simulation

Rupinder P. Jindal, Kumar R. Sarangee, Raj Echambadi, & Sangwon Lee

Designed to Succeed: Dimensions of Product Design and Their Impact on

Market Share This research examines the relationship between product design and market share, a topic of considerable significance that has not been addressed in the published literature. Drawing from diverse disciplines such as marketing, industrial design, and engineering, the authors conceptualize design as being composed of three distinct product-level dimensions: function, form, and ergonomics. Furthermore, the authors examine the interplay among these design dimensions and their impact on the market share of a product. Empirical results using integrated repeated cross-sectional data obtained from several different sources in the U.S. light vehicle industry reveal an important strategic trade-off concerning design capabilities. Firms can either “design for satisfaction,” by investing in both function and ergonomics, or “design for delight,” by investing in form design capabilities so as to reap share rewards. The authors also show that older-generation vehicles with superior form designs do much better in terms of share than corresponding older vehicles with higher levels of either function or ergonomics. Implications of these results for academic researchers and managers are discussed.

Keywords: product design, ergonomics, new products, mixed models, auto industry

Online Supplement: http://dx.doi.org/10.1509/jm.15.0036

P roduct design has been defined as both the art and the science of putting the technological, financial, opera- tional, and emotional pieces together in order to create

a differentiated product (Lojacono and Zaccai 2004). Due to intense competitive rivalries and diminishing importance of traditional sources of competitive advantage (Srinivasan et al. 2012), firms are now moving beyond employing design as merely a stage in the product development process to investing in design as a strategic tool (see Verganti [2006] for examples from firms such as Alessi, Apple, and Kartell). The strategic importance of design for practitioners is reflected in the growing number of design-oriented conferences and complete issues of popular magazines being dedicated to design (e.g., Bloomberg Businessweek [since 2013], Fast Company [since 2006], and Wired [since 2012]). Evidence has suggested that the emphasis

on design as a strategic capability is indeed a sound strategy; superior design has been acknowledged to be a critical determinant of performance (Homburg, Schwemmle, and Kuehnl 2015; Martin 2009).

In parallel, product design has been the focus of intense scholarly scrutiny across various fields, underscoring the cross-disciplinary theoretical appeal of the topic (see Bloch 1995; Rindova and Petkova 2007; Tosi 2012). But these distinctly different fields have taken disparate approaches, depending on their world views. For example, engineering has typically focused on the objective functional dimension of design, that is, technical features, architecture, and spec- ifications. Research in social sciences, on the other hand, has highlighted the role of a more subjective visual aesthetic dimension, that is, form design, that appeals to the senses (see Bloch 1995; Rubera 2015). In recent years, with user experience (UX) and user-centered design (UCD) gathering momentum, a distinct stream of literature in industrial design has focused on the role of ergonomic design (see De Albuquerque Campos, De Lima Lanutti, and Paschoarelli 2012). We integrate these disparate strands of literature in order to paint a holistic picture and posit that design is multidimensional, with function, form, and ergonomics as its three constituent dimensions.

This tripartite conceptualization of product design pre- sents a fascinating challenge for firms: how should these dimensions be configured to create a winning product? This research challenge, which has significant managerial implications in terms of development of design capabilities and resource allocation strategies for firms, has not been

Rupinder P. Jindal is Assistant Professor of Marketing, Milgard School of Business, University of Washington Tacoma (e-mail: [email protected]). Kumar R. Sarangee is Assistant Professor of Marketing, Leavey School of Business, Santa Clara University (e-mail: [email protected]). Raj Echambadi is Professor of Business Administration and James F. Towey Faculty Fellow, College of Business, University of Illinois at Urbana- Champaign (e-mail: [email protected]). Sangwon Lee is Assistant Professor of Marketing, Miller College of Business, Ball State University (e-mail: [email protected]). The authors thank Edward A. Blair and Pavan Rao Chennamaneni for their insightful advice and suggestions on the manuscript and Mary Kay Dailey for assistance with graphic design. The first author dedicates this article to his PhD mentor, the late Professor Erin Anderson. The first three authors contributed equally to the manuscript. Robert Palmatier served as area editor for this article.

© 2016, American Marketing Association Journal of Marketing ISSN: 0022-2429 (print) Vol. 80 (July 2016), 72–89

1547-7185 (electronic) DOI: 10.1509/jm.15.003672

addressed in the extant literature. We address this gap by studying the effects of interplay among the three design dimensions on market share, an important indicator of commercial success in the marketplace. The use of market share as an outcome variable is novel in the design literature because most scholarly work has primarily focused on indi- vidual consumer-level measures, such as consumer evalua- tions and satisfaction (see Chitturi, Raghunathan, and Mahajan 2007; Srinivasan et al. 2012).

The context of our study is the light vehicle industry in the United States, wherein design plays a major role in product success (Cho, Hasija, and Sosa 2015). Using repeated cross- sectional data integrated from multiple secondary sources, and controlling for price effects, we find the following: First, and not surprisingly, we find positive effects for all three dimensions of design: better form, function, and ergonomics yield higher market share, although the simple effect of form is not statistically significant. Second, we find interesting in- teractions among design dimensions. The interaction effect for form and function is negative, such that the effect of improved form design on market share diminishes when functional design is high (e.g., above average) versus when it is low (e.g., below average). The interaction effect for form and ergonomics is similarly negative, whereas the interaction effect for function and ergonomics is positive. Discussions with designers and auto industry experts suggest that these interactions arise from demand-side heterogeneity; in other words, different customers value the various design dimen- sions differently because of their inherent preferences. Some customers place high valueonform design but are less sensitive to functional and ergonomic design, whereas others place high value on function and ergonomics but are less sensitive to form.

Our study thus reveals an important strategic design capability trade-off for firms. The manufacturer of a vehicle with superior form design may receive a market share pen- alty from taking functional or ergonomic elements to the highest level as well; such a product could be considered “overdesigned.” Conversely, the manufacturer of a vehicle with superior functional design is likely to benefit from superior ergonomic design, and vice versa, but the manu- facturer will likely not realize market share benefits from adding superior form design. Given the size of this industry, the monetary implications of these interactions is substantial; for example, shifting market share up or down by a tenth of a point in the overall industry represents more than half a billion dollars.

Beyond the interplay among dimensions, we examine another important question. Given that an older vehicle in any time period usually has lower market share than it did when it was new, introducing new product generations for sustained differentiation in the marketplace is a critical, albeit expen- sive, strategy in the automotive industry. Can design help older products hold their share, and if so, is there a differen- tial impact among the subdimensions? Our results show that older vehicles with higher form design hold their share bet- ter than older vehicles with higher functional or ergonomic designs. The implication is clear: firms that focus on func- tionality and ergonomics need to introduce new-generation

products more frequently than firms focused on form, which can afford to wait longer.

The remainder of this article is organized as follows: First, we identify the three dimensions of design and discuss each briefly. Next, we describe the contingent effects and the methods used to address them. Third, we discuss the results. Finally, we note the limitations of our research, discuss the implications, and offer suggestions for future research.

Theory A Tripartite Conceptualization of Design

Sales and market share are aggregate-level outcomes that are indicative of value captured by a product. Consumer per- ceptions of value creation through the product, that is, how well a product’s utility aligns with a consumer’s own needs and expectations (Priem 2007), drives this value capture. Because product design encompasses translating technical solutions into products with features that enhance value for consumers (Veryzer 1995), it is one such firm capability that can affect firm-level outcomes. From a firm’s perspective, design is about integrating inputs from engineers, product developers, and industrial designers (Hagtvedt and Patrick 2014) and leveraging their specific design capabilities in order to generate appropriate value-creating products.

Although conventional wisdom may point to the ap- plicability of design only for high-priced premium prod- ucts, Landwehr, Labroo, and Herrmann (2011) show that design affects sales across all segments in the auto in- dustry. More importantly, their work shows that design aspects are independent of retail price and advertising and explain up to 19% more variance in sales. So what do these design capabilities specifically entail? A survey of the design literature across disciplines identifies three different dimen- sions (e.g., Bloch 2011; De Albuquerque Campos et al. 2012; Tosi 2012). Design pertains to visual attributes related to form (Bloch 1995), performance attributes related to functionality (Veryzer 1995), and user experience and comfort attributes related to ergonomics (Moon, Park, and Kim 2015). Although the multidimensionality of design has been acknowledged in marketing (Srinivasan et al. 2012), roles played by form and function have attracted the most attention (Chitturi, Raghunathan, and Mahajan 2007); ergonomics has re- ceived relatively limited attention.

In their three-dimensional specification that included function, form, and meaning, Srinivasan et al. (2012) sub- sumed ergonomics under both form (e.g., ability to rest arms comfortably while driving), and function (e.g., seat belt comfort). Their meaning dimension refers to significance and memorial associations about the product that are shared by the consumers. In a similar vein, Homburg, Schwemmle, and Kuehnl (2015) conceptualize a three-dimensional view of design that comprises functionality, form, and symbolic dimensions, wherein they subsume ergonomics explicitly under the functionality dimension. Their sym- bolic dimension refers to the self-image communicated to consumers (p. 44). Both meaning and symbolism relate to non-product-related needs of personal expression and occur

Designed to Succeed / 73

as a consequence of the consumption of a product (Keller 1993). Given that product-related design attributes possess the potential to change the meaning or symbolism of the product (Rubera 2015), we refrain from using meaning or symbolism as dimensions of product-related design.

We contend that form and performance alone are not useful if crucial aspects of the product related to user experience, that is, ergonomics, are not met. Because ergo- nomics has been demonstrated to be a distinct construct completely different from either form or function (Moon, Park, and Kim 2015), we conceptualize product design as composed of form, function, and ergonomics; furthermore, we examine how these constituent dimensions interact with one another and influence market share. Table 1 provides a broad overview of notable work done on design and places the current study in context. Aligning these design dimensions into appropriate product configurations that match consumer needs and expectations affects consumer-level evaluations and, thereby, market-level outcomes.

Functional, Form, and Ergonomic Design

Functional design refers to performance specifications that embody the utilitarian aspects of a product (Townsend, Montoya, and Calantone 2011). It delivers utilitarian value to customers by bundling together various attributes in order to provide adequate technical performance (Bloch 1995). In particular, functional design strives to accomplish prevention goals characterized by desires such as “being responsible” (Chitturi, Raghunathan, and Mahajan 2007, p. 703) and positively influences customer satisfaction (Srinivasan et al. 2012). Examples of functional design include the perform- ance of a vehicle in terms of ride smoothness, handling and stability, and braking performance (Norman 2005).

Form design blends chosen aesthetic characteristics of a product into a visual whole to appeal to, and please, human senses (Bloch 1995). For example, the form design of a Harley-Davidson Sportster encompasses the prominent V-configuration of its engine, the raked angle of its front shocks, the teardrop shape of its gas tank, and the way in which these elements work together as a visual whole (Bloch 1995). Form design typically strives to accomplish promo- tion goals that are usually characterized by desires such as “looking cool” (Chitturi, Raghunathan, and Mahajan 2007, p. 704). Form design creates hedonic value which, in turn, arouses consumer emotions and is thus capable of developing deeper relational outcomes (Noble and Kumar 2010). Indeed, prior research has established the positive influence of form on consumers through various routes such as intrinsic experiential value (Hirschmann 1983), visceral emotional reactions to the product (Rindova and Petkova 2007), and customer delight (Chitturi, Raghunathan, and Mahajan 2007). Thus, form is regarded as a significant driver of product differentiation (Cho, Hasija, and Sosa 2015) and competitive advantage (Veryzer and Hutchinson 1998).

Although function and form are critical and valid dimen- sions of design, the picture is incomplete without incorporating ergonomics. Ergonomic design involves incorporating user- centric factors in product design, including experiential aspects

of the product–user interface (Montini 2014), in order to maximize comfort for users during consumption (Bloch 1995). As such, ergonomic design emphasizes complementing the abilities of users while minimizing their limitations, rather than forcing them to adapt.1 Ergonomic design is about enhancing comfort and improving product usability—it “makes products fun and engaging and emphasizes beauty in interaction rather than appearance alone” (Bloch 2011, p. 379). For example, Herman Miller’s Aeron office chair is designed on the prin- ciples of ergonomics; it lowers spinal compression, promotes safety by reducing stress injuries, and consequently enhances seating comfort (Srinivasan et al. 2012).

The theory of internal processing algorithms suggests that consumer responses to design are determined by cog- nitive algorithms, that is, nonconscious, internalized pre- existing rules, such that individuals prefer designs that conform to those cognitive algorithms (Rubera 2015). By delivering positive user experiences and communicating operational effectiveness, ergonomic design ensures con- sistency with preexisting rules and generates positive eval- uations. Thus, ergonomic design serves both utilitarian and hedonic aspects (Noble and Kumar 2010). Examples of er- gonomic design in an automobile include quality of seats and various other comfort and convenience features (Kaljun and Dolsak 2011).

In summary, product design has three distinct dimensions— function, form, and ergonomics—each of which can positively influence product evaluations. Our theoretical development is based on the premise that form, function, and ergonomics are utilized by consumers as cues to evaluate products, form their preferences, and then make choices in accordance with those preferences that ultimately drive market share. We next examine the interrelationships among these dimensions to understand their substitutable and complementary effects on market share. Our conceptual framework is illustrated in Figure 1.

Interdependencies Among the Design Dimensions

Form and functional designs. Given that consumers expect a certain congruity between functional and form designs—they expect beautiful products to work better (Hagtvedt and Patrick 2014)—ostensibly, firms should be well-served by maximizing both dimensions. However, it is not a judicious strategy. The principle of “hedonic domi- nance” suggests that consumers attach relatively greater importance to form design after a baseline level of func- tionality has been met; in such cases, consumers tend to focus on products with superior form design because such features tend to enhance hedonic appeal and generate more excitement, thereby driving consumer choices (Chitturi, Raghunathan, and Mahajan 2007). Products with high levels of both form and function may not appeal to the mainstream market be- cause they tend to overshoot customers’ willingness to pay. As a result, they may end up catering only to a niche, high-end market, thereby making it difficult for manufacturers to maximize market share.

1See http://www.ergonomics.org.uk/learning/what-ergonomics/.

74 / Journal of Marketing, July 2016

T A B L E 1

R e p re s e n ta ti v e R e s e a rc h o n D e s ig n D im

e n s io n s

S tu d y

C o n tr ib u ti o n

D e s ig n

D im

e n s io n s

In te ra c ti o n s

B e tw

e e n

D im

e n s io n s

In c lu d e d ?

A g e o fV

e h ic le

G e n e ra ti o n

In c lu d e d ?

L e v e l o f

A n a ly s is

S a m p le

a n d D a ta

O u tc o m e

V a ri a b le (s )

B lo ch

(1 9 9 5 )

• G o o d fo rm

d e si g n a tt e m p ts to

o p tim

iz e

d e si g n g o a ls

fo r a sp

e ci fi c ta rg e t

m a rk e t w h ile

sa tis fy in g re le va

n t

co n st ra in ts .

• A w e ll- e xe

cu te d fo rm

d e si g n e vo

ke s

p o si tiv e p sy

ch o lo g ic a lr e sp

o n se

s fr o m

co n su

m e rs

a n d is

th u s a ke

y d e te rm

in a n t o f m a rk e tp la ce

su cc

e ss

.

F o rm

N o

N o

C o n ce

p tu a l

P sy

ch o lo g ic a l

a n d b e h a vi o ra l

re sp

o n se

s to

p ro d u ct

fo rm

R in d o va

a n d

P e tk o va

(2 0 0 7 )

• F o rm

d e si g n in fl u e n ce

s cu

st o m e r

p e rc e p tio

n s o f th e va

lu e p o te n tia

l o f

n e w

p ro d u ct s.

• G o o d fo rm

d e si g n e n d o w s th e p ro d u ct

w ith

cu e s th a tt ri g g e r p o si tiv e co

g n iti ve

a n d e m o tio

n a l co

n su

m e r re sp

o n se

s.

F o rm

N o

N o

C o n ce

p tu a l

C o n su

m e r va

lu e

p e rc e p tio

n s

C h itt u ri ,

R a g h u n a th a n ,

a n d M a h a ja n

(2 0 0 7 )

• W h e n a tt ri b u te s o f b o th

fu n ct io n a n d

fo rm

m e e to

r e xc

e e d th re sh

o ld va

lu e s,

co n su

m e rs

a tt a ch

g re a te r im

p o rt a n ce

to th e fo rm

a tt ri b u te .

• S u p e ri o r fo rm

d e si g n re su

lts in

h ig h e r

w ill in g n e ss

to p a y.

F o rm

a n d

fu n ct io n

Y e s (f o rm

· fu n ct io n )

N o

In d iv id u a l

le ve

l E xp

e ri m e n ta l

d a ta

fr o m

u n d e rg ra d u a te

st u d e n ts

C o n su

m e r

e m o tio

n s a n d

ch o ic e

C h itt u ri ,

R a g h u n a th a n ,

a n d M a h a ja n

(2 0 0 8 )

• P ro d u ct s w ith

g o o d fo rm

d e si g n th a t

m e e t o r e xc

e e d cu

st o m e rs ’ h e d o n ic

n e e d s e n h a n ce

cu st o m e r d e lig h t,

w h e re a s fu n ct io n a lly

su p e ri o rp

ro d u ct s

th a t m e e t o r e xc

e e d u til ita

ri a n

e xp

e ct a tio

n s e n h a n ce

cu st o m e r

sa tis fa ct io n .

• D e lig h tin

g cu

st o m e rs

im p ro ve

s cu

st o m e r lo ya

lty , a s m e a su

re d b y

w o rd

o f m o u th

a n d re p u rc h a se

in te n tio

n s, m o re

th a n m e re ly sa

tis fy in g

th e m .

F o rm

a n d

fu n ct io n

Y e s (f o rm

· fu n ct io n )

N o

In d iv id u a l

le ve

l E xp

e ri m e n ta l

d a ta

fr o m

u n d e rg ra d u a te

st u d e n ts

a n d

ve h ic le

o w n e rs

C o n su

m e r

sa tis fa ct io n ,

e m o tio

n s,

re p u rc h a se

in te n tio

n s,

a n d

w o rd

o f m o u th

R u b e ra

(2 0 1 5 )

• G o o d fo rm

d e si g n in fl u e n ce

s p ro d u ct

sa le s;

sp e ci fi ca

lly , it d im

in is h e s in iti a l

sa le s b u t in cr e a se

s sa

le s g ro w th

ra te

o ve

r tim

e .

• B o th

te ch

n o lo g ic a li n n o va

tiv e n e ss

a n d

b ra n d st re n g th

b o o st

th e im

p a ct

o f

fo rm

d e si g n o n sa

le s g ro w th

ra te .

F o rm

N o

N o

F ir m

le ve

l (m

o d e ly e a r)

M u lti ye

a r p a n e l

d a ta

o n ca

rs a n d

m o to rc yc

le s

In iti a l sa

le s a n d

sa le s g ro w th

Designed to Succeed / 75

T A B L E 1

C o n ti n u e d

S tu d y

C o n tr ib u ti o n

D e s ig n

D im

e n s io n s

In te ra c ti o n s

B e tw

e e n

D im

e n s io n s

In c lu d e d ?

A g e o fV

e h ic le

G e n e ra ti o n

In c lu d e d ?

L e v e l o f

A n a ly s is

S a m p le

a n d D a ta

O u tc o m e

V a ri a b le (s )

H o m b u rg ,

S ch

w e m m le , a n d

K u e h n l (2 0 1 5 )

• A sc

a le

to m e a su

re th e th re e

d im

e n si o n s o f p ro d u ct

d e si g n is

d e ve

lo p e d a n d va

lid a te d .

• B ra n d a tt itu

d e m e d ia te s th e

re la tio

n sh

ip s b e tw e e n d e si g n

d im

e n si o n s a n d w o rd

o f m o u th .

F o rm

, fu n ct io n ,a

n d

sy m b o lis m

N o

N o

In d iv id u a l

le ve

l D a ta

o n se

ve ra l

h o u se

h o ld

a n d

m u lti m e d ia

p ro d u ct s in

th e

U n ite

d S ta te s

a n d E u ro p e

P u rc h a se

in te n tio

n a n d w o rd

o f m o u th

T h is

st u d y

• E rg o n o m ic s is a cr iti ca ld

e si g n

d im e n si o n ,a

lo n g w ith

fo rm

a n d fu n ct io n ;

fi rm

s m u st

co n fi g u re

th e le ve ls o f th e se

d im e n si o n s fo r b e st

re su lts .

• O ld e r- g e n e ra tio

n p ro d u ct s w ith

su p e ri o r fo rm

d e si g n d o m u ch

b e tt e r

th a n th o se

o f co

rr e sp

o n d in g

g e n e ra tio

n s w ith

su p e ri o r fu n ct io n o r

e rg o n o m ic s.

F o rm

, fu n ct io n ,a

n d

e rg o n o m ic s

Y e s (f o rm

· fu n ct io n ,

fu n ct io n · e rg o n o m ic s,

a n d fo rm

· e rg o n o m ic s)

Y e s

F ir m

le ve

l (m

o d e ly e a r)

In te g ra te d

re p e a te d cr o ss

- se

ct io n a l d a ta

fr o m

th e U .S .

lig h t ve

h ic le

in d u st ry

M a rk e t sh

a re

76 / Journal of Marketing, July 2016

Considering that enhancing both functional and form dimensions to higher levels is suboptimal for share, firms would do well to strategically balance the two dimensions when designing a product. In mature competitive markets, products have to meet minimal regulated levels of func- tionality in order to compete in the marketplace. Once a product meets the levels required by the marketplace, research on limited attribute compensation has shown that a higher level of form design tends to compensate for a lower level of functionality (Hagtvedt and Patrick 2014). Superior form design does so by acting as a quality signal for the product, making up for the inferior functional attributes and, thereby, communicating value (Hoegg and Alba 2011). More importantly, owing to its conspicuous nature, superior form design is more observable, making the product more socially desirable and, thereby, enabling consumers to fulfill social objectives such as obtaining social status (Noble and Kumar 2008). Social desirability has also been shown to drive con- sumer choice by increasing product likability (Rindfleisch and Inman 1998). Thus, in a choice between two products with similar functionality, consumers prefer products with superior form design that generates higher-order emotions, such as exhilaration and excitement (Chitturi, Raghunathan, and Mahajan 2008).

Along similar lines, Hoegg, Alba, and Dahl (2010) suggest pairing products that possess lower levels of form design with higher perceived functionality; the apparent in- congruity between form and function enables consumers to elaborate on and compensate for lower form design through superior functional attributes. Such a product promotes positive prevention goals such as confidence and security and

satisfies utilitarian expectations (Chitturi, Raghunathan, and Mahajan 2008), thereby leading to better product evaluations. Accordingly, we propose the following:

H1: The positive relationship between form design and market share is weakened at increasing levels of functional design.

Functional and ergonomic designs. While the aim of functional design is to create a product that can perform tasks required to satisfy the utilitarian needs of consumers, ergo- nomic design is focused on enhancing user experiences during consumption and promoting comfort and ease of use. Both functionality and ergonomics lead to positive prevention goals such as confidence and security (Chitturi, Raghunathan, and Mahajan 2008) and improve outcomes such as product choice and satisfaction (Noble and Kumar 2008).

Higher levels of functionality can result in greater complexity due to intricate dependencies among the various elements (Lu and Suh 2009). Given that a complex product may be inconsistent with consumers’ existing values and experiences (Gourville 2005), a highly functional product can be more difficult for a consumer to evaluate, thereby requiring extensive mental effort and cognitive processing (Bloch 1995).The challenge for designers, then, is to achieve the delicate balance of offering highly functional products without cognitively overwhelming the consumers (Berlyne 1970) but, at the same time, helping consumers evaluate the product objectively. Product developers tend to utilize subtle design cues to effectively influence the performance expect- ations of consumers (Noble and Kumar 2010); ergonomic design is one such cue. Research has shown that consumers often utilize their usage experiences as signals not only

FIGURE 1 Conceptual Framework of Hypothesized Interactions

Design

H1 (−)

H2 (+)

H3 (−)

Market Share

Age of vehicle generation

Functional design

Form design

Ergonomic design

H4b (+)

H4a (+)

Designed to Succeed / 77

to understand the operational effectiveness (Creusen and Schoormans 2005) but also to understand and evaluate the functional attributes of a product (Crawford and Di Benedetto 2010). Therefore, higher levels of ergonomic design can not only better articulate a product’s comfort, ease of use, and usability through positive user experiences but can also signal a product’s utility by accentuating the strengths of functionality.

Introducing an excellent functional product without correspondingly high levels of ergonomics can negate the impact of superior functional design because consumers find the product difficult to use and, thereby, fail to ap- preciate the superiority of its functional design. Industry is replete with such examples: in 2007, Volkswagen Beetle failed to reap full benefits of its superior functionality be- cause of its relatively inferior ergonomics. Coupling poor functional design with higher levels of ergonomics is also unwise. Higher levels of ergonomics raise consumers’ expectations about the performance of the product (Noble and Kumar 2010), but a product with inferior functional design will not fulfill these expectations and will thereby lead to dissatisfaction.

The preceding arguments suggest a positive synergy between functionality and ergonomics such that high levels of experiential user satisfaction through superior ergo- nomics amplify positive consumer evaluations due to superior functionality. This synergy results in higher consumer evaluations and better market share. Accordingly, we propose the following:

H2: The positive relationship between functional design and market share is strengthened at increasing levels of ergonomic design.

Form and ergonomic designs. Both form and ergonomic designs communicate value and can influence customer preferences. Form design is visual and observable. At a broad level, superior form design is about mere exposure of novel stimuli (Hoegg and Alba 2011) and thus is much easier and more straightforward to perceive. Ergonomic design, on the other hand, is experience-centric and intensely personal. Con- sumers must interact with a product to evaluate its usability and ease of use.

We extend the limited attribute compensation argument suggested by Hagtvedt and Patrick (2014) to the interplay be- tween form and ergonomics. Akin to the rationale outlined with lower levels of functional attributes, superior form design can compensate for lower levels of ergonomic design by signaling social desirability (Rindfleisch and Inman 1998) and social status (Noble and Kumar 2010). As such, form design communicates product value (Hoegg and Alba 2011) and evokes emotional excitement (Chitturi, Raghunathan, and Mahajan 2008).

In a similar vein, superior ergonomics can compensate for lower levels of form design. From a technical perspective, enhancing ergonomics may lead to aesthetic characteristics, such as shape, size, texture, and color, being suboptimal (Lawson 2005). Given the lack of visual and hedonic appeal, consumers tend to weigh self-experiences heavily in their final evaluations (Hagtvedt and Patrick 2014) and accept that products designed to deliver the most comfort may not be

aesthetically the most appealing (Bloch 1995). In such cases, the ideal product is not necessarily that which possesses the best form design but, rather, the product that is most com- prehensible and usable (Bloch 1995, p. 18). By promoting usability and positive prevention goals such as safety and security, ergonomic design takes a different route than form design in affecting consumer preferences for the product. Thus, we propose the following:

H3: The positive relationship between ergonomic design and market share is weakened at increasing levels of form design.

The Impact of Design on Older- Versus Newer-Generation Vehicles

Empirical studies have shown a decline in sales as a product ages in the marketplace (Bayus 1998). Market demands and consumer preferences for design attributes evolve over time, requiring reconfiguration of design attributes. Competitors gradually catch up to, or even overtake, incumbent’s sales with better-designed products. So firms strive to be dynamic and launch newer generations of their products with superior designs to create better value for their customers and capture share. In this context, relevant questions arise: Can design help slow down the decline in market share as a product gets older? Are some design dimensions more effective than others at doing so?

Form design is conspicuous and becomes a part of the sensory environment for both users and nonusers (Bloch 1995). A novel form design generates high initial awareness and interest among consumers, leading to potential adoption (Rubera 2015). The positive word of mouth accompanying the increased adoption may lead to bandwagon effects, wherein consumers enjoy psychological benefits from using popular products (Hellofs and Jacobson 1999), leading to more consumers buying the product. As such, the popularity for superior form design in the marketplace creates opportunities for repeated exposure that leads to higher levels of familiarity with the product. Given that higher levels of familiarity are associated with higher levels of preference among consumers (Rindfleisch and Inman 1998), a product with superior form design is likely to engender high levels of consumer support. Superior form design is also hard to imitate. This difficulty in imitation arises in part because form design (legally known as “trade dress”) is protected, and firms can successfully sue imitators for design infringement (Gelb and Krishnamurthy 2008). Thus, a product with superior form design is likely to be perceived as unique for a longer time (Bloch 1995). Therefore, consumers are likely to prefer products with superior form design even when the products are no longer new.

Neither functional nor ergonomic design, by and large, is conspicuous. To be perceived, they have to be either expe- rienced by consumers or evaluated technically. The incon- spicuous aspect of these two design dimensions makes it difficult for them to articulate their potential benefits (Noble and Kumar 2010). For example, a new-generation sedan with notably high horsepower rating or comfort is unlikely to get noticed as much as a new-generation sedan with breathtaking new styling, which is highly visible and thus creates band- wagon effects. Also, both functional and ergonomic designs

78 / Journal of Marketing, July 2016

in older-generation vehicles are imitated relatively easily by other firms in a competitive marketplace, making sustained differentiation relatively difficult. Positive feedback from consumers and vicarious learning from suppliers and com- petitors hasten this imitation. So the uniqueness of functional and ergonomic design is likely to be limited to a shorter time frame compared with that of form design. As a result, con- sumers are less likely to prefer older-generation vehicles with superior functionality or ergonomics—newer vehicles may offer similar or better functional and ergonomic attributes. Accordingly, to compare the effects of design dimensions, we propose the following:

H4: Compared with (a) functional design and (b) ergonomic design, the decrease in market share of an older-generation product with a high level of form design is smaller.

Method Data

We collected data from the U.S. light vehicle industry, which consists of cars, crossover-utility vehicles (CUVs), sport-utility vehicles (SUVs), vans, and pickups. Product design plays a key role in the light vehicle industry, as exemplified by several prestigious awards, such as Car Design Awards Global (decided by industry executives worldwide) and World Car Awards (decided by automotive journalists worldwide). We obtained data primarily from four sources—J.D. Power and Associates, Automotive News, TNS Media Intelligence, and MSN Autos—and integrated them into a single repeated cross-sectional data set. We also referred to supplementary data sources for certain variables, which are detailed in the “Description of Variables” section. Our integrated data set consists of 937 observations over a period of six years, from 2002 to 2007, and represents vehicles from 33 brands (see Table 2) and 22 categories (see Table 3).

J.D. Power and Associates, a global marketing informa- tion services company, conducts several survey-based annual studies for which it collects data from new vehicle owners and lessees. For these surveys, the company mails questionnaires in the first half of the year to a randomly selected and geo- graphically distributed sample of almost half a million verified respondents. Such data obtained are the source of J.D. Power’s well-recognized “Power Circle Ratings,” publicly available data that we used in our empirical analysis.

One of these studies, named the APEAL (Automotive Performance, Execution, and Layout), collects informa- tion on more than 80 vehicle attributes and measures owner delight with their new vehicles after 90 days of ownership—that is, how gratifying a new vehicle is to own and drive. The company aggregates the attributes into four measures of owner evaluation.2 We use two of these

measures—style and comfort—to capture form design and ergonomic design, respectively. Another study, named the IQS (Initial Quality Study) collects information on more than 100 vehicle attributes and focuses on the problems ex- perienced by owners during the first 90 days of ownership. Often referred to in the industry as “things gone wrong,” this study is the industry benchmark for new-vehicle quality. The company aggregates these attributes into six measures of owner evaluation of mechanical and design quality of the vehicle. We use one of these measures—powertrain design quality—to capture functional design.

TABLE 2 Brands in the Data

Brand Number of Models Market Share (%)

Acura 7 1.17 Audi 5 .40 BMW 7 1.20 Buick 9 1.75 Cadillac 7 .88 Chevrolet 21 10.28 Chrysler 8 1.87 Dodge 6 3.36 Ford 13 10.22 GMC 6 2.59 Honda 8 6.72 Hummer 2 .17 Hyundai 7 2.22 Infiniti 5 .49 Jaguar 3 .19 Jeep 5 2.07 Kia 7 1.48 Land Rover 5 .08 Lincoln 6 .68 Mazda 8 1.41 Mercedes Benz 9 1.24 Mercury 8 .98 Mitsubishi 5 .50 Nissan 13 4.44 Pontiac 9 1.82 Porsche 4 .11 Saturn 7 1.13 Scion 2 .42 Subaru 5 1.06 Suzuki 6 .19 Toyota 13 9.10 Volkswagen 5 1.36 Volvo 6 .81

Notes: Number of models for each brand indicates the number of models available in the data set. The data do not include all the models for all the brands, and information for some models is available only for a portion of the years in the data set. Some brands replaced a particular model with another during the duration of our data set. For example, Ford replaced the Windstar minivan with the Freestar minivan in 2004, and Mazda replaced the Protégé sedan with the Mazda 3 sedan in 2004. In all such cases, we considered the two models to be the same vehicle for the purpose of model estimation. Each brand’s share of the total market changes every year. Market share shown here is for 2004. On average, our data set ac- counts for 72%–75% of the total industry sales each year. For example, in 2004 the total industry sales was 16.9 million vehicles, and the sales of vehicles included in our data set add up to 12.2 million.

2We have data at the measure level but not at the attribute level. J.D. Power and Associates, well recognized in the market research industry, uses its industry knowledge and factor analysis to form measures from individual attributes. As discussed in the Web Appendix, we also conducted several qualitative exercises and a customer survey to validate these measures.

Designed to Succeed / 79

Automotive News, founded in 1925, is a key source of news and data for auto industry executives across vehicle manu- facturers, original equipment suppliers, and dealers. We ob- tained data on vehicle sales and number of dealerships from this source. We have sales data (in units) at the model and year levels for all vehicles in our data set. For example, the data show that Cadillac sold 54,846 CTS sedans in 2006. We have dealer data at the brand and year levels; for example, the data show that Cadillac was available at 1,469 dealerships at the end of 2006.

The firm TNS Media Intelligence, founded in 1946 and now a part of Kantar Media, is a key source of advertising spending data across industries. In the case of the auto industry, it measures advertising spending across 18 tele- vision, radio, magazine, and Internet media types. We obtained annual advertising data at brand and year levels; for example, the data show that Cadillac spent $224 million on advertising in 2006.

The website MSN Autos (http://www.msn.com/en-us/autos/), owned by Microsoft, is an online resource for researching light vehicles. Among other information, it provides detailed historical pricing data; for example, it shows that the retail price of the basic trim for Cadillac CTS in 2006 was $29,270.

We obtained the prices of all vehicles in our data set from this source.

Description of Variables

Dependent variable. We use the annual market share of each vehicle as the dependent variable. We measured this value as the ratio of focal vehicle’s annual sales to total annual industry sales.

Independent variables. We use four independent varia- bles. First, functional design, which represents the utilitarian aspects of product design, is evaluated using the “powertrain design quality” measure from the IQS study. It takes into account powertrain issues such as smoothness of trans- mission and ride, handling and stability of the vehicle, and responsiveness of the steering system and brakes. Because the survey is based on the first 90 days of ownership, the measure focuses on evaluation of these attributes in the short run rather than in the long run. Given our interest primarily in design-oriented aspects, rather than deterio- ration in functionality over time, the measure captures the essence of the construct. The item was reported on a seven- point scale from 2 to 5, with gradations of .5, where a higher value indicates higher evaluation.

Second, form design, which represents the hedonic as- pects of product design, is evaluated using the “style” mea- sure from the APEAL study. It is based on the evaluation of such attributes as exterior front- and rear-end styling, pro- file styling, interior styling, and exterior and interior colors. The item was reported on a seven-point scale from 2 to 5, with gradations of .5, where a higher value indicates higher evaluation.

Third, ergonomic design, which represents the aspects of product design that improve a product’s usability and enhance pleasure of consumption, is evaluated using the “comfort” measure from the APEAL study. It is based primarily on the evaluation of such attributes as the ease of getting in and out of the vehicle, leg room and head room, seats, seat belt comfort and adjustability, and quality of seating materials. The item was reported on a seven-point scale from 2 to 5, with gradations of .5, where a higher value indicates higher evaluation.

Fourth, to obtain the age of vehicle generation, we gath- ered information from several sources available publicly, including trade magazines, company websites, and third- party websites such as Edmunds.com and Kelley Blue Book to determine the years during which each vehicle launched a new generation. Because almost all the new- generation vehicles were launched toward the end of a cal- endar year, we considered the subsequent calendar year as the first year for the generation. For example, if a new generation vehicle was launched in October 2002, we con- sidered 2003 to be its first year. Most of the sales toward the end of 2002 were likely for the previous-generation vehicle because manufacturers usually offer heavy incen- tives to unload the inventory of a previous-generation vehicle before a year’s end. Moreover, sometimes there is a gradual rollout of a new-generation vehicle such that many markets do not even have the new-generation vehicle in significant numbers in the initial few months of launch. Descriptions of

TABLE 3 Vehicle Categories in the Data

Vehicle Category Number of Models Example

Low-price small carsa 7 Hyundai Accent High-price small carsa 21 Honda Civic Low-price midsize carsa

7 Kia Optima

High-price midsize carsa

21 Honda Accord

Large carsa 7 Toyota Avalon Midsize specialty carsb 11 Mazda RX-8 Low-price luxury cars 18 BMW 3 Series Midprice luxury cars 10 BMW 5 Series High-price luxury cars 4 BMW 7 Series Luxury sports cars 8 BMW Z4 Luxury specialty carsb 7 Mercedes CLK Small CUVs 7 Honda Element Midsize CUVs 16 Honda CRV Midsize luxury CUVs 13 Acura MDX Midsize SUVs 24 Ford Explorer Large SUVs 7 Ford Expedition Midsize luxury SUVs 7 Land Rover LR3 Large luxury SUVs 5 Lincoln

Navigator Small vans 16 Honda Odyssey Large vans 3 Ford Econoline Small pickups 11 Ford Ranger Large pickups 5 Toyota Tundra

a“Cars” includes sedans, wagons, and hatchbacks. b“Specialty cars” includes coupes and convertibles. Notes: Vehicles are categorized according to Ward’s Automotive

Yearbook, which segments vehicles on the basis of their body style, size, and price. Number of models for each category indicates the number of models available in the data set. The data do not include all the models for all the categories, and information for some models is available only for a portion of the years in the data set. CUV = crossover-utility vehicle; SUV = sport-utility vehicle.

80 / Journal of Marketing, July 2016

all variables, along with their operationalizations and sources, appear in Table 4.

Control variables. We controlled for several factors that might influence a vehicle’s market share. First, recognizing that the sales in the last year of a vehicle’s generation could be higher because of stock-clearing incentives and deals, we controlled for this effect by using a binary measure that we call “last year of a generation.” In the previous example, this measure would take the value 1 in 2002 and 0 otherwise.

Second, we controlled for any potential effects of vehicle’s relative price. This variable is obtained as the ratio of focal vehicle’s price to the average vehicle price in the industry (Buzzell and Wiersema 1981). However, price of a vehicle varies depending on several factors, such as trim level, optional features, rebates, financial incentives, and trade-in programs.3 To keep prices comparable, we referred to the manufacturer’s suggested retail price (MSRP) of each vehicle’s base trim level.

Third, we controlled for any potential effects of market coverage. Market share of a brand might be associated with the number of dealerships at which it is available. Thus, we accounted for the dealer share each brand had in the market (Bell, Keeney, and Little 1975), which is obtained as the ratio of focal brand’s number of dealerships to the total number of auto dealerships.

Fourth, we controlled for any potential effects of ad- vertising. Arguably, higher advertising spending might be associated with a larger market share. Thus, we accounted for the advertising share each brand had in the market, which is obtained as the ratio of focal brand’s advertising expense to the total advertising expense in the industry.

Finally, we controlled for any potential effects of a vehicle’s ratings on its sales. There is ample trade evi- dence to suggest that customers consider such ratings before purchasing a new vehicle. More favorable ratings can bring certain vehicles to customers’ attention and may even in- fluence purchase decisions. There are usually two kinds of ratings in the auto industry: ratings by editors and staff of third-party sources such as Edmunds.com and Consumer Reports, and ratings by vehicle owners and lessees. Editors can be considered experts, whereas most of the owners are relative novices. We control for potential influences of both editor ratings and consumer ratings.

We obtained data on vehicle ratings from Edmunds.com. The company was the first to launch an automotive infor- mation website, in 1995, and it is considered the premier online resource in the industry. From 2002 to 2007, each year Edmunds.com announced “Editors’ Most Wanted Vehicles” according to votes cast by its staff writers and “Consumers’ Most Wanted Vehicles” according to readers’ votes. We used information about Editors’ Most Wanted Vehicles to measure editor rating of vehicles and information about Consumers’

Most Wanted Vehicles to measure consumer rating of vehi- cles. These are both binary measures that take the value 1 if a particular vehicle was recognized in a particular year. De- scriptive statistics and the correlation matrix for all variables appear in Table 5.

Model and Estimation We estimated a linear mixed model to accommodate the repeated cross-sectional nature of the data (Stroup 2013):

y = Xb + Zg + e,(1)

where y is an n · 1 vector of the response variable; X is an n · p matrix of independent variables; b is a p · 1 parameter vector of fixed effects; Z is an n · r design matrix for random effects; g is an r · 1 parameter vector of random effects, normally distributed with mean 0 and variance G; and e is an n · 1 random error vector, normally distributed with mean 0 and variance R.

The variance of y thus is V = ZGZ9 + R, which can be modeled by setting up the random effects design matrix Z and specifying covariance structures for G and R. We specify the vehicle’s category and brand, and their cross-level, as random effects and a variance-components covariance structure for both G and R (Stroup 2013). We use restricted maximum likelihood to construct an objective function and maximize it over all unknown parameters (SAS 2008; Stroup 2013). The corresponding log-likelihood function is

lRðG, RÞ = - �n - p

2

� logð2pÞ - 1

2 logjVj

- 1 2 log

��X9V-1X�� - 1 2 r9V-1r,

(2)

where p = rank (X) and r = y - XðX9V-1XÞ-X9V-1y. We minimized this function using a ridge-stabilized

Newton–Raphson algorithm. The estimation runs iteratively: initial values of the variance–covariance matrix are used to compute bG and bR, which allow computation of bb and bg. These are used to determine new values of the score vector and the Hessian matrix and update the variance–covariance matrix. At convergence, parameters and random effects are estimated as

bb = �X9bV-1X�-X9bV-1y, and(3) bg = bGZ9bV-1�y - Xbb�.(4)

Regarding specification of independent variables, we use an attraction-type specification for three independent variables, namely, price, distribution strength, and advertising. This is driven by market share theorem, which states that market shares are proportional to shares of total marketing effort (Barnett 1976; Bell, Keeney, and Little 1975; Buzzell and Wiersema 1981).4 We use grand-mean-centered values for the remaining continuous variables.5 Furthermore, we built

3Our data pertain to the entire U.S. auto industry, and all variables in the data set vary annually. However, rebates and financial incentives usually pertain to specific trim levels, vary geographically, vary seasonally,and change frequently (sometimes at weekly intervals). So any attempt to aggregate financial incentives would offer ques- tionable insights at best. Thus, we do not explicitly account for the role of incentives.

4We thank an anonymous reviewer for this suggestion. 5We also estimated a model in which we used attraction-type

specification (primarily relative indices) for the remaining con- tinuous variables. The results for the hypothesized interactions are substantively similar.

Designed to Succeed / 81

the model incrementally and use Bayesian information cri- teria (BIC) to evaluate incremental nested models (Stroup 2013). The results appear in Table 6.

Results To begin our analysis, we first specified a model with an intercept and dummy variables for years only. Next, we included random intercept effects to account for variation in market share because of any potential category- or brand- specific factors. The comparison of these two nested models indicates significant improvement in the model fit (ΔBIC(2) = 620, p < .01).

Empirical best linear unbiased predictors (BLUP) for random effects indicate that vehicles in some categories had significantly higher or lower market shares. For example, midsize specialty cars (e.g., Mazda RX-8), large SUVs (e.g.,

Ford Expedition), luxury specialty cars (e.g., Mercedes CLK), and luxury sports cars (e.g., BMW Z4) had sig- nificantly lower market shares. In contrast, low-price small cars (e.g., Hyundai Accent), high-price small cars (e.g., Honda Civic), large pickups (e.g., Toyota Tundra), and low- price luxury cars (e.g., BMW 3 Series) had significantly higher market shares. Results also indicate a significant variation in market shares of vehicles belonging to differ- ent brands. Vehicles belonging to Chevrolet, Dodge, Ford, GMC, Honda, Jeep, Nissan, and Toyota had higher market shares than vehicles belonging to Audi, Buick, Cadillac, Jaguar, Mazda, Mercury, Mitsubishi, Pontiac, Subaru, Suzuki, Volkswagen, and Volvo had. These results seem generally plausible.

Next, we included a third random intercept effect to ac- count for potential variation in market share because of any category- and brand-level cross-interactions (labeled Model 1

TABLE 4 Variable Description and Data Sources

Variable Label Definition Operationalization/Measure Data Source

Market share Annual market share in the entire U.S. light vehicle market

Ratio of focal vehicle’s annual sales to total annual industry sales (%)

Automotive News

Functional design

Evaluation of vehicle based on issues with powertrain performance during driving

“Powertrain design quality” measure from IQS study (seven-point interval scale), based on attributes such as smooth transmission, smoothness of ride, handling and stability of the vehicle, and responsiveness of the steering system and brakes

J.D. Power and Associates (JDP)

Form design Evaluation of vehicle based on hedonic aspects

“Style” measure from APEAL study (seven- point interval scale), based on attributes such as exterior front- and rear-end styling, profile styling, interior styling, and exterior and interior colors

JDP

Ergonomic design

Evaluation of vehicle based on comfort, convenience, and user friendliness

“Comfort” measure from APEAL study (seven- point interval scale), based on attributes such as seats, seat belt comfort and adjustability, quality of seating materials, leg and head room, and ease of getting in and out of the vehicle

JDP

Age of vehicle generation

Number of years since the generation was first launched in the U.S. market

Age with respect to first full calendar year of a vehicle’s availability

Multiple sourcesa

Last year of a generation

Year when vehicle was replaced by a newer generation

Dummy variable indicating last calendar year of the vehicle generation

Multiple sourcesa

Relative price Price of the vehicle indexed to the average price in the industry

Manufacturer’s suggested retail price (MSRP) of a vehicle’s basic trim divided by the average MSRP of basic trims for all vehicles

MSN Autos

Dealer share Proportion of dealerships at which the brand is available for purchase

Number of dealerships at which a brand is available divided by the total number of dealerships

Automotive News

Advertising share A brand’s annual advertising spending as a proportion of all brands’ annual advertising spending

Annual advertising spending by a brand divided by the total annual advertising spending by all the brands

TNS Media

Editor rating Whether a vehicle won the Edmunds Editors’ Most Wanted award

Binary variable (0 = no; 1 = yes) Edmunds.com

Consumer rating Whether a vehicle won the Edmunds Consumers’ Most Wanted award

Binary variable (0 = no; 1 = yes) Edmunds.com

aIncludes publicly available sources such as trade magazines, company websites, and third-party sites such as Edmunds.com and Kelley Blue Book.

82 / Journal of Marketing, July 2016

in Table 6). We accounted for this variation because some brands are perceived as significantly stronger or weaker in certain vehicle categories. For example, results show that Honda has significantly higher market share in high-price midsize cars (Honda Accord) and high-price small cars (Honda Civic) but has significantly lower market share in midsize SUVs (Honda Pilot) and small pickups (Honda Ridgeline). Similar differences exist for several other brands, including BMW, Chevrolet, Chrysler, Dodge, Ford, GMC, Hyundai, Jeep, and Toyota. The comparison of these two nested models indicates significant improvement in the model fit when this random intercept effect is included (ΔBIC(1) = 355, p < .01).

Tests of Hypotheses

Next, we added control variables and the main effects of the three design dimensions (form, function, and ergonomics; labeled Model 2 in Table 6). These additions produced sig- nificant improvement in model fit (ΔBIC(10) = 144, p < .01). Finally, we added the hypothesized interaction terms to create the full model, Model 3. Adding these terms significantly improved the model fit (ΔBIC(6) = 25, p < .01), indicating that the interaction effects are jointly significant. The results from Models 1, 2, and 3 appear in Table 6.

Our discussion will focus on Model 3, the full model. This model shows significant simple effects for functional design (b = .025, p < .05) and ergonomic design (b = .027, p < .10). A vehicle with functional design one standard devia- tion above average has .024% higher market share, and a vehicle with ergonomic design one standard deviation above average has .025% higher market share. Given the size of the light vehicle industry in the United States, wherein gaining market share of .01% in 2006 would have translated into additional sales of almost $50 million, these results are managerially significant. The simple effect of form design on market share, at mean levels of other design dimensions, is not statistically significant (b = .011, n.s.). Results show a significant neg- ative effect for the age of vehicle generation (b = -.017, p < .01). As a generation gets older, its market share

declines—on average, it loses .017% market share every year, which reinforces the need for firms to launch new generations.

Examining the interaction effects reveals that the rela- tionship between form design and market share is negatively moderated by functional design (b = -.041, p < .01) in sup- port of H1. To get an in-depth sense of the interaction, we used simple slope analysis, which overcomes the need to create subgroups of continuous independent variables (Aiken and West 1991). Per convention, we assumed base values (i.e., 0) for all categorical variables and mean values for all continuous variables. Results (see Figure 2, Panel A) show that increase in form design is associated with increase in market share when functional design is lower. When fun- ctional design is higher, however, increase in form design is associated with decrease in market share; a vehicle with both higher form and higher functional design may start hitting the limits of affordability. Overall, either increasing function when form is low or increasing form when function is low is beneficial in terms of market share increases.

Consistent with H2, the relationship between functional design and market share is positively moderated by ergo- nomic design (b = .026, p = .07). However, as the p-value indicates, this effect is only marginally supported. Results of simple slope analysis (Figure 2, Panel B) show that increase in functional design is associated with increase in market share when ergonomic design is high. When ergonomic de- sign is low, however, increase in functional design has no significant association with market share. Thus, improving functional design seems to provide greater market share returns only when ergonomic design is high.

Consistent with H3, the positive relationship between ergonomic design and market share is negatively moderated by form design (b = -.027, p < .05). Results of simple slope analysis (see Figure 2, Panel C) show that increase in ergo- nomic design is associated with increase in market share when form design is low. When form design is high, how- ever, increase in ergonomic design has no significant asso- ciation with market share. Thus, increasing ergonomics provides greater market share only when form design is low.

TABLE 5 Descriptive Statistics and Correlation Matrix

Variable M SD Min Max 1 2 3 4 5 6 7 8 9 10 11

1. Market share (%) .47 .46 .005 2.591 1.00 2. Functional design 3.44 .94 2 5 .00 1.00 3. Form design 3.25 .92 2 5 -.31 .30 1.00 4. Ergonomic design 3.31 .91 2 5 -.20 .27 .58 1.00 5. Age of vehicle generation 4.10 2.58 1 11 -.05 .01 -.24 -.26 1.00 6. Last year of generationa .10 0 1 .05 .01 -.07 -.07 .22 1.00 7. Relative price 1.00 .51 .339 3.377 -.37 .36 .50 .51 -.04 -.03 1.00 8. Dealer shareb .04 .03 .004 .105 .35 -.04 -.43 -.40 .21 -.05 -.26 1.00 9. Advertising shareb .04 .03 .002 .111 .50 -.03 -.36 -.30 .13 -.00 -.31 .56 1.00 10. Editor ratingc .10 0 1 .05 .02 .15 .16 -.07 .01 .12 -.11 -.03 1.00 11. Consumer ratingd .10 0 1 .10 .08 .20 .21 -.08 .01 .11 -.11 .01 .28 1.00

aProportion of vehicles in the last year of their generation. bBased on brand-level values. cProportion of vehicles recognized as Editor’s Most Wanted by Edmunds. dProportion of vehicles recognized as Consumer’s Most Wanted by Edmunds. Notes: Correlations greater than .06 (absolute value) are significant at the .05 level (n = 937).

Designed to Succeed / 83

Results show that the negative relationship between age of vehicle generation and market share is moderated by form design (b = .012, p = .02). There appears to be no such moderation effect for functional design (b = ‒.006, n.s) or ergonomic design (b = .001, n.s.), in support of both H4a and H4b. Simple slope analyses for vehicles with higher levels of each design dimension (see Figure 3) show that older vehicles with higher form design do relatively better in terms of retaining market share than older vehicles with either higher functional or higher ergonomic design.

Other Results

Price has a negative impact on market share—an increase of .01 in the relative price index is associated with a decrease of .001% market share. Both dealer share and advertising share have a positive impact on market share—an increase of .01%

in a brand’s dealer share is associated with an increase of .038% in market share, and an increase of .01% in a brand’s advertising share is associated with an increase of .029% in market share. Results also indicate that both editor and consumer ratings of vehicles impact sales: a vehicle recog- nized by an editor award has approximately .056% higher market share than a vehicle not recognized by an editor award, and a vehicle recognized by a consumer award has about .054% higher market share than a vehicle not recog- nized by a consumer award.

Robustness Checks

First, although the effects of interactions between design dimensions—primarily driven by theory—are not expected to vary across vehicle categories or brands, it could be argued that their main effects might differ. For example, function

TABLE 6 Results from Linear Mixed Model Estimation

Model 1 Model 2 Model 3

Coefficient SE Coefficient SE Coefficient SE

Hypothesized Interactions H1: Form design · Functional design -.041*** .014 H2: Functional design · Ergonomic design .026* .014 H3: Ergonomic design · Form design -.027** .013 H4: Form design · Age of vehicle generation .012** .005 Functional design · Age of vehicle generation -.006 .004 Ergonomic design · Age of vehicle generation .001 .005

Covariates Functional design .027*** .010 .025** .010 Form design -.003 .017 .011 .018 Ergonomic design .035** .016 .027* .016 Age of vehicle generation -.020*** .004 -.017*** .004 Last year of generation .038 .027 .036 .026 Relative price -.156*** .040 -.127*** .041 Dealer share 3.744*** 1.151 3.793*** 1.138 Advertising share 2.680*** .922 2.897*** .913 Editor rating .045 .030 .056* .030 Consumer rating .037 .030 .054* .030

Dummy Variable for Year Year = 2002 .085*** .027 .082*** .026 .071*** .026 Year = 2003 .067*** .026 .070*** .025 .058** .025 Year = 2004 .043* .026 .046* .024 .040* .025 Year = 2005 .012 .025 .005 .024 .001 .024 Year = 2006 .023 .024 .020 .023 .018 .023

Intercept .272*** .056 .232*** .074 .218*** .074

Covariance Parameter Estimates Intercept Brand .045*** .015 .013* .008 .013* .008 Vehicle category .023*** .010 .014** .007 .016*** .007 Brand · Vehicle category .062*** .009 .067*** .009 .065*** .009

Residual .050*** .003 .043*** .002 .042*** .002

BIC 284.8 140.9 116.0

ΔBIC 143.9*** 24.9***

*p < .1 (two-tailed test). **p < .05 (two-tailed test). ***p < .01 (two-tailed test). Notes: Number of observations = 937; number of brands = 33; number of vehicle categories = 22.

84 / Journal of Marketing, July 2016

could be more important in SUVs, ergonomics more im- portant in sedans, and aesthetics more important in luxury vehicles. Or function could be more important in Japanese brands, ergonomics more important in domestic brands, and aesthetics more important in European brands. We estimated two additional models: one with random effects of design dimensions at the vehicle category level and the other with random effects of design dimensions at brand level. The BIC showed no significant improvement, and the covariance pa- rameter estimates indicated that the effects did not differ significantly across categories or brands.

Second, although we use grand-mean-centered values of design dimensions in the final model, we also estimated an additional model in which we used relative indices for all remaining continuous variables, in line with market share theorem. The results stayed substantively similar.

Third, it could be argued that a parent company’s market power or resource richness could provide certain unobserved advantages to some vehicles. Although we included the effects of both advertising share and dealer share, as well as random effects at the brand level, to account for such factors, we estimated an additional model in which we explicitly included a measure of parent company’s resource richness. We measured resource richness by calculating the focal par- ent company’s overall market share. For example, in 2006, General Motors had an overall market share of almost 19% in the United States. Results showed that the overall market share of a parent company was not significantly associated with the market share of any particular vehicle.

Fourth, we have used the prices of basic trims as the measure of vehicle price. It could be argued that all basic trims are not comparably equipped; for example, some basic trims may be equipped with automatic transmission, whereas others may be equipped with manual transmission and offer automatic transmission only as an option. It is, however, a difficult proposition to obtain prices of per- fectly comparable vehicle trims, given the large number of available options and features, along with a wide range of consumer preferences. From 2002 to 2005, Automotive News provided pricing information on basic trims, along with prices for four options, where applicable: automatic transmission, air conditioning, antilock brakes, and sunroof/ moonroof. For these four years only (2002–2005), we esti- mated an additional model in which we measured price as the sum of prices for basic trim level, automatic transmission, air conditioning, and antilock brakes. The results obtained were substantively similar.

Fifth, although we controlled for relative price, it could be argued that there are trade-offs related to specific design dimensions and price. We estimated an additional model in which we included interaction terms of relative price with each design dimension. The effects of hypothesized inter- actions remained statistically significant. Thus, interactions between design dimensions seem to affect market share in- dependent of the effect of their interactions with price.

Sixth, it could be argued that the effects of advertising can carry over into subsequent years. For this reason, we esti- mated three additional models. In the first case, before cal- culating brand’s advertising share, we obtained an arithmetic

FIGURE 2 Simple Slope Analysis of Interactions Between

Design Dimensions

A: Effect of Interaction Between Functional Design and Form Design on Market Share

B: Effect of Interaction Between Ergonomic Design and Functional Design on Market Share

.26

.31

.36

.41

.46

.51

.56

Low High

M a

rk e

t S

h a

re

Functional Design

.073***

–.023

.14

.21

.28

.35

.42

.49

.56

Low High

High function Lo

w fu

nc tio

n H

ig h

er go

no m

ic s

Low ergonomics

High form

Lo w

fo rm

M a

rk e

t S

h a

re

Form Design

–.067**

.089**

C: Effect of Interaction Between Form Design and Ergonomic Design on Market Share

.20

.25

.30

.35

.40

.45

.50

Low High

M a

rk e

t S

h a

re

Ergonomic Design

.076**

–.022

*p < .10 (two-tailed test). **p < .05 (two-tailed test). ***p < .01 (two-tailed test). Notes: Simple slope values are shown beside each plotted segment.

Low levels of each dimension are two standard deviations below average; high levels of each dimension are two standard deviations above average.

Designed to Succeed / 85

average of both the focal brand’s spending as well as the total industry spending for the current year and two previous years. In the second case, we obtained a weighted average using weights of .50 for current year spending, .33 for the previous year’s spending, and .17 for spending two years back. In the third case, we used weights of .17 for current year spending, .33 for the previous year’s spending, and .50 for spending two years back. We got substantively similar results in all three cases.

Discussion Product design has been acknowledged as a critical source of competitive advantage (Martin 2009) and has thus become a significant topic of examination for scholars and practitioners alike. Evidence of the impact of various design dimensions on market share would help validate how these dimensions can be utilized as strategic levers for product differentiation. Recent work by Landwehr, Labroo, and Herrmann (2011) demonstrates that incorporating design aspects improves the performance of sales forecasting models, thereby underscoring the importance of design for positive product-level outcomes. We contribute to, and extend, the design literature in three distinct ways. First, we point to, and integrate, ergonomics as a critical dimension of design. Second, we examine the interplay between the various design dimensions and their impact on market share. While most existing work has typi- cally focused on consumer-based evaluative measures, our focus on market share is new to the design literature. Third, given that new generation products are increasingly expensive to develop, we examine the impact of these design dimensions

on older-generation vehicles. Our study does so by conducting empirical tests on an integrated repeated cross-sectional data set from the U.S. light vehicle industry.

This article heeds the calls to better identify and dif- ferentiate the essential components of product design (Srinivasan et al. 2012; Veryzer and Hutchinson 1998). We do these things by advancing a richer, tripartite concep- tualization of product design that includes form, function, and ergonomics as the three critical dimensions of design. In the process, we validate ergonomics as an important com- ponent of the design construct. This represents a significant addition to the extant literature, which has predominantly focused on function and form (Chitturi, Raghunathan, and Mahajan 2007, 2008). To the best of our knowledge, ours is the first study to demonstrate a positive empirical relation- ship between ergonomics and market share.

Our examination of the interplay between the specific design dimensions revealed an interesting design capability trade-off for firms. Firms can either “design for satisfaction,” by meeting utilitarian needs on both function and ergo- nomics, or “design for delight,” by delivering on form and meeting the hedonic needs of consumers (Chitturi, Raghunathan, and Mahajan 2008). Manufacturers would do well to develop their competencies according to their innate strengths. For example, a firm with dominant strengths in form design would do well to continue to be the best in aesthetics rather than try to compete on function; straddling both form and function may lead to confused positioning in the marketplace.

The underpinnings for the capability trade-off are obtained from the interplay among the various design dimensions. Given that consumers explicitly consider these interrelationships in their decision-making process, these findings provide prescriptive guidance to practitioners for increasing market share. Specifically, form and function possess a substitutable relationship such that a product with excellent form design can be successful with a lower level of functional design, and vice versa. For example, for a customer who values form design as highly important, a product with superior form design will be highly valued as long as the product meets the minimal regulated levels of functionality required to operate in the market. It appears that consumers exhibit a certain level of tolerance and tend to compensate for lower levels of one of these two dimensions as long as the other dimension is high (see Hagtvedt and Patrick 2014). Therefore, simultaneous investments in both form and function may actually be counterproductive because the product may end up being “overdesigned.” From a demand-side perspective, a product that is strong on both form and function is likely to hit the limits of afford- ability and may be relegated to a niche status.

The same logic extends to form and ergonomics as well. Maximizing both form and ergonomics, however, is suboptimal for firms. Enhancing ergonomics imposes for- midable technical constraints that can affect aesthetic charac- teristics such as shape, size, texture, and color (Lawson 2005), so a balance between the two dimensions is required. Con- sumers appear to be willing to live with lower-than-average looks as long as the product compensates with higher levels of

FIGURE 3 Simple Slope Analysis of Interactions Between Age of Vehicle Generation and Each Design Dimension

.370

.375

.380

.385

.390

.395

.400

Low Generation Age (Newer)

High Generation Age (Older)

M a

rk e

t S

h a

re

Age of Vehicle Generation

–.017***

–.006

–.023***

Ergonom ics

FunctionForm

*p < .10 (two-tailed test). **p < .05 (two-tailed test). ***p < .01 (two-tailed test). Notes: Simple slope values are shown beside each plotted segment.

Levels of each design dimension are one standard deviation above average.

86 / Journal of Marketing, July 2016

user experiences, or they are willing to buy a great-looking product that is lower on the comfort dimension. Simultaneous investments on both form and ergonomics may produce lower returns that are not commensurate with higher costs.

Our results also indicate that functionality and ergonom- ics have synergies. Better user experiences accentuate per- ceptions of functionality; a high-performing product with great user experience is likely to be highly valued by con- sumers. Conversely, neither a poorly performing product with great comfort features nor a high-performing product that is uncomfortable to use is likely to find much traction in the market. Given that increasing both dimensions positively impacts market share, simultaneous investments in both function and ergonomics dimensions are worthwhile. We have provided a summary of the managerial takeaways from these findings in Figure 4, which also provides a visual de- piction of the interplay among various design dimensions along with the prototypes in light vehicle market.

Next, we examined the impact of these design dimensions on older-generation products. Broadly, our results suggest

two distinct paths. For firms that focus on function and er- gonomics, frequently launching newer-generation products with both improved functionality and improved ergonomics is critical. Function and ergonomics are easily imitable, so sustained differentiation is difficult. Moreover, functional and ergonomic benefits must be analyzed and evaluated or experienced personally and are hard to communicate. In con- trast, firms that are focused on superior form design can af- ford to wait longer to introduce newer-generation products. The conspicuous nature of form design favors bandwagon effects, and consumers like to be associated with popular products. Moreover, it is difficult to successfully and con- tinuously imitate form design. As a result, a firm focused on superior form design has more time to develop a newer- generation product with even better aesthetics. From a prag- matic viewpoint, this knowledge about the different effects of design dimensions can help firms prolong the profitability and tenure of each product and generate appropriate rents before the product becomes obsolete and is replaced by a newer generation.

FIGURE 4 Product Prototypes Across the Design Dimensions and Managerial Prescriptions

Interplay Between Dimensions Managerial Takeaway Form and Function

Invest in your “core” design capability; do not overinvest in both form and function.

Increasing function when form is lower or increasing form at lower levels of function are beneficial strategies. Higher levels of both function and form lead to lower returns.

Harness the synergies between function and ergonomics.

Increasing both function and ergonomics yields superior returns. A lower level of one dimension accompanied by a higher level of the other hurts share.

Function and Ergonomics

Form and Ergonomics Invest in either form or ergonomics; do not overinvest in both form and ergonomics.

Increasing ergonomics at lower levels of form or increasing form at lower levels of ergonomics benefits firms, whereas increasing both form and ergonomics is suboptimal in terms of market share.

–2 SD

0 +2 SD

.15

.25

.35

.45

.55

–2 SD –1 SD 0 +1 SD

+2 SD

M a rk

e t

S h

a re

–2 SD

–1 SD

0 +1 SD +2 SD

.20

.30

.40

.50

.60

–2 SD –1 SD

0 +1 SD

+2 SD

M a rk

e t

S h

a re

–2 SD

0

+2 SD .2

.3

.4

.5

–2 SD –1 SD

0 +1 SD

+2 SD

M a rk

e t

S h

a re

Function

Function

Form

Form

Ergonomics

Ergonomics

High

High

High

High

High

High

Chevy Malibu 2006 Nissan Murano 2003

Chrysler 300 2006Chevy Aveo 2006

VW Beetle 2007

VW Touareg 2007Suzuki Forenza 2007

Honda Odyssey 2002 Volvo S60 2002

Kia Rio 2002 Ford Mustang 2004

Honda Accord 2003

Low

Low

Low

Low

Low

Low

Designed to Succeed / 87

Limitations and Future Research

Although this research contributes to an enhanced under- standing of product design, the empirical analysis has a few limitations that open up several avenues for further study. First is the issue of generalizability. This study is based on a single industry. Moreover, the light vehicle industry is in the mature stage of its life cycle, and the results might be idio- syncratic to this phase. Therefore, it would be worthwhile to examine whether the findings of this study can be replicated in industries in both the growth and the mature phases. Specifically, it would be useful to determine whether the interplay between different dimensions stays consistent across various stages of product life cycle. Also, our data pertain only to the U.S. market. Future studies should add to this literature by examining these issues in diverse markets and identifying and comparing any differences in the impact of design dimensions across geographic contexts.

A second limitation pertains to some of our measures. The use of market share as the dependent variable implies that manufacturers rely simply on consumer evaluations of trade- offs between design dimensions as the key factor in customers’ purchase decisions; it ignores the inherent cost considerations. Future studies might incorporate more cost-based measures. Our measures of dealerships and advertising spending are at the brand level, not at the vehicle level. Furthermore, we used only one source, Edmunds.com, for data about vehicle ratings. Obtaining information at the vehicle level from a variety of sources would further help in establishing validity of these findings. Also, we considered all new-generation vehicle launches as the same in this study. Future research should investigate the impact of interplay between different dimen- sions of product design on market shares of different kinds of new-generation products, such as radical and incremental new products (Chandy and Tellis 1998).

It also should be noted that since these data were col- lected, the light vehicle industry has undergone tremendous change in response to economic recession, gas prices, reg- ulations, evolving consumer tastes, and market share shifts among manufacturers. We have, however, no reason to believe that these changes would affect the hypothesized pattern of effects, which are primarily driven by theory, but specific effect sizes may well be different.

It would also be interesting to study whether consumer re- sponses to the three dimensions of design differ. For example, does functional design lead to higher consumer awareness, form design lead to more consumer attention, and ergonomic design lead to more trial? Evidence of differences in the nature of consumer responses (e.g., cognitive vs. affective) would be a significant contribution to the design literature. It is also well known that customers vary in their preferences for, and ability to cope with, product novelty (Rindova and Petkova 2007). For example, experts tend to have more knowledge about a product category, whereas novices tend to have less knowledge. Thus, it will be valuable to investigate how the impacts of function, form, and ergonomics on prod- uct evaluations vary depending on user expertise. It would also be useful to identify different customer segments within an industry and determine whether the complementary and substitutable relationships between the design dimensions hold across disparate customer segments.

Finally, researchers might also consider competitive and temporal dynamics that could potentially reveal noteworthy findings. For example, do competitive pressures enhance the importance of one design dimension over another in the short run or over a period of time? Also, an examination of the persistence of design dimensions using time-series data would be a worthwhile endeavor. Answers to these strategic questions involving competition and temporal dynamism are essential to the creation of dynamic design strategies that provide sus- tainable competitive advantage. Results of such studies can provide prescriptive guidance to help product developers uti- lize design appropriately in competitive milieus. More im- portantly, they can establish design as a dynamic capability that helps companies sense market opportunities and seize leadership outcomes for long-run competitive advantage.

In conclusion, we augment extant research, which has highlighted the impact of design on individual-level attitudes and behaviors (Chitturi, Raghunathan, and Mahajan 2007, 2008), with aggregate product-level outcomes. We provide a multidimensional view of design and demonstrate that the interplay between the three constituent dimensions, namely, form, function, and ergonomics, influences market share. Good design enables firms to create value for and capture value from customers, thereby leading to positive firm-level outcomes.

REFERENCES Aiken, Leona S. and Stephen G. West (1991), Multiple Regression:

Testing and Interpreting Interactions. London: Sage Publications. Barnett, Arnold I. (1976), “More on a Market Share Theorem,”

Journal of Marketing Research, 13 (February), 104–09. Bayus, Barry L. (1998), “An Analysis of Product Lifetimes in a

Technologically Dynamic Industry,” Management Science, 44 (6), 763–75.

Bell, David E., Ralph L. Keeney, and John D.C. Little (1975), “A Market Share Theorem,” Journal of Marketing Research, 12 (May), 136–41.

Berlyne, D.E. (1970), “Novelty, Complexity, and Hedonic Value,” Perception & Psychophysics, 8 (5), 279–86.

Bloch, Peter H. (1995), “Seeking the Ideal Form: Product Design and Consumer Response,” Journal of Marketing, 59 (July), 16–29.

——— (2011), “Product Design and Marketing: Reflections After Fifteen Years,” Journal of Product Innovation Management, 28 (July), 378–81.

Buzzell, Robert D. and Frederik D. Wiersema (1981), “Modelling Changes in Market Share: A Cross-Sectional Analysis,” Stra- tegic Management Journal, 2 (1), 27–42.

Chandy, Rajesh K. and Gerard J. Tellis (1998), “Organizing for Radical Product Innovation: The Overlooked Role of Willingness to Can- nibalize,” Journal of Marketing Research, 35 (November), 474–87.

Chitturi, Ravindra, Rajagopal Raghunathan, and Vijay Mahajan (2007), “Form Versus Function: How the Intensities of Specific Emotions Evoked in Functional Versus Hedonic Trade-Offs Mediate Product Preferences,” Journal of Marketing Research, 44 (November), 702–14.

88 / Journal of Marketing, July 2016

———, ———, and ——— (2008), “Delight by Design: The Role of Hedonic Versus Utilitarian Benefits,” Journal of Marketing, 72 (May), 48–63.

Cho, Hallie S., Sameer Hasija, and Manuel Sosa (2015), “How Important Is Design for the Automobile Value Chain?” Working Paper No. 2015/86/TOM, INSEAD.

Crawford, Merle and Anthony Di Benedetto (2010), New Products Management, 10th ed. Boston: McGraw-Hill.

Creusen, Marielle and Jan Schoormans (2005), “The Different Roles of Product Appearance in Consumer Choice,” Journal of Product Innovation Management, 22 (1), 63–81.

De Albuquerque Campos, L.F., J.N. de Lima Lanutti, and L.C. Paschoarelli (2012), “Product Functions: Interfaces with Ergo- nomic Design,” Work (Reading, Mass.), 41, 960–63.

Gelb, Betsy D. and Partha Krishnamurthy (2008), “Protect Your Product’s Look and Feel from Imitators,” Harvard Business Review, 86 (10), 36.

Gourville, John T. (2005), “Note on Innovation Diffusion: Rogers’ Five Factors,” Background Note 505-075, Harvard Business School, Cambridge, MA.

Hagtvedt, Henrik and Vanessa M. Patrick (2014), “Consumer Response to Overstyling: Balancing Aesthetics and Functionality in Product Design,” Psychology and Marketing, 31 (7), 518–25.

Hellofs, Linda L. and Robert Jacobson (1999), “Market Share and Customers’ Perceptions of Quality: When Can Firms Grow Their Way to Higher Versus Lower Quality?” Journal of Marketing, 63 (January), 16–25.

Hirschmann, Elizabeth (1983), “Aesthetics, Ideologies, and the Limits of the Marketing Concept,” Journal of Marketing, 47 (July), 45–55.

Hoegg, JoAndrea and Joseph W. Alba (2011), “Seeing Is Believing (Too Much): The Influence of Product Form on Perceptions of Functional Performance,” Journal of Product Innovation Management, 28 (3), 346–59.

———, ———, and Darren W. Dahl (2010), “The Good, the Bad, and the Ugly: Influence of Aesthetics on Product Feature Judgments,” Journal of Consumer Psychology, 20 (4), 419–30.

Homburg, Christian, Martin Schwemmle, and Christina Kuehnl (2015), “New Product Design: Concept, Measurement, and Consequences,” Journal of Marketing, 79 (May), 41–56.

Kaljun, Jasmin and Bojan Dolsak (2011), “Artificial Intelligence in Aesthetic and Ergonomic Product Design Process,” in MIPRO, 2011 Proceedings of the 34th International Convention. Opatija, Croatia: IEEE Conference Publications, 959–64.

Keller, Kevin Lane (1993), “Conceptualizing, Measuring, and Managing Customer-Based Brand Equity,” Journal of Mar- keting, 57 (January), 1–22.

Landwehr, Jan R., Aparna A. Labroo, and Andreas Herrmann (2011), “Gut Liking for the Ordinary: Incorporating Design Fluency Improves Automobile Sales Forecasts,” Marketing Science, 30 (3), 416–29.

Lawson, Bryan (2005), How Designers Think, 4th ed. Burlington, MA: Elsevier.

Lojacono, Gabriella and Gianfranco Zaccai (2004), “The Evolution of the Design-Inspired Enterprise,” MIT Sloan Management Review, 45 (3), 75–79.

Lu, Stephen C.-Y. and Nam-Pyo Suh (2009), “Complexity in Design of Technical Systems,” CIRP Annals: Manufacturing Technology, 58 (1), 157–60.

Martin, Roger L. (2009), The Design of Business: Why Design Thinking Is the Next Competitive Advantage. Boston: Harvard Business Press.

Montini, Laura (2014), “How Better Design Can Get People to Try Your Brand,” Inc., (accessed January 6, 2015), [available at http:// www.inc.com/laura-montini/infographic/the-future-of-experience- design.html].

Moon, Hakil, Jeongdoo Park, and Sangkyun Kim (2015), “The Importance of an Innovative Product Design on Customer Behavior: Development and Validation of a Scale,” Journal of Product Innovation Management, 32 (2), 224–32.

Noble, Charles H. and Minu Kumar (2008), “Using Product Design Strategically to Create Deeper Consumer Connections,” Busi- ness Horizons, 51 (5), 441–50.

——— and ——— (2010), “Exploring the Appeal of Product Design: A Grounded Value-Based Model of Key Design Ele- ments and Relationships,” Journal of Product Innovation Man- agement, 27 (5), 640–57.

Norman, Donald A. (2005), Emotional Design: Why We Love (or Hate) Everyday Things. New York: Basic Books.

Priem, Richard L. (2007), “A Consumer Perspective on Value Creation,” Academy of Management Review, 32 (1), 219–35.

Rindfleisch, Aric and J. Jeffrey Inman (1998), “Explaining the Familiarity-Liking Relationship: Mere Exposure, Information Availability, or Social Desirability?” Marketing Letters, 9 (1), 5–19.

Rindova, Violina P. and Antoaneta P. Petkova (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), 217–32.

Rubera, Gaia (2015), “Design Innovativeness and Product Sales’ Evolution,” Marketing Science, 34 (1), 98–115.

SAS (2008), SAS/STAT 9.2 User’s Guide. Cary, NC: SAS Institute. Srinivasan, Raji, Gary L. Lilien, Arvind Rangaswamy, Gina M.

Pingitore, and D. Seldin (2012), “The Total Product Design Concept and an Application to the Auto Market,” Journal of Product Innovation Management, 29 (S1), 3–20.

Stroup, Walter W. (2013), Generalized Linear Mixed Models. Boca Raton, FL: CRC Press.

Tosi, F. (2012), “Ergonomics and Sustainability in the Design of Everyday Use Products,” Work (Reading, Mass.), 41, 3878–82.

Townsend, Janell D., Mitzi M. Montoya, and Roger J. Calantone (2011), “Form and Function: A Matter of Perspective,” Journal of Product Innovation Management, 28 (3), 374–77.

Verganti, R. (2006), “Innovating Through Design,” Harvard Business Review, 84 (12), 114–22.

Veryzer, Robert W., Jr. (1995), “The Place of Product Design and Aesthetics in Consumer Research,” in Advances in Consumer Research, Vol. 22, Frank R. Kardes and Mita Sujan, eds. Provo, UT: Association for Consumer Research, 641–45.

——— and J. Wesley Hutchinson (1998), “The Influence of Unity and Prototypicality on Aesthetic Responses to New Product Designs,” Journal of Consumer Research, 24 (4), 374–94.

Designed to Succeed / 89

1

WEB APPENDIX

Designed to Succeed: Dimensions of Product Design and Their Impact on Market Share*

Rupinder P. Jindal, Assistant Professor of Marketing, Milgard School of Business, University of Washington Tacoma, 1721 S. Jefferson, Tacoma, WA 98402, (253) 692-5885, [email protected] Kumar R. Sarangee, Associate Professor of Marketing, Department of Marketing, Leavey School of Business, Santa Clara University, 500 El Camino Real, Santa Clara, CA 95053, [email protected] Raj Echambadi, Professor of Business Administration and James F. Towey Faculty Fellow, Department of Business Administration, College of Business, University of Illinois at Urbana-Champaign, 1206 South Sixth Street, Champaign, IL 61820, [email protected] Sangwon Lee, Assistant Professor of Marketing, Department of Marketing, Miller College of Business, Ball State University, 2000 W. University Ave., Muncie, IN 47306, [email protected]

VALIDATION OF MEASURES

Although JD Power uses its industry knowledge and factor analysis to form APEAL and

IQS measures from data on vehicle attributes, we conducted several additional studies to validate

if these attributes indeed belong to the measures to which the company allocates them. First, we

conducted personal interviews with consultants at four well-known global design firms. Second,

we conducted a focus group among eight buyers of new vehicles. Third, we conducted an in-

depth interview with a senior sales manager with over 28 years of auto sales experience across

several manufacturers – both domestic and foreign as well as both non-luxury brands and luxury

brands. There was a high degree of convergence between the field perspectives from these

2

qualitative exercises and the attributes measured by JD Power validating the design measures

used in this paper.

In addition, we designed a questionnaire that contained both the measures from JD Power

used in this study and items from a scale proposed by Moon, Park, and Kim (2015) intended to

measure functional, form, and ergonomic dimensions of design. We collected data from 201

buyers of new vehicles via Amazon Mechanical Turk (MTurk) website. MTurk has become an

increasingly popular medium for researchers to collect data and several studies have shown that

the data obtained via MTurk are reliable and comparable to data obtained from subject pools

(e.g., Goodman, Cryder, and Cheema 2013).

We presented the list of attributes from JD Power and asked respondents to classify these

attributes into one of the three design dimensions. Results of the Chi-square tests available from

the authors supported the categorization of attributes followed by JD Power. Next, we asked

respondents to what extent did the three design dimensions affect their decision to buy the

vehicle. We followed this up with twelve questions – four for each design dimension – from

scales proposed by Moon, Park, and Kim (2015). We constructed a multitrait-multimethod

(MTMM) matrix from the scores we obtained (Campbell and Fiske 1959). As shown in table

A.1, each validity coefficient (0.79, 0.64, and 0.62) is higher than values lying in its column and

row in the same hetero-method block as well as values in the heterotrait-monomethod triangles.

Thus, MTMM results supported the construct validity of the design measures.

3

______________________________________________________________________________

Table A.1 MULTITRAIT-MULTIMETHOD (MTMM) MATRIX FOR CONSTRUCT VALIDITY

Study measuresa Scale measuresb Form Ergonomics Function Form Ergonomics Function

Study measures

Form 1.00c Ergonomics 0.42 1.00c Function 0.35 0.54 1.00c

Scale measures

Form 0.79 0.34 0.31 0.92d Ergonomics 0.38 0.64 0.37 0.45 0.82d Function 0.44 0.48 0.62 0.42 0.53 0.84d

Notes: a Method 1 based on attributes listed by JD Power b Method 2 based on scales from Moon, Kim, and Park (2015) c Correlations d Cronbach’s alpha values

Finally, we tested the predictive power of our model using three hold-out samples. We

randomly chose vehicle models from the original dataset to form each hold-out sample. In the

first case, we randomly set aside almost 20 percent of the observations to form a hold-out sample

which left us with an in-sample composed of the remaining 80 percent of the data. We formed

two more hold-out samples in this manner – one with almost 13 percent of the data as hold-out

and 87 percent of the data as in-sample and another with almost 30 percent of the data as hold-

out and 70 percent of the data as in-sample. We estimated the full model using each in-sample

separately and then used the fixed and random coefficients from each estimation to predict the

market shares in the corresponding hold-out sample. To compare the in-samples and the hold-out

samples we calculated several measures such as the root mean square error (RMSE), mean

absolute deviation (MAD1), and median absolute deviation (MAD2) (Armstrong and Collopy

1992). Overall, the results validate our model specification (see table A.2).

4

______________________________________________________________________________

Table A.2 RESULTS FROM IN-SAMPLES AND HOLD-OUT SAMPLES

Sample RMSE MAD1 MAD2 Full sample (n=937) 0.24 0.14 0.08 In-sample 1 (n=813) 0.24 0.14 0.08 Hold-out sample 1 (n=124) 0.20 0.13 0.07 In-sample 2 (n=745) 0.25 0.15 0.08 Hold-out sample 2 (n=192) 0.20 0.12 0.07 In-sample 3 (n=657) 0.24 0.14 0.08 Hold-out sample 3 (n=280) 0.25 0.16 0.09

a

REFERENCES FOR WEB APPENDIX

Armstrong, J. Scott and Fred Collopy (1992), “Error Measures for Generalizing About

Forecasting Methods: Empirical Comparisons,” International Journal of Forecasting, 8

(June), 69–80.

Campbell, Donald T. and Donald W. Fiske (1959), “Convergent and Discriminant Validation by

the Multitrait-Multimethod Matrix,” Psychological Bulletin, 56 (2), 81–105.

Goodman, Joseph K., Cynthia E. Cryder, and Amar Cheema (2013), “Data Collection in a Flat

World: The Strengths and Weaknesses of Mechanical Turk Samples,” Journal of Behavioral

Decision Making, 26 (3), 213–224.

Moon, Hakil, Jeongdoo Park, and Sangkyun Kim (2015), “The Importance of an Innovative

Product Design on Customer Behavior: Development and Validation of a Scale,” Journal of

Product Innovation Management, 32 (2), 224–232.

Copyright of Journal of Marketing is the property of American Marketing Association and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.