 .

Wheelwright, S .C . & Clark, K .B . (1992) Competing through manufacturing . Harvard Business Review, 70(2), pp . 70–82 .

Zirger, B .J . & Hartley, J .L . (1996) The effect of acceleration techniques on product development time . IEEE Transactions on Engineering Management, 43(2) (May), pp . 143–152 .

About the authors

Rubén Huertas-García is a professor in Marketing Management in the Economics and Business Administration Department, Barcelona University (UB) . His research interests are applied statistics in business administration and experimental designs in marketing . Carolina Consolación-Segura is a professor in Business Administration at Technical University of Catalonia (UPC), School of Telecommunications Engineering of Barcelona (ETSETB) . Her research interests are in management and marketing activities associated with the design of services .

Address correspondence to: Rubén Huertas-García, Economics and Business Organization Department, Faculty of Economics and Business, Barcelona University, Main Building, Tower 2, 3rd floor . Diagonal 690; 08034-Barcelona (Spain) .

Email: rhuertas@uoc .edu

designing new products and services

Documents