baker (pp. 274-307)

35
PREGRADO UESAN Asignatura: Gestión de Proyectos de Lanzamiento de Productos (Inglés) Lectura: Baker, M. (2007). En: Product strategy and management. (pp. 274-307). 2a ed. Harlow, England: Pearson Educación. Profesor Responsable: Lu Chang-Say, Estuardo. El presente material se pone a disposición de manera gratuita, para uso exclusivo de los alumnos de pregrado de la Universidad ESAN y es sólo para fines académicos, de acuerdo con lo dispuesto por la legislación sobre los derechos de autor. Decreto Legislativo N° 822. En tal sentido, se deja constancia, que la difusión de este documento bibliográfico, está expresamente prohibida, por estar destinado únicamente para uso académico en el presente curso.

Upload: sandracarolina

Post on 09-Feb-2016

226 views

Category:

Documents


1 download

DESCRIPTION

Gestión de proyectos de lanzamiento de producto

TRANSCRIPT

Page 1: Baker (pp. 274-307)

PREGRADO UESAN

Asignatura: Gestión de Proyectos de Lanzamiento de

Productos (Inglés)

Lectura: Baker, M. (2007). En: Product strategy and management. (pp. 274-307). 2a ed. Harlow, England: Pearson Educación.

Profesor Responsable: Lu Chang-Say, Estuardo.

El presente material se pone a disposición de manera gratuita, para uso

exclusivo de los alumnos de pregrado de la Universidad ESAN y es sólo

para fines académicos, de acuerdo con lo dispuesto por la legislación

sobre los derechos de autor. Decreto Legislativo N° 822. En tal sentido,

se deja constancia, que la difusión de este documento bibliográfico, está

expresamente prohibida, por estar destinado únicamente para uso

académico en el presente curso.

Page 2: Baker (pp. 274-307)

10 Concept development and testing

Learning objectives

1. To position needs as the central issue in concept testing.

2. To explain the link between product characteristics and customer needs.

3. To describe the process of concept testing.

4. To outline the decisions managers need to make to design a concept test.

5. To discuss data collection methods for concept testing.

On completion of this chapter you will:

1. Understand more clearly the difference between screening ideas and testingconcepts.

2. Appreciate the underlying relationships between the benefits customers seekfrom products and the characteristics that make up products.

3. Be able to describe the process of concept testing for new products.

4. Understand the array of concept testing techniques that may be employed bycompanies.

5. Appreciate the breadth of data collection methods for concept testing.

6. Be able to design a concept test for a particular new product.

PROS_C10.qxd 3/14/07 7:09 PM Page 274

Page 3: Baker (pp. 274-307)

Chapter 10 • Concept development and testing 275

INTRODUCTION

In Chapters 8 and 9, the focus of discussion related to how companies generateand screen ideas for further development. At the next stage, which this chapterexamines, the screened ideas are developed into more fully specified concepts,which are evaluated for their appeal to the potential market segments for whomthey are intended. Concept development and testing involves designing and pre-senting representations of the proposed new product to a sample of its intendedcustomers. However, as has been discussed previously, these stages are by nomeans discrete, and much of what happens during concept development is con-ditioned by the nature of the idea generation and screening which has alreadytaken place. The objective of concept development and testing is to estimate mar-ket reaction to a product idea before committing substantial funds to its physicaldevelopment. In other words, this stage may be viewed as an extension of ideageneration and screening, in that those ideas screened to be of potential value aredeveloped and screened further in order to determine which idea specificationshave greatest appeal to potential customers, in the light of competitive offerings.The more detailed the information gathered at this stage, the greater the chanceof reducing total development costs, as alternative concepts can be derived andtested more cheaply than alternative, prototype, products. If a full commitment ismade to market research at this stage, the prototypes developed will, in theory,more closely match customer needs and preferences, requiring fewer costlyamendments later on.

The dual tasks of concept development and testing are intimately related. The first articulations of the new product concept will be derived from the activ-ities carried out during idea generation and screening, but these articulations can be reworked in the light of customer reaction, and retested, until an acceptableconcept is modelled, which can be progressed to physical development. This ‘in-stage’ iteration is shown in Figure 10.1.

The concept testing cycle highlighted is repeated until the company identifiesthe concept which exhibits the acceptable level of appeal, both internally andexternally.

The number of iterative cycles at this stage depends on the way in which theoriginal ideas were generated and screened. For example, a very large number of ideas screened may still not have been subjected to any formal evaluation by the market, in which case there is greater scope for an initial concept formula-tion to be unsatisfactory as a number of versions for the target market may haveto be drawn up. These initial formulations will have to be modified and retested.If, on the other hand, ideas are generated and screened on the basis of marketresearch techniques such as perceptual maps, the most appealing dimensionsmight well be present in the first concept formulations and fewer iterations areconsequently required.

As was discussed in Chapter 9, there is a lack of consensus regarding the delin-eation of screening and concept testing in texts on NPD. In some cases, discussion

PROS_C10.qxd 3/14/07 7:09 PM Page 275

Page 4: Baker (pp. 274-307)

276 Part II • New product development

of screening encompasses some of the concept testing techniques described inthis chapter, depending on how lengthy and costly screening is presumed to be.For our purposes, concept development and testing is taken to begin with avastly reduced number of ideas, that is after the initial evaluation or screeninghas taken place. The actual labels given to these stages are of less importance thanthe notion that as the NPD process unwinds, both the development of ideas andthe dimensions which can be evaluated are more specific, therefore yieldinggreater amounts of increasingly reliable information from the market. Any con-fusion that arises in terminology is largely due to the fact that the ‘stages’ of NPDare not tidy, linear and discrete.

A final point of introduction is to emphasize the parallel strands of develop-ment and evaluation that run throughout the NPD process. This is shown inFigure 10.2, building on the ideas introduced in Chapter 9 (see Figure 9.1).

As before, each development stage feeds into an evaluation stage. These evaluation stages have dual purposes: evaluation of the technical aspects of thenew product and its market appeal, based on market needs and preferences. Thisreinforces the ideas introduced in Chapter 9 (see Figure 9.2). As we have seenpreviously, market needs and preferences are germane to some methods used foridea generation and screening, which reinforces the idea that these early stagesare closely linked, in both objective and method, as they are focused on evolving

FIGURE 10.1 The iterations in concept development and testing

PROS_C10.qxd 3/14/07 7:09 PM Page 276

Page 5: Baker (pp. 274-307)

Chapter 10 • Concept development and testing 277

a concept attractive to the market, which can progress on to physical develop-ment and eventual market launch.

The remainder of this chapter is divided into three sections. The first looks atthe various purposes of concept testing, the second deals with the managerialdecisions to be made in developing concept tests, and the final section discussesthe specific types of tests that might be used.

THE PURPOSES OF CONCEPT TESTING

As explained above, concept testing has an overriding purpose: to estimate customer reaction to an idea before developing the physical product. However,

FIGURE 10.2 Parallel strands of development and evaluation

PROS_C10.qxd 3/14/07 7:09 PM Page 277

Page 6: Baker (pp. 274-307)

278 Part II • New product development

at this stage in the development cycle, a number of supporting objectives are implied:

• To profile the market:

– current buying pattern– existing segments– customers’ view of the products available

• To assess likely purchase intention and position the product:

– trial and repeat purchase– barriers to changing brands

• To make improvements to the new product concept:

– overall product concept– features of the product concept.

Each of these different sets of objectives implies that different concept tests areappropriate, although the various ways in which data are collected may oftenpermit several ‘tests’ or ‘measures’ to be made at once. The sets of objectives arelinked, largely through the concepts of customer needs and preferences. Theselinks are shown in Figure 10.3.

FIGURE 10.3 The purposes of concept testing

PROS_C10.qxd 3/14/07 7:09 PM Page 278

Page 7: Baker (pp. 274-307)

Chapter 10 • Concept development and testing 279

The centrality of needs

Figure 10.3 shows that while the three sets of objectives are themselves inter-related and interdependent, they are commonly informed by a thorough under-standing of customer needs and preferences. Therefore, these necessarily formthe core of all forms of concept testing. However, eliciting information regardingneeds is not always straightforward, particularly if developers relate them to a new technology. The ‘need’ for compact discs was not apparent until laser tech-nology was sufficiently developed to read digital imprints and amplify the musicat a level of quality previously unknown. The need for personal stereo systemswas not a recognizable need to the market before Sony designed, developed andlaunched the Walkman. There were, however, indications that the need forportable music existed, in the practice of carrying around radios and cassette-players. Moreover, the durability of this need beyond specific product character-istics is evident in the global adoption of MP3 and iPod music technologies. To sum up, needs have varying levels of visibility. King (1987) categorizes needsinto three types:

1. Basic needs – those which a customer will assume the product satisfies. Forexample, a consumer assumes that a vacuum cleaner will clean carpets.

2. Articulated needs – those which a consumer can express readily. These are oftenmet by at least one current solution, or can be easily imagined as being met.For example, a consumer might imagine a vacuum cleaner that shampoos acarpet.

3. Exciting needs – those that will delight and surprise a customer. These are usually not met by current products available and customers might find them difficult to articulate. For example, some customers might be excited by avacuum cleaner that distributed anti-stain solution and fibre conditioner!

Any new project concept must tap into all levels of needs. The fulfilment ofbasic needs is, of course, a prerequisite; the satisfaction of articulated needs char-acterizes the basis upon which the intended new product might compete withothers in the category; whilst addressing exciting needs provides a platform uponwhich to differentiate the new product and attract new customers. It should beremembered, however, that Baker’s maxim ‘the act of consumption changes theconsumer’ indicates the dynamism of the experience of needs. Once satisfied,needs migrate towards the ‘basic’ status.

In other words, a product concept which at the time of launch satisfied bothbasic and articulated needs will soon slip in status, to the part where it satisfiesonly basic needs. Examples of this can be found in passenger cars and personalcomputers alike. Concepts comprising features designed to reach articulated oreven exciting needs, such as airbags and ‘wysiwyg’ screens, soon became viewedas ‘standards’. How long before GPS in cars is standard, too?

PROS_C10.qxd 3/14/07 7:09 PM Page 279

Page 8: Baker (pp. 274-307)

280 Part II • New product development

Concept testing is not always appropriate in every situation. For example, con-cepts expressing new art and entertainments are difficult to test, as the success or failure of the final result is difficult to determine, or indeed to re-create. J.K. Rowling’s hugely successful Harry Potter and the Philosopher’s Stone wasturned down by numerous publishers with her agent taking well over a year tosign a deal with Bloomsbury. She was, of course, a first-time novelist, so the com-mercial impact of her writing would have been more difficult to assess. In thefilm world, attempts to account for the uncertainties of what drive success have mixed results. The Great Waldo Pepper had several tried and tested successingredients: proven director, Hollywood star in the form of Robert Redford, forexample. Commercially, it failed. A second category is where the concept is based on a new technology which the market does not understand and for whichthere are no reference points. Consumer acceptance of virtual-reality leisureproducts may fall into this category. A third category of concepts that is difficultto test reliably describes those expressing a new physical experience such as tasteor smell.

Despite these exceptions, concept testing is widely accepted to be an importantelement in successful NPD. If concept testing is deemed worthwhile and pro-cedures are aimed at all three sets of purposes, then the information collected, centred as it is on needs, will inform later stages in the NPD process, such as testmarketing and launch. This is an important issue, as it shows how, relativelyearly on, information regarding eventual segmentation and positioning is beingcollected.

The next section describes the relationship between needs and product conceptsthat must be understood if the purposes of concept testing are to be achieved.

The relationship between needs and products

In Chapter 1, we reviewed formal definitions of a product, most of which agreethat a product is the focus of an exchange process wherein customers exchangecash for satisfaction. In order to explicate the levels on which this satisfactionmay or may not be delivered, marketing literature uses a multi-level view ofproducts comprising:

• Core benefit: the fundamental service or benefit being purchased

• Generic product: the material properties which deliver the core benefit

• Augmented product: the ‘package’ which includes additional services andbenefits that distinguish the company’s offering from competitors

• Potential product: the augmentations and transformations that the productmight undergo in the future.

Viewed this way, a complete product concept would have to cover at least thefirst four of these levels. Even if each level were not to be individually subjected

PROS_C10.qxd 3/14/07 7:09 PM Page 280

Page 9: Baker (pp. 274-307)

Chapter 10 • Concept development and testing 281

to concept tests, the internal picture is not complete without an indication of howthe new product concept performs at this level. Recently, however, Saren andTzokas (1994) have questioned the utility of this view of the product in a waywhich is particularly relevant for our discussion of the link between user needsand product concepts. The basic criticism is that conceptualization of the productsuch as Kotler’s will attempt to define the product in isolation from the customer–supplier context:

It regards the product as an autonomous unit which can be de-constructed and understood in its basic elements. But this is fundamentally wrong since the mereunderstanding of the existence of a product for consumption is realised, not by its material nature (which sometimes need not be there at all), but by its symbolicmeaning that the society and individual consumers and producers have ascribed to itby means of culture, use or experience and their interaction with each other. (Sarenand Tzokas, 1994, p. 904)

Instead, the authors propose that a product is the outcome of a ‘continuous tri-partite signification process’ between buyers, suppliers and the object. Thisoutcome – the pluri-signified product – is shown in Figure 10.4.

The significance of this alternative view is that it highlights the need for product developers to focus not only on needs and expectations but also on theirunderlying determination and on the relationship between the customer, the

FIGURE 10.4 The pluri-signified product (source: Saren and Tzokas, 1994, p. 911)

PROS_C10.qxd 3/14/07 7:09 PM Page 281

Page 10: Baker (pp. 274-307)

282 Part II • New product development

object and the supplier. This, it is argued, is of benefit to those engaged in newproduct development, as it introduces elements into concept testing which areusually ignored (company reputation, whether or not consumers usually buy thecompany’s products) and over which the company has considerable control.

WHAT IS A NEW PRODUCT CONCEPT?

A concept which can be subjected to testing is more than an idea, newly emerged.An example of this might be where an idea, from a food company, is to enter the ‘health snack market’, which is growing and for which there are docu-mented consumer needs. This is still a vague idea, which must be turned into a more specific concept. There are several ways in which such a concept might be developed.

A health snack might take several forms: packet soups, vitality drink or a con-fectionery health bar, to name but three. These are still ideas. How could theirpotential be discussed – either internally or in the marketplace? First, there has tobe some specification of the ingredients, in order for the company itself to assessthe feasibility, likely competition and fit with strategy. Second, testing consumerneed for an acceptance of a ‘healthy snack soup’ does not give them much idea ofhow this product might benefit them or how it might be different from othersnack soups on the market. They might, however, indicate a ‘trial’ willingness,depending on their view and experience of the company concerned.

Concept tests are ultimately tests of perception, which are affected byinfluences such as past experience of a product or company, as well as the charac-teristics of any specific concept. Before going on to discuss the methods availablefor concept testing, however, it is important to define exactly what is being tested.As suggested above, it is difficult at this stage in the NPD process to test all thelevels of a product concept, largely because they have not usually all been devel-oped. For example, the ‘augmented’ product – the package of benefits includingretailed services – may not be developed until a later stage in the product devel-opment cycle. Further, although research has shown that new products withpotential longevity often have the seeds of the second generation within the con-ceptualization of the first, it is rarely a ‘testable’ factor. In other words, the con-cept is a promise of the benefits which will be derived from certain productcharacteristics or attributes. These attributes are produced from the manipulationof certain material properties, either of which might be transformed by new ormodified technology. These relationships may be described as the ‘mechanics’ ofthe new product concept, and are shown in Figure 10.5.

Why are these mechanics important, when the satisfaction of the need is theend result which is of interest to the customers? The answer is that rarely willcustomers be able to give an accurate assessment of their view of a new ‘promiseof benefits’ unless they can compare it with how those benefits will deliver satis-faction. This allows them to make comparisons with current offerings, where

PROS_C10.qxd 3/14/07 7:09 PM Page 282

Page 11: Baker (pp. 274-307)

Chapter 10 • Concept development and testing 283

they exist. For example, potential buyers of industrial heating and ventilation sys-tems cannot readily and reliably evaluate the promise of ‘greater comfort control,more cost savings’ without an idea of at least the attributes that will deliver thepromise, and more usually a sketch of the material properties of the system, theirfunctioning (manipulation) that will shape these attributes. In consumerdurables, the benefit of dry as well as clean dishes from a dishwasher can only beevaluated, with some idea of the attributes which will accomplish the benefits. In this case the nature of the attribute would need to be explained in terms ofmaterials, say a hot-air blower. Finally, a promise of benefits may be evaluated if the attributes are amplified in terms of manipulation, rather than materialproperties. For example, research has shown that in the UK ‘meat pie’ sector,women dislike jelly in cold pies (Grocer, 9 July 2005, p. 31). A process of pro-duction that reduced jelly content might be evaluated alongside the resultantjelly-free meat pie. A final example of a new concept is the ‘Guinness Surger’, afirst innovation since the 1989 launch of widget cans. The Surger is intended todeliver ‘a perfect pint’ in the home. It comes with a plug-in unit that sends anultra sonic sound through the beer, to replicate the way in which the pint settleswhen poured from draught in a bar. Although already launched (Marketing Week,February 2006, p. 5), it remains to be seen whether the promise of ‘the perfectpint’ is in the minds of consumers, realized by the Surger unit and whether thiswill help revitalize the declining sales of Guinness. What is of relevance here isconsideration of how the concept might have been tested before launch, to gaugeits potential for meeting its objectives.

These examples are driving towards the conclusion that a concept must be presented in such a way that the benefit, and one of the components of the poten-tial product’s attributes, must be apparent to the potential customer. Consider,for example, trying to have customers evaluate a ‘new’ concept four-hour trainjourney from Edinburgh to the English side of the Channel Tunnel. While the

FIGURE 10.5 The mechanics of a new product concept

PROS_C10.qxd 3/14/07 7:09 PM Page 283

Page 12: Baker (pp. 274-307)

284 Part II • New product development

benefit is clear, it is an idea which cannot be tested because the market wouldimmediately ask questions about how this service might be accomplished. If theirreactions to the idea were positive, but we added that the journey included a flight from London-City to the Tunnel, their reactions might change. In sum, a new product concept is not complete until it is articulated in terms of both attributes and benefits. Consumers can then evaluate whether they feel the benefits fulfil the need and whether the attributes deliver the promisedbenefits. This concept, however, is but a starting point. In order for concept testing to begin, a number of decisions must be made. These are the subject of the next section.

THE PROCESS OF CONCEPT TESTING

As outlined above, concept testing is a test of perception, and as such isinfluenced by all that influences perception, such as:

• Information given about the new product concept. The degree to which the concept is explained and linked to other product levels (for example theaugmented product) will affect customer perceptions.

• Timing of the evaluation. The more time given to a customer to evaluate a concept influences the response. If only a short time is given, only first impressions will be forwarded.

• Context for the concept. The context for a product evaluation may be definedin many ways, vis-à-vis competition, place of consumption or use of the prod-uct, with reference to certain market segments and positioning.

• The nature of the product concept itself. This relates to the point where a product concept fits on the continuum between ‘emotive’ and ‘functional’. Ingeneral, the more emotive a product concept, the more information is requiredto explain the concept. Of course, as has already been stated, highly emotiveproducts such as perfume may not concept-test well or easily.

These overriding considerations influence the choices outlined below in deal-ing with the process of concept testing, which is generally made up of threephases: (1) definition of specific objectives; (2) types of concept presentation; and(3) method for gathering data.

Defining objectives

The overall purpose of a concept test is likely to vary from one concept toanother. Earlier, broad purposes of concept testing were introduced: to assesslikely purchase intentions, to make improvements to the new product concept,

PROS_C10.qxd 3/14/07 7:09 PM Page 284

Page 13: Baker (pp. 274-307)

Chapter 10 • Concept development and testing 285

and to profile the market. The relative importance of each purpose will influencethe type of concept test carried out. Each is examined briefly below.

Assessment of purchase intent (PI)

It might be argued that this is the primary purpose of concept testing, so thatthose concepts with poor potential may be weeded out. All the techniques usedare in some way related to key concepts in buyer behaviour generally and to those associated with the adoption of innovation. Both these subjects were covered in Chapter 3. The most common way to assess purchase intention is toprovide a description of the product and ask respondents participating in the testto check the appropriate box, saying whether they:

� definitely would buy� probably would buy� might or might not buy� probably would not buy� definitely would not buy

It is usual to consider the top two boxes’ responses in deciding the concept’spotential. The percentage of respondents checking these can then be compared toany existing category norms. Alternatively, the percentage may be subjected toadjustments based on different levels of consumer awareness, distribution levelsand other levels of promotional support. In either case, marketing managers musttake a view on how likely the percentages are to be a realistic prediction of a realproduct launch. Many use past experience to adjust the percentage up or down.Research by Taylor, Houlahan and Gabriel (1975) suggests that a concept shouldreceive 80–90% of the responses in the ‘top two boxes’ to encourage furtherdevelopment work. Further, studies by Morrison (1979) and Kalwari and Silk(1982) suggest that purchase intention does correlate with actual purchases.

A vital consideration, particularly relevant in assessing market potential for fast-moving consumer goods, is to take account of ‘repeat buying’. Clearly,high trial will not always be followed by levels of repeat buying which wouldrealize potential.

Improving the product concept

Inherent to concept testing is concept development. Since respondents are askedto evaluate a concept, there is ample opportunity to adjust the details of the con-cept in order to pinpoint which concept has greatest appeal.

This development work may be carried out with respect to the overall product,in which case an indication of consumers’ perceptions of a number of concept-specifics may be evaluated (Dolan, 1993):

• the concept’s uniqueness vis-à-vis other products available

• the concept’s believability

PROS_C10.qxd 3/14/07 7:09 PM Page 285

Page 14: Baker (pp. 274-307)

286 Part II • New product development

• the ability of concepts to solve a customer’s problem

• inherent interest in the product concepts

• value for money.

These questions may be termed diagnostic. Moore (1982) defined diagnosticquestions as those which shed light on reasons behind the intention-to-purchasequestions, implying that diagnostic questioning tends to accompany intention-to-purchase questions. However, as one of the benefits of concept testing overproduct testing is the flexibility with which concepts can be altered and retested,diagnostic questions have a role in their own right in attaining the concept-to-market requirements.

In addition to diagnostic questions which elicit feedback regarding the entireconcept, some may focus on specific attributes of the product concept. Again thismay be investigated in relation to intention-to-purchase in a specific way, or togauge preference among different attributes or the perceived importance of oneattribute over another.

Market profiling

The third major purpose (objective) of concept testing is to assess the character-istics of likely buyers and non-buyers. At the most basic level, this will involve thecollection of demographic information, but may extend to include informationsuch as typical buying criteria, psychographic profiles, product usage patterns,and buying processes used. This kind of information helps to interpret analysesof the intention-to-purchase date, thereby leading to refinement of the conceptunder development and preliminary information for targeting, positioning, pro-moting and launching the product.

It is important that the balance of these three possible objectives of concepttesting is specified at the outset. Depending on these objectives – what the testsare to achieve – their nature will change. Diagnostic objectives may require bothqualitative and quantitative tests to be carried out. The actual types of questions,whether focused on one concept or on more than one, whether the concept statements given to consumers include a ‘positioning’ signal, essentially flowfrom the objectives.

Initial choices

In addition to defining these broad objectives for the concept tests, the remit mustbe specified in relation to two important initial choices: points of comparison andtarget markets.

Points of comparison

Faced with a new product concept, a potential customer has to decide how toassess whether the attributes deliver the promised benefits. Those benefits must

PROS_C10.qxd 3/14/07 7:09 PM Page 286

Page 15: Baker (pp. 274-307)

Chapter 10 • Concept development and testing 287

be set in the context intended by the developer. If, for example, we wish to testthe concept of cheese-flavoured potato bites, customers must understand againstwhat they are evaluating. A number of possibilities arise. First, it could be that nopoints of comparison are sought, that it is the intrinsic notion (core idea) of cheeseplus potato in a bite size which needs to be evaluated. The danger with such anapproach is that different customers do use competitive perspectives to maketheir evaluation. Some may have cheese-flavoured crisps in mind, others have‘take-away’ finger foods such as chips or nachos, while still others may comparethe concept with the huge variety of frozen potato variations available as a conveni-ence meal complement. If the developer defines the cheese-flavoured potato biteas a new snack concept, then the points of comparison must encompass othersnack opportunities. Failure to specify these points of comparison means thatdevelopers cannot understand what the concept is accepted (or rejected) as. Thishas implications, not only for the decision to continue or kill the development,but also for how the product is eventually positioned. Moreover, there is theincreased influence, in the UK, of the Food Standards Agency and its concern forfood promoted to children to bear in mind when considering positioning andtesting of the concept.

Target markets

The envisaged target market affects the marketing context into which the new product is placed. For example, cheese-flavoured potato bites may have several target audiences: children, food providers for children – itself a huge category which might be subdivided further, to include parents (working andnon-working), nursery and infant schools, crèches – and snack-eating adults, whoagain might be subdivided into health-conscious and convenience-conscious.Each of these targets would require a slightly altered type of test, in search ofwhat is rated and the points of comparison. Another issue relating to the targetmarkets introduces the concept of ‘expert consumers’. Faced with the criticismthat ‘confronted with a radically new technology, customers do not understandwhat needs the technology could satisfy’ (Tauber, 1974, p. 22), some advocate theuse of ‘innovators’ as the target market for concept tests (Ortt and Schoormans,1993) or ‘expert’ consumers (Schoormans, Ortt and de Bont, 1995). This is paral-leled in the industrial market by the work of von Hippel (1988) and Foxall (1988)who advocate the involvement of lead-users (experts) in new product develop-ment. The selection of the target segment does not happen once only throughoutthe NPD process. As we have seen in Chapter 7, the strategic context for NPDsets some broad parameters in which to consider target markets, whilst both ideageneration and screening often involve research with a specific target market.Indeed, ‘gaps in the market’ necessarily define target (and non-target) segments.Further on in the NPD process, the stages of product testing and test marketingrequire further specification of target markets.

Once these broad issues have been delineated, the more specific but equallyimportant issues of deciding on how to present the concept and collect the dataare broached. The next section deals with the first of these.

PROS_C10.qxd 3/14/07 7:09 PM Page 287

Page 16: Baker (pp. 274-307)

288 Part II • New product development

Types of concept presentationOnce the basic research objectives have been set, the next step is to define whatand how to present to customers. There are commonly seven ways to present aconcept to the market for evaluation:

• verbal presentation• monochrome line drawing• colour line drawing• photograph• storyboard• mock-up• computer simulation.

A verbal presentation is the most basic type of representation. Verbal presenta-tion generally gives an incomplete picture of the product concept (which mayresult in unreliable findings). Consider the following statement describing a‘smart radiator valve control’:

The new radiator valve control consists of a small (4 cm × 4 cm × 5 cm) cuboid,which is easily fitted to the existing valve by a simple screwing action. It has anattractive push button surface which controls the amount of heat generated bythe radiator. The control is programmable by temperature, hour and day of theweek, and can shut itself off when draughts (for example from open windows)are detected.

A monochrome line drawing can be used on its own (see Figure 10.6) or withaccompanying text.

A colour line drawing can be used to portray more of the feel of the product (seeFigure 10.7).

A photograph of a mock-up is normally used at the concept development phase.Also, where a ‘new’ product is introduced from a foreign market, photographs(or even product packs) may be available to evaluate the concept.

A storyboard is frequently used to evaluate advertising concepts. It allows apresentation of the product in the context for which it is designed.

A mock-up of the pack, shape or general form can be used – but at the conceptstage this is not a ‘working’ mock-up. This type of representation gives a cus-tomer a better feel for the product he or she is being asked to evaluate.

A computer simulation – or graphic – is used to represent the product. This hasthe advantage of being able to be rotated, enlarged, with the overall designs andfeatures highlighted as the evaluator wishes.

Influences on the choice of presentationThere is no agreed ‘best’ way to represent a product for the evaluation of thebasic concept, although there is increasing research on the matter. As we gothrough the six ways described above, the amount of information increases and

PROS_C10.qxd 3/14/07 7:09 PM Page 288

Page 17: Baker (pp. 274-307)

Chapter 10 • Concept development and testing 289

so there is an effect on the answers produced. Dolan (1993) contends that a moreembellished format of presentation will elicit more realistic evaluation. Wherethere is more information needed to convey the concept of the new product toprospective consumers, the greater the value of using pictorial or mock-up repre-sentation, since more information can be understood visually than verbally. In

FIGURE 10.7 Positioning and mounting of the electronic radiator controller

FIGURE 10.6 Electronic radiator controller

PROS_C10.qxd 3/14/07 7:09 PM Page 289

Page 18: Baker (pp. 274-307)

290 Part II • New product development

addition, where a product is likely to be viewed from one angle only, a pictorial(or verbal) description will suffice. For example, a hob designed to be built-in isgenerally viewed and used from one position only (above) and might be readilyrepresented by a picture. On the other hand, a free-standing cooker has severalpoints from which it is viewed and used – the hob, the grill, the oven, and so on, andtherefore it may be more comprehensively represented by a mock-up. A numberof issues affect the choice of method of representation. These are dealt with below.

The product’s frame of reference and product field

The more familiar the concept, the easier it is to conceptualize what it is like.Product concepts are generally viewed in terms of what is already in the cus-tomer’s mind – past experience of related brands and products. Therefore eachproduct type is judged by a specific frame of reference which differs betweenproduct fields. Thus, the form of presentation must take the field frame of referenceinto account. Generally, the product field frame of reference depends on whether aproduct concept is familiar or unfamiliar. Where product concepts are unfamiliar,there tends to be a greater need for the amount of information given. An exampleof the effect of this can be seen with the hypothesized concept testing of a new fruitbeer. Table 10.1 shows what consumers might think of various representations ofthe product in terms of its perceived size, its use and for whom it might be suitable.

The table shows how the opinions might alter, depending on the type of pre-sentation. If a company launched the product on the strength of the information

TABLE 10.1 Concept test results for fruit beer

Presentation Consumer evaluation of

Format Size Use Target market

1. Verbal No clear indication Mid-day refreshment → 18–35 female

2. Black and white drawing Half pint tumbler Evening drink → 45+ femaleand verbal description

3. Colour drawing and Highball glass Evening drink → 25–35 femaleverbal description Mid-day refreshment 35–50 male/female

(summer)

4. Colour photograph Pint glass Beer substitute 35–50 male; outdoor oriented

5. Pack shot, showing Small bottle Picnic accompaniment → 30–50 male and female33cl bottle

PROS_C10.qxd 3/14/07 7:09 PM Page 290

Page 19: Baker (pp. 274-307)

Chapter 10 • Concept development and testing 291

received after the first or second product concept tests, the resultant sales mightbe very different from those achieved if the company carried on to the fourthproduct concept test. This example underlines the fact that concept tests are evaluating perception, and the potential role of the test itself in determining thereliability and/or validity of the data. It follows also that the actual product must,once developed, be in accordance with the concept presentation or a differentmarket reaction might be obtained at a later stage in the NPD process. In mostcases, however, only one or two presentation formats will be chosen owing tocost constraints, so great care must be taken with the eventual choice.

Functionality of the product

A second factor influencing the product field frame of reference is where theproduct lies on the emotive–functional continuum mentioned earlier. In this case,the more emotive the product, the greater amount of information is required bycustomers attempting to evaluate it. This is because emotive products are soldmore on image, connotation, personality and so on. It follows, therefore, that alow level of concept information will not adequately represent the ‘attributes plusbenefit’ of the product. Consider, as an extreme example, the texts that go alongwith emotive products such as aftershave and perfume. The following are takenfrom descriptions of two major scents.

• Amarige de Givenchy: ‘wonderfully feminine, lively and elegant, warm-hearted and generous, Amarige is an expression of intense happiness’.

• Jazz: ‘Jazz starts with vibrating aromatic notes, moving onto a core of floral elegance over a base of wood, resinous warmth’.

For more functional products, the more information is given, the greater theamount of variation in customer/market responses. This happens because thedescriptions may vary. For example, here are two statements about vinyl foldersfor loose-leaf student notepads:

• ‘The folder is hard wearing, with a metal clip mechanism, and is expandable,holding up to one ream of A4 paper.’

• ‘The folder comes in a variety of colours and laminated finishes, holds up toone ream of A4 paper and costs £1.50.’

The second description gives the person who is evaluating the product differ-ent information against which the evaluation can be made. This means theresponses to the second description contain different variables that evaluatorsmight choose to comment on. The decision regarding how much and whatdescription to provide depends on the dimensions for which the developerwould like to receive comment. The more variables, the greater the range ofpotential comments.

PROS_C10.qxd 3/14/07 7:09 PM Page 291

Page 20: Baker (pp. 274-307)

292 Part II • New product development

Purpose of the test

The amount of information given to evaluators in a concept test also depends on thepurposes of the test, which may favour the use of a ‘core idea’ statement or a ‘posi-tioning concept’ statement. The question of which of these approaches to adopt isdebated. In favour of positioning concept statements are the following arguments:

• The assessors (respondents) relate the concept to current products more easily,resulting in a more lifelike assessment.

• Concept assessments are generally better, i.e. more favourable.

In favour of core statements are the following counter-arguments:

• The purpose of a concept test is to assess reactions, not persuade respondents.

• There is a need to elicit unbiased responses regarding the product, so thereshould be as little promotional appeal as possible.

• There is research showing how different copywriters’ texts may affect conceptscore results (Haley and Gatty, 1971).

The evidence in favour of each approach suggests that neither is demonstrablybetter than the other. Moore (1982, p. 287) suggests that ‘the amount of positioningand sell is a function of how great the benefit is, how well it is understood, howsocially acceptable it is to admit a certain need and how emotional the need is. Asa general rule, concept writers should use the minimum amount of sell required.’

Traditionally, the literature on product testing suggests that both a persuasivetone and visual stimuli increase concept ratings (Tauber, 1972; Holbrook andMoore, 1981) but more recently, Lees and Wright (2004) found that if an earlyconcept screen is all that is required, then the addition of a line drawing to a concept statement makes very little difference to consumer evaluations.

Accuracy of information presented

The information should not be misleading. Accurate information is dependentupon the information presented to respondents in a concept test. This is particu-larly important with respect to size, colour and key features intended by thedeveloper to differentiate the product from those of competitors. For example,information about a car’s features must show, either by explanation or visually,whether or not the model includes an airbag, if this feature is an important aspectof the car’s differentiation or positioning.

Comprehensiveness of information presented

Equally, the concept test must present all the relevant details. Respondents mustbe able to evaluate the concept against all the features of the product, not just the

PROS_C10.qxd 3/14/07 7:09 PM Page 292

Page 21: Baker (pp. 274-307)

Chapter 10 • Concept development and testing 293

more attractive features. For example, the concept test for a jug kettle mustdescribe not only its capacity and overall shape and dimensions, but all features,including whether or not there is a water-level indicator, an automatic switch-offmechanism, whether it is cordless and the location of the on–off button as anyone of these might cause the reaction to the product to be negative.

Consistency of information presented

All the information has to be consistent. Any graphic representation should notconflict with verbal descriptions, or the validity and reliability of the evaluationswill be rendered suspect due to an element of confusion. At best respondentsmight articulate any confusion; at worst they will tend to use one representationupon which to base their evaluations. The researchers will then not be able toknow whether they are comparing reactions to the ‘same’ representation wheninterpreting the results of the concept test.

Time allocated to convey/understand information

The amount of time given to a respondent to assess the concept tends to increasethe amount of criticism they will generate. Judgement must therefore be exer-cised when considering the time element. If the amount of time given for the concept test is too short, then the overall reaction might be more positive than ina real purchase situation. A useful clue here is to consider the nature of purchasesin the relevant product category. Low-involvement purchases which demand little effort in the purchase situation will need correspondingly less time in a concept test than products typically requiring a complicated search process.

There is relatively little evidence to shed light on which modes of presentationare used most frequently. In a recent study of one type of concept test, conjointmeasurement, over 75% of studies used verbal descriptions (Wittink, Vriens and Burhenne, 1994). More recently, Crawford and di Benedetto (2000) have sug-gested that the most common form of concept presentation is a concept statementplus line drawing. Studies by Page and Rosenbaum (1992) report that theSunbeam Appliance Company has adopted a form of concept testing which com-bines freehand sketches and lists of three to five bullet points to describe the useof products, their features and benefits. This type of testing allows the companyto evaluate a larger number of product concepts than by using either mock-ups orfully functional models, as it is much less expensive. An example of the type ofconcept card used is given in Figure 10.8, depicting a ‘spin fryer’.

These procedures cost around £1,000 per concept tested, whereas testing con-sumer reactions even to mock-ups costs £3,000 for the mock-up alone. This said,the expenses of the company show that the concept tests are not sufficient, by them-selves, to predict final market performance, since the way in which the final productis developed may change aspects of its appeal. Concept testing is not, therefore, asubstitute for product testing, which is examined in Chapter 12, but is a way offurther assessing potential before expensive physical development takes place.

PROS_C10.qxd 3/14/07 7:09 PM Page 293

Page 22: Baker (pp. 274-307)

294 Part II • New product development

The Internet now offers much greater scope for visual depiction, text, 3-D viewand indeed animation in concept testing. Moreover, as we shall discuss later inthe chapter, the costs of carrying out concept testing in this way can be dramatic-ally reduced. Research by Dahan and Srinivasan (2000) used the example of bicycle pumps and tested the preference of consumers using several differentvisual depictions of the products and differing amount of animation. Theirresults showed that this type of representation can provide a close match to thetype of results which might be expected from physical product testing.

The next section focuses on the choice of data collection techniques.

RESEARCH METHODS FOR CONCEPT TESTING

The research methods used for concept testing are numerous. Among the centralissues are data collection methods, location of the research and question format.

Data collection methods

The methods for collecting the data may be qualitative or quantitative.Qualitative procedures such as in-depth interviewing of either individuals orgroups of consumers are generally recognized to be of benefit, particularly forgenerating and refining concept statements, or indeed the concepts themselves.As with qualitative research in general, its intensive focus allows respondents the

FIGURE 10.8 Concept test results for the spin fryer (source: Page and Rosenbaum, 1992)

PROS_C10.qxd 3/14/07 7:09 PM Page 294

Page 23: Baker (pp. 274-307)

Chapter 10 • Concept development and testing 295

time and freedom to express likes, needs, preferences and opinions, without forc-ing a reply, as in quantitative methods. This, in turn, yields high levels of validityand delivers greater understanding to the developers in the dimensions alongwhich eventual customers will judge their product. This is particularly relevant,since concepts are likely to be multi-attribute. The depth approach afforded byqualitative methods allows a fuller understanding of the relative appeal of theseattributes as well as their collective contribution to the ‘whole’ concept. However,qualitative techniques require skilled fieldworkers and researchers (as do quanti-tative techniques, as discussed below) to avoid the pitfalls of leading consumersor misinterpreting their meaning. In addition, their intensiveness usually eclipsesthe possibility of carrying out research with large samples, which is problematicwhere the purpose of the tests is to generate representative estimations of marketpreferences and possible market shares.

Quantitative techniques, on the other hand, which cover survey techniquessuch as structured ‘personal’ interviewing, mail and telephone surveys, are lessintensive and can therefore be conducted on a much larger, random scale, withthe resultant data being more easily generalizable. In reality, however, becauseconcept testing is about assessing attitudes, perceptions, intentions and so forth,it is valuable for researchers to be able to explain issues and probe customers, sopersonal, structured interviewing is the most commonly used data collectionmethod. Mail surveys in particular have a number of drawbacks: it is difficult toobtain full answers to any open-ended question regarding product dimensions;response rates to mail surveys are generally low, casting doubt on the representa-tiveness of the responses obtained; respondents usually have no opportunity to ask questions about the concept; and, finally, the problem of the effects ofquestion wording and question order on the responses given cannot be easilyavoided. Telephone interviews suffer from limited opportunities to use anygraphic representation, unless combined with mail contact, and the interviewusually has to be kept rather short.

Research, however, by Jamieson and Bass (1989) showed that with multiple-wave telephoning, despite a reduction in numbers of respondents on each wave,the profile of respondents remained constant and the market share predictionlevels were satisfactory. There is still, however, little evidence that concept testingis carried out via telephone interviews elsewhere.

Location of tests

Given that the most common data collection method for concept testing tends tobe the personal, structured interview, decisions regarding the location of suchtests have to be made. The choice is usually made between carrying out the inter-views at the respondent’s home, or in a neutral location such as a shopping mall,a public thoroughfare, or a hotel seminar room.

Industrial companies may also execute concept tests at trade fairs or exhibitions,despite the drawback that competitors are present. Public places, such as the

PROS_C10.qxd 3/14/07 7:09 PM Page 295

Page 24: Baker (pp. 274-307)

296 Part II • New product development

street, shopping malls or hotel seminar rooms (where people from the street may beinvited), are usually varied to cover different times of the day as well as differentdays of the week, in order to generate samples that are as representative as possible.

Question format

In addition to decisions regarding the type of data collection strategy to be used,the types of question asked must be decided upon, in terms of their content andform. As discussed above, their precise content is in large measure dependentupon the objective of the concept test. Therefore, where the tests aim to identifythe proportion of likely users, the questions will be different from where tests are focusing on diagnostic issues. The types of question that relate to differentpurposes will be returned to shortly. There is, however, an overriding choice governing the type of question. Specifically, developers must choose betweenmonadic and comparison questions.

Monadic or comparison questions

Monadic questions are those which ask respondents about ‘one’ concept (hencemonadic). If the concept testing phase as a whole is focused on several new con-cepts, then each one would be evaluated by one sample only, with comparisonsbeing made after the data have been collected. No one respondent would beasked about more than one concept. Figure 10.9 shows how three new conceptsmight be tested using a monadic approach.

The three product concepts, A, B and C, are assessed by three groups ofrespondents. These three groups, or samples, are matched. This means they aresimilar samples in forms of demographic characteristics, or indeed of character-istics which might be used to segment the market for the product, once launched.The groups of respondents are asked to rate the (one) product they are testing in

FIGURE 10.9 The monadic approach in concept testing

PROS_C10.qxd 3/14/07 7:09 PM Page 296

Page 25: Baker (pp. 274-307)

Chapter 10 • Concept development and testing 297

terms of overall likes and dislikes, and in terms of likes and dislikes of specificattributes. These questions would satisfy the ‘diagnostic’ aims. For the purchaseintention aims, each sample would be asked how likely they would be to pur-chase the product. The results from each group are then compared and inferencesare drawn regarding which product concept has achieved highest acceptance.

The argument in favour of monadic testing posits that the test captures thereality of the purchase situation, since, faced with a choice of products, con-sumers compare products in their own minds. However, the disadvantage is that it is difficult to assess whether differences in scores for products are due tothe fact that different people are testing them, rather than due to a consensusview that one product concept is better or worse than another. Using the samepeople to test different product concepts resolves this problem, and is called comparison or comparative testing.

In comparison testing, the same group of respondents is asked about two ormore concepts. Several groups of respondents may be used in order to check thereliability of the results. Figure 10.10 depicts this approach.

One advantage of comparative testing is that since the same group of respond-ents is being used to test different new product concepts, the differences in aver-age scores for each can be more reliably allocated to real perceived differences,rather than differences among the groups of assessors. In addition, some peoplefavour comparative concept testing as they believe it reflects more accurately thepurchase situation where a potential buyer assesses a new concept in relation toones he or she already knows.

Question wording

Both the diagnostic and PI objectives can be pursued by both monadic and comparison questions. For purchase intention to be assessed monadically, thequestion wording might take the form of a checklist, as set out earlier under

FIGURE 10.10 The comparative approach in concept testing

PROS_C10.qxd 3/14/07 7:09 PM Page 297

Page 26: Baker (pp. 274-307)

298 Part II • New product development

‘Assessment of purchase intent’, or it might use a probability or point allocationquestion. A probability question for the spin fryer shown in Figure 10.8 might be:‘Please circle the number from 0 to 10 on the scale below which best indicateshow likely you are to buy the product. If you would definitely not buy the spinfryer, circle 0. If you would definitely buy it, circle 10. If your opinion is notdefinite, please indicate a number between 1 and 10 which indicates how likelyyou would be to buy, or not buy, the spin fryer.’

A point allocation system is suited to comparison tests. In this case, severalconcepts of the spin fryer might be presented to respondents, along with othercurrently available deep-fryers. Then respondents may be given points, counters,or even fake money. The question posed to respondents might be: ‘Based on thedrawings and statements you have just seen, allocate these 10 counters across thebrands in proportion to how likely you would be to purchase them.’

____________ ____________ ____________ ____________

Concept A Concept B Brand C Brand D

For diagnostic purposes, it will be recalled that a number of concept specificsmay be evaluated, including uniqueness and believability, the concept’s ability tosolve a problem, inherent interest in the product concept, and value for money.Questions which might be used to elicit diagnostic information on these dimen-sions for a Memo-flash-drive, which is named, hypothetically Me-Flash:

Concept statement: ‘The Memo-flash-drive is a functioning flash drive with a built-in digital voice recorder on to which up to 1 minute of spoken notes, messages and aide-mémoires can be recorded.’ It connects to other hardware toproduce a list which can be saved. (Concept A)

Ability to solve a problemHow relevant is the concept described above in solving your personal organiza-tion needs?

Highly Not at allrelevant relevant

BelievabilityHow believable is the claim that the Memo-flash-drive can record 1 minute of spokennotes which can be transferred to your hand-held, laptop or desktop or phone?

Completely Not at allbelievable believable

PROS_C10.qxd 3/14/07 7:09 PM Page 298

Page 27: Baker (pp. 274-307)

Chapter 10 • Concept development and testing 299

UniquenessCheck the box that most closely fits the description you just read:

An ordinary A uniqueflash-drive flash-drive

Inherent interestPlease write, in the space provided, how interested you are in the concept of aflash-drive which will record spoken memoranda.

________________________________________________________________________

________________________________________________________________________

________________________________________________________________________

Or: please check the box that best describes how interested you are in the flash-drive described above:

Extremely Not at allinterested – interested –it is a good it is a pooridea idea

Value for moneyIf the flash-drive were offered to you at £29.99, how would you describe its valuefor money?

Good value Poor valuefor money for money

These questions might also be modified to compare the Memo-flash-drive withother concepts such as the Flash-drive-alarm-clock. In this case, a comparativequestion would include a second concept statement:

Concept statement: ‘The Flash-drive-alarm-clock is a fully functioning flash-drivewith a built-in digital 24-hour clock and programmable alarm function.’ (Concept B)

The question regarding the comparative ability of the two concepts to solve aproblem might read as follows:

How relevant is each concept described above in showing your personal organiza-tion needs? Indicate their relevance by writing A and B in the appropriate boxes.

PROS_C10.qxd 3/14/07 7:09 PM Page 299

Page 28: Baker (pp. 274-307)

300 Part II • New product development

Highly Not at allrelevant relevant

Finally, market profiling requires that a number of questions be included toallow developers to learn which groups of respondents like and dislike the con-cept. The questions will usually cover demographics and psychographics, as wellas what products are usually bought and what the usual choice criteria are.

The above examples of the kind of questions used are intended only as broadsuggestions. In fact, several techniques are commonly applied in concept testing,which are essentially specific combinations of purpose, objectives, approach(monadic vs. comparison), data collection technique, question wording and analysis. Three of these are the subject of the next and final section of the chapter:conjoint analysis, hybrid method and the ‘House of Quality’.

SPECIFIC TECHNIQUES FOR CONCEPT DEVELOPMENT AND TESTING

This chapter has concentrated thus far on describing the purposes and decisionsof concept testing in general. As explained above, some techniques described in‘idea generation’ such as perceptual maps may have their uses also in designingand refining new product concepts.

Below we briefly discuss three more specific techniques associated withdesigning new products: conjoint analysis, hybrid conjoint analysis and the‘House of Quality’. As with other techniques described in previous chapters,there is no rule confining these only to the ‘concept development and testingstage’. They may be usefully employed at earlier stages in developing and screen-ing new product ideas, and adaptations may be of use in product testing. Thislargely depends on the type of new product process typically followed by a par-ticular company and the type (newness) of the new product being developed. It isperfectly conceivable, for example, that the three techniques described belowcould be applied in product testing, where the ‘new’ product is a modification of,or an extension to, an existing product, since the physical presentation of differ-ent attributes would not be too costly. On the other hand, where innovations aremore radical, the presentation of alternative designs and attributes is best kept tothe ‘concept’ stage if costs are to be kept low.

Conjoint analysis

The key purpose of conjoint analysis is to evaluate how potential customers judge products in term of their attributes and to examine the extent to which

PROS_C10.qxd 3/14/07 7:09 PM Page 300

Page 29: Baker (pp. 274-307)

Chapter 10 • Concept development and testing 301

they might ‘trade’ one ‘attribute’ for another. The term covers a number of datacollection and analytical techniques, which allow researchers to optimize thecombination of product attributes in terms of the overall evaluations that poten-tial customers make.

In order to explain how it works, we will use a much simplified example of a small manufacturer of decorative containers for food and beverages. One oftheir ranges uses wood veneers, which are fashioned into lightweight drums of a variety of sizes. Recently, the company wished to enter the market for specialitycheeses, a product category where packaging is an important indicator of quality.In order to design their product for this market, the company had two funda-mental design decisions to make. The first related to the nature of the identifica-tion of the contents: labels or screen-printing. Screen-printing is more expensive.The second design decision related to the nature of the finish given to the wood, where three possibilities exist: natural, stained or lacquered. Thus, thereare 3 × 2 = 6 combinations. In order to illustrate how conjoint analysis might provide insights into how these attributes might be combined, we can take thehypothetical individual rankings on each of the six combinations:

Product identification

Container finish Labelled Screen-printed

Natural 3 2Stained 5 1Lacquered 6 4

The preferred combination, therefore, from this individual is the screen-printed, stained container, while the least favoured combination is the lacquered,labelled container. These rankings are then assigned a utility value which, it isassumed, describes the perceived utility of the specific combinations, and fromwhich the perceived utilities of each attribute can be deduced. Thus, the rank-ings of the overall product variations (in this case, six) allow the researchers todisaggregate the scores to derive the perceived utility of each attribute. (Theseutilities are also known as part worths.)

In assigning utilities, we can code the rankings so that the least attractive com-bination gets 0, and the most highly ranked gets 5:

Product identification

Container finish Labelled Screen-printed

Natural 3 4 Average = 3.5Stained 1 5 Average = 3Lacquered 0 2 Average = 1

Average = 1.3 Average = 3.7

PROS_C10.qxd 3/14/07 7:09 PM Page 301

Page 30: Baker (pp. 274-307)

302 Part II • New product development

Since both product identification attributes are rated with each of the finishattributes, it is possible to use the average score of each attribute across the ratings as an indication of its ‘utility’ or ‘part worth’. In this case, the individual’sutility value for each attribute is as follows:

Identification: Label 1.3, Screen-printing 3.7Container finish: Natural = 3.5, Stained = 3, Lacquered = 1

From this we can see that lacquer is the least favoured finish and that the pre-ferred product identification method is screen-printing. However, were this indi-vidual typical of a market or market segment, the company might wish to knowto what extent a lacquered finish might be acceptable along with a screen-printedproduct identification. Using the average utility scores we can estimate this. The decrease in utility value from having a lacquered finish is at least 3 − 1 = 2.The increase in utility value from having screen-printed product identification is3.7 − 1.3 = 2.4, which is larger than the decrease in utility value perceived by thelacquered finish. Therefore, a trade-off might be possible.

The simplicity of this exercise is misleading. In reality, trade-off or conjointanalysis programmes are capable of handling far more levels of far moreattributes of concepts. This can be done in one of two ways. The full-profilemethod may be used, where potential customers respond to complete descrip-tions (or full profiles) of the product. However, the term full-profile is rather misleading, since it is rarely possible to ask potential customers to rate all thepossible combinations of attributes. Even if only four attributes were beingaddressed, where one attribute consists of two levels and the other three consistof three levels each, the number of combinations would be 2 × 3 × 3 × 3 = 54. If itwere even possible to have respondents evaluate all 54, it would be tedious in theextreme for them. It is more usual, therefore to use a subset of these concepts.This is called a fractional factorial design. It is still a ‘full-profile method’, becauseit is possible to contain all of the attributes in the study within the description.

The second method is called the paired comparison method, where respon-dents choose between two concepts, giving the one they prefer. Again, fractionalfactorial designs can be used, to avoid having to ask respondents to judge everypossible combination of pairs. Whilst a detailed discussion of analytical methodsfor these techniques is beyond the scope of this book, a brief résumé of the deci-sion steps in conjoint analysis is appropriate. Dolan (1993) presents a useful tableof the decision stages in conjoint analysis, shown in Table 10.2.

For stages 1 and 2, the factors influencing the decision have already been discussed for concept testing in general. With respect to conjoint analysis in particular, it is important that all the salient attributes – those that might have animpact on the eventual decision to buy the product – are included. For stage 3,consideration must be given to the extent to which respondents can provide morereliable judgements by ranking preferences on attributes of alternative concepts,or by rating the alternatives independently in respect of the attributes. Thisrelates back to the discussion regarding monadic and comparative tests and thedegree to which a rating, on, say, ‘liking’ of an attribute can be made accurately

PROS_C10.qxd 3/14/07 7:09 PM Page 302

Page 31: Baker (pp. 274-307)

Chapter 10 • Concept development and testing 303

without specific reference points being set by the researcher. Stage 4, deciding onthe criterion, is related to the overall ‘purpose’ of the study, as discussed above,but carries the additional decision as to whether the criterion to be measured is‘preference’ or ‘purchase intention’ (PI). The distinction is not a trivial one; whilea respondent may prefer the aesthetic appeal of a Bang and Olufsen television, hemay at the same time have no intention to buy one because of its price. Stage 5, choiceof analytical method, is related to the chosen input data, as described above, andthe purpose for which the data are intended. At the stage of concept evaluation, fourpurposes are implied for conjoint analysis: analysis of attributes, competitiveanalysis, market share prediction and market segmentation. It is not the intentionof this section to describe in detail the analytical methods, which are discussedfurther in Dolan (1993), Urban and Hauser (1993) and Moore and Pessemier (1993).

The extent of usage of conjoint analysis has not been subject to investigation inEurope. However, Wittink and Cattin (1991) surveyed the use of the technique inover 1,000 projects during a five-year period from 1981 to 1985. Their findingsshowed that a majority of users were in consumer goods manufacturing. Themost common purposes for which the technique was used included conceptidentification, competitive analysis, pricing, market segmentation and reposition-ing, and multiple purposes were used for studies.

The most common form of data collection was the personal interview, and themost common form of construction was ‘full profile’. Both rating and rankingresponse scales were most typically used.

The decisions involved in conjoint studies affect the accuracy of results. One of the common criticisms of conjoint analysis in concept testing is that data mightbe somewhat artificial as they are typically gathered in isolation from existingproducts. This limits its use in market planning because it is unrelated to marketstructure (Acito and Hustad, 1981). On the other hand, an article by Katz (2004),drawing on industry experience has suggested that once a concept has been

TABLE 10.2 Decision stages in conjoint analysis

Stage 1: Determine relevant attributes

↓Stage 2: Choose stimulus representations (i.e. form of representation)

↓Stage 3: Choose response type (i.e. rankings or ratings)

↓Stage 4: Choose criterion (i.e. liking, preference, likelihood of purchase)

↓Stage 5: Choose data analysis method

Source: Dolan, Robert, Managing New Product Development Process, 1st Edition, © 1993 p. 116.Reprinted by permission of Pearson Education, Inc., Upper Saddle River, NJ.

PROS_C10.qxd 3/14/07 7:09 PM Page 303

Page 32: Baker (pp. 274-307)

304 Part II • New product development

defined, with all the features that are believed to address a customer’s need, it is useful to assess how they will be traded off against competing features by the target market, and at what price. Katz further suggests that it can be usedeffectively as a forecasting tool in this regard.

In addition to the traditional type of conjoint analysis, there are several pro-ponents of a modified procedure called ‘hybrid methods’ that have appeared(Green, 1984; Acito and Hustad, 1981). This is briefly described below.

Hybrid conjoint analysis

The hybrid procedure is intended to provide market share estimates for new andexisting products, examining the impact of changes in product attributes, priceand competitive reaction and taking into account ‘subjective’ factors such as manufacturer’s reputation, reliability in delivery, and service levels given to cus-tomers. In addition, it is designed to anticipate the extent of brand substitution,also called cannibalization, since overall company revenues will not be enhancedif a new product’s sales merely take the place of those provided by existing prod-ucts. In order to deliver these insights, the procedure must not only specify thenew product concept and select the experimental design, it must also define theset of relevant competing products. This involves, in the case of an entirely newproduct, finding which alternatives currently satisfy the needs the new productaims to satisfy.

In order to describe this in more detail, consider the example of two majormanufacturers of industrial air cleaners. These products are usually ceiling-mounted. They draw in air, which is passed through a filter to remove dirt anddust particles, and the air is expelled once ‘clean’. These products, when not usedas part of an integrated air-conditioning system, are most suited for the ‘Horeca’market (hotels, restaurants, cafés) across Europe. They tend not to be suited tooffices or domestic use because of the rather noisy motor used to suck in andexpel air. The manufacturers, whom we will call Atlas and Gaia (not their realnames), each manufactures a ‘standard’ product, which has one power level (forair suction) and which costs around £3,000 to buy. Recently, Atlas introduced areduced-noise version, which has reduced Gaia’s market share. The reduced-noise version costs £3,500. Gaia, therefore, has been considering how to developproducts which will strengthen their competitive position regarding this. Designengineers have found an alternative solution, which would increase the powerlevels, allowing greater volumes of air to be cleaned per hour, making the cleanermore effective. They have also identified ways of making the higher-powercleaner less noisy than the current standard cleaner they offer to the market.There are different cost implications for both options. Gaia must look at theimpact of the new product on its other standard product as well as the potentialimpact on the competitor’s products.

The procedure to examine these relationships begins with a description of all the potential and existing products. To begin with, there are two existing

PROS_C10.qxd 3/14/07 7:09 PM Page 304

Page 33: Baker (pp. 274-307)

Chapter 10 • Concept development and testing 305

‘standard’ products selling at £3,000. Second, there is Atlas’ new product whichexhibits the same power levels as the standard products, at a reduced level ofnoise. Its price is £3,500. The new concepts proposed by Gaia have either (1)increased power and reduced noise levels, costing £4,000, or (2) standard powerand reduced noise levels, costing £3,500. Then respondents would be asked torate their purchase intentions among existing products, as well as their prefer-ence based on different combinations of attributes of the concept. This allows thecalculation of preference or PI weighting for attributes, which can be used to cal-culate potential market shares. These can then be compared with the PI on exist-ing procedures to examine the effect of the new concept on Gaia’s aim, andcompeting existing products.

This procedure entails different results than a conventional conjoint analysis.Specifically, its focus is not exclusively the trade-offs consumers or buyers willmake between one attribute and another. It extends to consideration of theimpact of the new product on other sales in the market. In addition, the resultsare not centred on an ‘individual’s’ value system, but on an aggregate ‘intentionto buy’ or preference indicators, which may be measured on a number of differ-ent scales, including ranking, rating and constant sum techniques.

As with standard, conjoint analysis, there is a limitation in respect of the numberof features that may be tested. In addition, the customer assesses the productgiven the representations developed by the company. It follows, therefore, thathighly abstract representations may lead to results which should be interpretedwith caution.

The House of Quality (Quality Function Deployment, QFD)

The final technique considered is part of Quality Function Deployment (QFD)introduced in Chapter 6. The philosophy embedded in the ‘House of Quality’,which is the first series of steps in QFD, is the identification, structuring and prioritization of customer needs, allowing the early development of concepts to beneed-focused, rather than attribute-focused, in the first instance. In Chapter 6, webriefly described how the ‘House of Quality’ is useful in identifying needs. It isimportant to realize that QFD is a holistic technique which does not benefit in practice from being ‘split up’. However, its concurrent parts do fit well withvarious steps of the NPD process as it is commonly represented, and therefore itis the contribution on prioritizing needs that is related to concept developmentand testing. Furthermore, Katz (2004) affirms its usefulness as a concept defining-testing methodology which can be usefully undertaken to identify and prioritizewhich design specifications are most deserving of investment. Prioritizing needshelps the developer to balance the cost of fulfilling the need with that of benefit tothe customer. Griffin and Hauser (1993) describe the process of prioritizing needsas one of identifying the ‘importance’ of primary, secondary and tertiary needs asdescribed in Chapter 6.

PROS_C10.qxd 3/14/07 7:09 PM Page 305

Page 34: Baker (pp. 274-307)

306 Part II • New product development

The first ‘House of Quality’ covers a number of market-based techniqueswhich are traditionally spread across the early stages of the NPD process. Thedefinition of importances is a prerequisite for design work, but ultimately it is thedesign team which identifies the product attributes which will fulfil those needs.The techniques for measuring importances are described below.

Needs can be prioritized using a number of methods, including direct ratingscales, constant show scales and anchored scales. In Griffin and Hauser’s (1993)comparison, each method yielded similar rank-ordering (importance) of needs.On the other hand, the technique of using ‘frequency of mention’ did not appearto be a good surrogate for importance. They also found that where the import-ance weights of attributes are derived statistically (called revealed techniques),there was little association of these with product preference or interest in the concept. In addition, they found that revealed estimates of importance did nothave high face validity. In other words, revealed estimates did not make sense.

Pullman, Moore and Wardell (2002) carried out research to compare the effects of using conjoint analysis and QFD. Their point of departure is thatwhereas engineers, technical development and quality experts within firms havetended to use QFD, the use of conjoint analysis is more used by marketers to design new feature sets. Focusing on a new all-purpose climbing harness, they found that conjoint analysis made it easier to compare the most preferredfeatures while QFD better facilitated the identification of features that had bothpositive and negative features. In sum, then, both QFD and conjoint analysis havetheir place in concept development, testing and refinement, but perhaps withslightly different foci.

SUMMARY

This chapter has introduced a number of issues concerned with the developmentand testing of new product concepts. It began with an explanation of where the stage‘fits’ in the NPD process, and where overlaps and reiterations exist in relation toprevious stages. Next, the various purposes of concept testing were explained,with emphasis on the centrality of customer needs and preferences to these purposes. The nature of a new product concept was defined and the process anddecisions in designing a concept test were described. Finally, three more specifictechniques associated with concept development and testing were introduced.

After the information is collected and analysed, a major decision whether tocontinue with the development must be made, before making a commitment to full physical development of the product. The physical development of theproduct is a costly business, and, as mentioned previously in this book, the more‘up-front homework’ that can be done the more efficient and effective the entiredevelopment will be.

The next chapter, therefore, details the kinds of financial decisions that must be made.

PROS_C10.qxd 3/14/07 7:09 PM Page 306

Page 35: Baker (pp. 274-307)

Chapter 10 • Concept development and testing 307

QUESTIONS

1. Describe the differences between testing concepts and screening ideas.

2. Examine the differences between product characteristics and customer needs.

3. Explain how the components of a product might or might not satisfycustomer needs.

4. Why are needs important in concept testing?

5. What are the different objectives that concept testing might meet?

6. Describe the factors that affect the choice of product presentation in a concept test.

7. Compare the data collection methods for concept testing.

8. What are the advantages and disadvantages of monadic and comparativetesting?

9. Describe the concept of part-worths in conjoint analysis.

10. Do you consider QFD to be an improvement on other concept testingmethods? Why or why not?

PROS_C10.qxd 3/14/07 7:09 PM Page 307