johannes ennings
TRANSCRIPT
Electronic Submission Identification Number : xxxxx
1
Technological Evolution
at the Producer–Consumer Interface Draft 2 (January 5, 2000)
JOHANNES M. PENNINGS
The Wharton School
University of Pennsylvania 2000 Steinberg Hall – Dietrich Hall
Philadelphia, PA 19104 – 6370
phone (215) 898 – 7755 fax (215) 898 – 0401
and
HANN OHL KIM
Australian Graduate School of Management University of New South Wales
Sydney 2052 Australia
phone +61 – 2 – 9931 – 9552
fax +61 – 2 – 9313 – 7279 [email protected]
Electronic Submission Identification Number : 12140
Submitted to Technology and Innovation Management Division,
Academy of Management Meetings Toronto
2000
Electronic Submission Identification Number : xxxxx
2
ABSTRACT
This paper examines the emergence of dominant designs in the tennis racket industry.
Unlike other studies, we consider both the competitive dynamics among firms
comprising the industry, as well as the choice behavior of consumers. We assume that
innovations are usually surrounded by a good deal of causal ambiguity, giving producers
opportunities for shaping consumer choice behavior, but on the other hand according a
greater weight to visible and legitimate actors on the demand side who can alleviate
ambiguity about new products. In short, we very much stress the need to jointly consider
both supply and demand conditions and to view producers and consumers in interaction
with each other. Furthermore, we hold the selection environment in which new and
existing designs come and go, to include both economic (e.g., number of competitors)
and institutional (e.g., regulation) factors.
We distinguish several dominant designs, based on material head, width and
length of racket. Some designs are superior than other ones, based on physical and
consumer ratings, but quality is ambiguous, and even so, better designs do not always
win.
The main thrust of the paper revolves around the degree of professional endorsement of
new racket designs which is associated with the likelihood replication by competitors.
While controlling for other variables, this relationship is indeed empirical. Obtaining
data from trade publications, consumer magazines, and laboratory tests, we examined the
effects of environmental, firm and product characteristics. The window of observation
was 1975-1999, but some variables were missing for some time intervals. Using a
generalized maximum likelihood estimation procedure, the findings suggest that next to
endorsement, advertising and product quality are prominent in accounting for the success
and failure of new product introductions. The implication is that we should not only
focus on firm attributes, its strategy or size of market, but we should also discover what
consumers want and how firms can satisfy their needs with a good product.
Keywords: dominant design, technological trajectory, consumers, tennis racket.
Electronic Submission Identification Number : xxxxx
3
Technological Evolution at the Producer–Consumer Interface
I. INTRODUCTION
Technological evolutions or “technological trajectories” (Dosi, 1982) are usually
attributed to competitive conditions among industry incumbents and new entrants.
Schumpeter (1934) was among the first to have spelled out the conditions that undermine
the status quo in an industry, the improvement or demise of a dominant design and its
replacement by a new one. Dominant designs get exchanged as innovative investments
produce new components and/or new architectures.
In this paper, we explore the emergence of dominant designs in the tennis racket
industry. However, unlike other studies, we consider not only the conduct of
competitors, but include also the choice behavior of consumers. While firms can certainly
change or solidify consumer preferences regarding new products and services, we should
consider those preferences in their own right. We highlight the importance of consumer
preferences in accounting for successful product innovations.
In the following paragraphs, we develop a framework for innovation that centers
on the producer-consumer interface. First, we consider a concern with that interface,
rather than with the more conventional, exclusive focus on producers for understanding
the dominant design emergence. The views surrounding producers’ role in shaping
technological evolutions have been well documented. Next, we explore some aspects
that figure prominently in the interface, most notably advertising as input for legitimizing
innovations by producers and consumers, and examine that interface as a selection
environment that comprises both market and institutional conditions. The results support
Electronic Submission Identification Number : xxxxx
4
our hypothesis that consumer preference management is critical in technological
evolution.
II. A PRODUCER–CONSUMER INTERFACE FRAMEWORK
Whether in traditional industrial economics (Sutton, 1991) or economic history
(Sahal, 1981), it is widely assumed that the dominant position in the industry is an
inevitable consequence of technical efficiency of an ex post successful alternative.
However, technological change towards superiority is not self-evident as technical
efficiency considerations constitute only one of the factors that shape a technology break-
through (e.g., David, 1985; Arthur, 1989; Podolny and Stuart, 1995; Anderson and
Tushman, 1990). A new technology might have multiple, but uncorrelated benefits, or a
mixture of positive and negative outcomes (e.g., Hoch and Ha, 1986). However, research
is typically slanted towards the emergence of new dominant designs, with the benefit (or
drawback!) of hindsight, and little knowledge exists about efforts at establishing new
designs that faltered.
II. 1. Producers
The role of firms in shaping technological trajectories is well documented. The
definition of the market, and the competitive interdependencies among its firms are the
central focus of industrial economics. Markets typically are viewed as firms supplying
goods and services that are close substitutes and thus have high cross-elasticities
(Robinson, 1933). Market boundaries derive from the gaps in demand such that products
within the same market have higher cross elasticities of demand than products belonging
to different markets. The market might be segregated cross-sectionally or inter-
temporally. When innovations produce new products having diminished cross elasticities
Electronic Submission Identification Number : xxxxx
5
with obsolete ones, we observe the emergence of a new market, or convergence from
existing ones, e.g., the video game industry and consumer owned communication
technologies, or TVs and PCs (Greenstein and Khanna, 1997).
It has been suggested that the early and late stages of an industry resemble the
‘entrepreneurial’ and ‘routinized’ regimes, respectively (e.g., Utterback, 1995). This
qualitative shift in technological regimes occurs with the emergence of a dominant
design. When the industry standard has been firmly established, incumbent firms are in a
better position to command a lion’s share of the market because their capabilities in
process innovation confer competitive advantages over entrants. The impact of dominant
designs has been documented in various industries such as typewriter, automobile,
television, and transistor (Utterback and Suarez, 1993).
Many innovations are introduced by new entrants. However, when incumbents
posseses complementary assets, i.e., the specialized supporting resources necessary for
successful commercialization of an innovation, they have competitive advantage over
new entrants introducing an innovation (Teece, 1986). Tripsas (1997) indicates that
incumbents in the typesetting industry, endowed with certain complementary assets (such
as a proprietary font library), might overcome the death of a dominant design to which
they were wedded. These firms redeployed those assets during the arrival of a new
design by attaching them to new architectures. In the tennis racket industry, firms
introducing new designs may be at an advantages as they are already endowed with
complementary assets such as marketing and distribution capabilities.
This “producer” literature thus dwells on the relationship between the nature of
innovation, firm characteristics, and industry dynamics. Not much attention is given to
Electronic Submission Identification Number : xxxxx
6
the process by which certain innovations come to be widely adopted by competitors and
potential users.
The diffusion process plays an important role in technological evolution. It is rare
for an innovator to solely satisfy the aggregate market demand. Furthermore, the
evolutionary path of technological development significantly changes depending on the
extent to which an innovation is imitated by competitors: the decline of Macintosh in
personal computing is an oft-cited example. Therefore, the process of imitation or
diffusion is the mechanism by which new technologies come to acquire economic
significance. According to Nelson and Winter (1982), few studies formally treat the
imitation stage as part of a more general theory of technological change. Technological
diffusion is a process that spreads successful varieties of products and processes through
an economic structure and displaces wholly or partly the existing varieties (Rogers,
1995). Early diffusion studies attempted to find empirical regularities (e.g., Grilliches,
1957; Mansfield, 1961) as illustrated by the well-known S-shaped curve.
II. 2. Consumers
The role of consumers is generally not considered when tracking and accounting
for the emergence, divergence or convergence of technologies. Markets are defined by
the set of producers whose products have high cross-elasticities. The role of consumers at
the end of the value chain and their interaction with upstream firms is disregarded,
although the original meaning of markets goes back to the Middle Ages when buyers and
sellers congregated in the town market, a well bounded physical locale. This imagery is
also prevalent among sociologists (e.g., White, 1981) who define markets as self-
reproducing structures of roles (of buyers and sellers). Today, markets are mostly abstract
Electronic Submission Identification Number : xxxxx
7
and inferred from demand and supply structures. To give adequate status to consumers,
we need to move away from the exclusive focus of firms in shaping those demand/supply
structures. Consumer choice behavior should not be treated as a given. Rather, we could
focus on the buyer-supplier relationships, and view a market as an ongoing network that
evolves around a product or service. Several ingredients for a framework suggest
themselves now. As the term implies, innovation involves novelty and, therefore,
embodies uncertainty. We begin with a review of choice models accommodating
uncertainty. Then we examine related issues including advertising and the selection
environment for new products and services, which bifurcates into “market” and
“institutional” selection environments. Under conditions of uncertainty and limited
rationality, the “clearing” of a product by the market gets complemented by non-market
selection processes.
Behavioral Decision Theory
Unlike classical economists, we cannot assume that consumers are rational--rather
their rationality is bounded due to their limited computational capabilities. Following the
early writings by Simon (e.g., 1955), March (1978) has elaborated on rational decision
making. He gives a qualified endorsement to rational choice models due to the absence of
a clearly defined preference structure: “Human beings have unstable, inconsistent,
incompletely evoked, and imprecise goals at least in part because human abilities limit
preference orderliness” (March, 1978: 598). One reason for these “ambiguous tastes” is
that their limited cognitive capability affects information processing about future
preferences of consequences of current actions just as it affects information processing
about “future consequences of current actions” as Simon (1955) emphasized. Human
Electronic Submission Identification Number : xxxxx
8
beings (and institutions) “intelligently choose to have ambiguous tastes” (March, 1978:
598) partly because they recognize that our tastes are constructed and interpreted through
various experiences. Along this line, he argues that human beings view “inconsistency as
a normal, and necessary, aspect of the development and clarification of tastes” (March,
1978: 599). Compared with established products and the habit formation that has taken
its course, new products provoke a higher level of search and greater intensity of choice
behavior.
The combination of bounded rationality (Simon, 1955; March, 1978) and novelty
of innovative products poses a challenging task to consumers. First, consumers
experience difficulties in getting relevant information on the overall quality of new
products or the benefit of new product features. In many cases, we observe a lack of
consensus over whether a particular innovation improves the performance of the product
and who stands to benefit. In the absence of accurate information about new products we
should not assume that consumers can make rational choices as stipulated in traditional
economic models (Simon, 1955).
Another challenge surfaces around the capacity of new products in meeting
specific needs. Innovations typically add new features to existing products. Innovations,
by definition, introduce products that are new to the market. When innovations are
incremental, consumers deal with a relatively simple problem of comparing cost and
benefit. However, when more radical or discontinuous, we cannot assume the existence
of a well-defined preference structure for judging the presence or absence of an
innovation’s benefits. Since such novel products are foreign to the market, it is unlikely
that consumers possess already well developed preference structures. Consumer
Electronic Submission Identification Number : xxxxx
9
evaluation of new products then inevitably accompanies guesses regarding their future
preferences. Innovation produces two classes of uncertainty, uncertainty about quality
and uncertainty about preferences (Ellsberg (1961). Such uncertainty requires us to view
the acceptance of a new product through a lens that encompasses also non-market
attributes.
Ambiguity, Innovation, and Advertising
As innovations entail ambiguity, and firms are deemed quite influential in shaping
cognitions regarding a novel product’s features, it is of considerable interest to explore
how firms set the tone in a producer-consumer interface. Facing new features and
dimensions of the new product, consumers cannot depend on their prior knowledge
structure for the product and consumption patterns: they rather need to modify previous
cognitive schemas or develop a new set of preferences. During the early stages of a
product’s life cycle, the causal ambiguity awaits diminution, while at the end of the
cycle, consumers have acquired the requisite cognitive categories for evaluating a
product’s features. Product quality evaluation under ambiguity can be more appropriately
represented as a learning process than as a choice problem: consumer evaluation of new
features is a learning process through which they update their knowledge about the nature
of the product (Veryzer, 1998). The process through which consumers assess product
quality is posited as “an intuitive hypothesis-testing process whereby one of the most
important sources for tentative hypotheses is advertising (Hoch and Ha, 1986). While
economists treat advertising as a source of information (Nelson, 1970, 1974;
Schmalensee, 1978), consumers often do not seem to believe what is claimed in
advertising because they recognize the possibility that firms have incentives to lie about
Electronic Submission Identification Number : xxxxx
10
their products. Recent studies in marketing rather treat advertising as a source for
tentative hypotheses (Hoch and Ha, 1986). Consumers take an eclectic approach to
advertising because they use advertising as a source of new information but at the same
time they are skeptical about its validity. So temporary hypotheses emerge in consumers’
minds by advertising and consumers wait to either believe or discard it until they have a
chance to test the product themselves. When confirmed, they accept but when
disconfirmed by experience, they reject the hypotheses.
Then, the next question is how consumers interpret the information from
experience with specific hypotheses in their mind. Human information processing can
take a form of either concept-driven or data-driven (Bobrow and Norman, 1975; Gavetti
and Levinthal, 2000). The former, also called on-line or top-down information
processing, denotes the pattern by which prior expectations and knowledge guide which
information will be selected and how it will be interpreted. That is, people do not pay
equal attention to all information. Rather they put more weight on the information that is
consistent with their prior expectations and beliefs. The information is interpreted as
consistent with their hypothetical relationship.
In contrast, people sometimes adopt the opposite, the data-driven, bottom-up
information processing. In this case, prior expectations do not play a critical role.
Individuals update and revise their prior beliefs based on new information by interpreting
the information at its “face” value: prior expectations do not “interfere” with the
supposedly correct message of the information.
The above argument opens up the possibility of firm intervention into consumer
product evaluation of new products. Since innovations often break the established
Electronic Submission Identification Number : xxxxx
11
perception of the product and its function, they challenge the quality and usefulness of
old versus new products and unravel the extant preferences for making a commensurate
evaluation. Therefore, the management of consumer perception of innovations is critical
for their success. Schumpeter stressed the active role of the entrepreneurial firm by
arguing that “changes in consumers’ tastes are brought about by the action of producers”
(Schumpeter, 1939: 73). Then, the question is “how.” As discussed above, one way that
the firm can actively form a working hypothesis about its innovations in consumer’s
mind is advertising.
The tennis racket industry appears to be highly typical for such a setting. We can
explore that market vis-a-vis other markets, say the market for squash rackets or golf
clubs, but we might also focus on successive markets where product and process
innovations trigger discontinuous markets over time (e.g., Utterback and Abernathy,
1975), say, the market for wooden rackets versus the market for aluminum rackets. As we
will see, the tennis racket industry shows successive dominant designs that became the
industry standard, but also numerous designs that never reached a minimal critical mass
to merit such a status. Nor should we assume that the selection environment in a market
like that of tennis rackets consists largely of the market’s invisible hand, as we often also
detect the presence of visible hands, most notably governmental, legal, regulatory and
other institutional forces.
II. 3. Market versus Institutional Selection Forces
For analytical purposes we might distinguish between market and institutional
forces as comprising a novel product’s selection environment. It is plausible to argue
that a product’s causal ambiguity might give comparatively more weight to institutional
Electronic Submission Identification Number : xxxxx
12
forces. We find evidence for such pronouncements in complex and simple technologies
that vie for the dominant design--as illustrated by NTSC and Qwerty standards
respectively. For example, the Qwerty keyboard might be ergonomically awkward, but
unlikely to entangle some of the characters’ hammers and thus not slow down the typing
task. The NTSC television standard became the US standard because RCA was well
connected with the Federal Communication Commission, a regulatory agency that could
more readily be manipulated by RCA and not by RCA’s competitors which lacked
RCA’s social capital (Collis, Bane and Bradley, 1997). The NTSC adoption was not
exclusively predicated on its picture video quality as any international traveler can attest.
Summarizing, when causal ambiguity prevails, producers might signal superiority,
thus shaping the technological evolution, for example through advertising, demonstration
projects and other modes of signaling. Producers enjoy considerable leverage in
manipulating the debate around products, especially ambiguous ones. Many markets are
socially constructed and reproduced. This follows not only from the above statement that
markets and their products comprise self-regulating systems of roles that are observable
at the producer-consumer interface. Social construction is most salient when we
encounter high levels of uncertainty, and unequivocal reference points are lacking.
III. THE TENNIS RACKET INDUSTRY
The sporting goods industry in general, and the tennis racket industry in
particular, provides an attractive opportunity for studying technological innovation.
While some sporting goods have been immune to change – most notably the baseball bat
and the field hockey stick, others have shown dramatic changes – golf clubs and tennis
rackets among them the most visible ones. The study of the tennis racket should lend
Electronic Submission Identification Number : xxxxx
13
itself to an examination of evolving dominant designs and is very suitable to deal with
the above issues.
The tennis racket is a relatively simple product with a rather uncomplicated
architecture and transparent product components. Its technological development is easily
discernible with several distinct changes that permit a rather unambiguous punctuation in
developmental stages. The development into different stages also supports improvement
in product qualities, although as we will see, such improvements are shrouded by
uncertainty. Apparently, those improvements have not always produced superior
performance among athletes, using the new artifacts. Thus, there remains ambiguity
regarding product improvements.
Ironically, the institutional context has also produced considerable resistance to
innovation and has been reluctant to embrace new product designs, however evident or
erroneous the product enhancements seem to be. New equipment might enable players to
generate more power with equivalent amounts of force expended, reduce variability in
accuracy with the same skill level, and reduce various safety concerns without tampering
with the sport’s integrity. Or it might trigger compassionate debates about the veracity of
product improvement enhancement, the advertising claims by presumably deceitful firms,
and the bribing of influential opinion leaders. In any event, the greater than 1000 product
introductions and a correspondingly large number of patents granted, renders a
comparison with fruit flies appropriate: this setting provides large numbers of relatively
simple observations, and, therefore, is ideal for examining supply and demand conditions
that destroy and embellish successive dominant designs.
III. 1. The Tennis Racket’s Evolution
Electronic Submission Identification Number : xxxxx
14
The tennis racket has quite a long and rich history. While the oldest racket-shape
instrument documented in the literature goes back to the 16th century, it is rather similar
to the modern racquetball instrument in that it is about seven to ten inches long. The
modern tennis racket, which established the oval face and 68 cm (27-inch)-long frame
standard, was developed in 1873. Despite the long history, significant technological
discontinuities began to take place since 1967 when the first metal tennis racket was
introduced. Figure 1 illustrates the concentration of innovative activities in the past three
decades. The number of patents issued by the U.S. Patent and Trademark Office in the
class 273/73(Amusement Devices: Games/name) and D21/212 (Games, Toys, and
Sporting Goods/name) steeply increased in the early 1970s. The number of patents
containing keywords "tennis racket" or "tennis racquet" also shows a similar pattern.
[ Insert Figure 1 about here ]
Improvements in materials and its processing technology have played a
significant role in the advancement of racket development. New materials such as steel,
aluminum, fiberglass, graphite, boron, and titanium gradually replaced wood because the
former is stiffer and more durable than the latter. Interesting developments also took
place in racket design, and some of them were possible due to advanced materials.
Among others, three innovations stand out.
The first is the oversize racket in 1976 developed by Prince, a ball machine
company at the time. It has the face of 784 cm2 (110 square inches) compared to the 460
cm2 (70 square inch) face of the traditional racket. While it might appear trivial, the
over-size idea is regarded as the most revolutionary in the industry's history because it
broke the mind of designers free from the tradition. The second is the wide-body racket
Electronic Submission Identification Number : xxxxx
15
of Wilson in 1987. Its frame has a 39mm-thick cross section beam at the thickest part,
while the traditional racket has a constant beam of only 19mm. The last is the extra-
length racket, which ranges from 71cm (28 inches) to 81cm (32 inches). The length of
the racket had remained at 68cm (27 inches) since the late 19th century until the first
commercially successful 71cm (28-inch) racket was introduced in 1995. These three
design innovations along with the development in materials have produced a series of
changes in dominant designs in the industry (Figure 2).
[ Insert Figure 2 about here ]
In addition to these three dominant technologies, numerous new technologies
have been introduced that achieved only moderate success. Pro Kennex developed a
racket that has about 100 micro vials in the racket frame which adds extra momentum to
the ball and reduces the transmission of racket shock and vibration (Ashley, 1993). A
number of companies have developed rackets with a string pattern other than the
traditional two-way horizontal/perpendicular string pattern. A two-way diagonal by
Victor Sports and Völkl, and a three-way string pattern by MadRaq are notable. These
rackets are claimed to enlarge the sweet spot. Snauwaert, BioSports, and Sentra
developed ergonomically designed rackets that place the sweet spot differently and
reduce stress on the wrist of players. Figure 3 illustrates a few of those unusual
innovations.
[ Insert Figure 3 about here ]
III. 2. Ambiguity
The success of an innovation – its imitation by competitors and its acceptance by
consumers – is implicitly regarded as an inevitable consequence of its technical
Electronic Submission Identification Number : xxxxx
16
superiority. It is widely assumed that consumers as well as producers are capable of
perceiving benefits of an innovation, and technological evolution follows its own logic.
However, economic agents of both sides of the market experience a significant amount
of uncertainty and ambiguity when new technologies and products are introduced to the
market.
There are two types of ambiguity that render the detection of an innovation
benefit tenuous. The first results from the fact that the benefit of an innovation is often
not evident. For example, Ellsberg (1961) claimed that innovations pose a challenging
task to decision-makers due to the lack of relevant information. Presumably, experts are
better equipped to form appropriate judgments, yet the heuristics they employ preclude
the suspension of decision biases (Kahneman and Tversky, 1979). While the oversize
racket is now unanimously praised as the most important innovation in the history of the
tennis racket industry, it faced resistance even from industry experts. In a review article,
an industry expert intimated by commenting that,
“After playing the Prince racket for a while, I must admit that I was a bit disappointed. The racket showed very little torsional rigidity, excessive vibration on off-center hits, and a marked tendency to "spoon" or bend at the head of the racket. Innovation in any sport is healthy, but I feel the Prince is more of a gimmick, not a true step forward in high performance racket technology. It is best used by beginners or novices.” (From Fiott, 1978: 66, italics added)
Our expert regards the benefit of an oversize racket as a simple psychological advantage.
However, it turned out that an oversize racket has a definitive advantage over a
traditional size racket because the former enlarges the sweet spot on the racket (Brody,
1979).
Electronic Submission Identification Number : xxxxx
17
The story behind a recent ITF (International Tennis Federation) ruling of banning
the racket longer than 74cm (29 inches) also illustrates the ambiguity and uncertainty
caused by an innovation. The ITF, the rule-governing body of tennis, changed the
maximum length of the racket from 81cm (32 inches) to 74cm (29 inches) because it felt
that high speed, powerful serves produced by extra-length rackets would diminish the
popularity of tennis as a spectator as well as a participant sport. However, an experiment
conducted by Tennis magazine shows that the real benefit of extra lengths comes from a
higher accuracy rather than increased power and speed (Table 2).
[ Insert Table 2 about here ]
The second type of ambiguity results from a poor or moderate correlation
between product attributes. A low correlation generates considerable discretion for
consumers and induces uncertainty about optimum choices (Hoch and Ha, 1986). In the
case of tennis rackets, three product attributes – stiffness, stability, and power – are
employed as representative characteristics. Table 3 provides the correlations between the
three attributes measured by instruments and a panel of players. It shows that the three
attributes have only moderate to low correlations. Furthermore, a low inter-subject
correlation indicates that it is difficult for players to agree on the characteristics of a
particular racket.
[ Insert Table 3 about here ]
III. 3. Producer-Consumer Interface through Professional Endorsement
When ambiguity prevails, the sense making activity becomes critical in
determining the path of technological development. It has been widely documented that
individuals tend to stay with a current option when the benefit of a new alternative is not
Electronic Submission Identification Number : xxxxx
18
clear (Kahneman and Tversky, 1979; Thaler, 1980). Samuelson and Zeckhauser (1988)
show that people weigh potential losses from switching from the product to a new one
larger than the same amount of potential gains from a new product: they call it the "status
quo" bias. This phenomenon creates a barrier that entrepreneurs should overcome
through various activities. According to Schumpeter,
“...the great majority of changes in commodities consumed has been forced by producers on consumers who, more often than not, have resisted the change and have had to be educated up by elaborate psychotechnics of advertising.” (From Schumpeter, 1939: 73)
This Schumpeterean observation point has an important implication in the present
argument. Given the uncorrelated relationship among performance attributes, the
superiority of any tennis racket is equivocal. When its benefits are elusive, consumers
are reluctant to switch. This tendency is reinforced due to the conservatism in the tennis
community. An innovator's commercialization efforts take on extra significance. While
development of technically superior products is important, it is the beginning of the
entrepreneurial process rather than its end.
Racket manufacturers resort to professional tennis players in providing a strong
message to consumers about their products. Having top players use the product, as they
are highly visible and respected in the tennis community, can increase the legitimacy of a
new product. Also, the confusion regarding the performance of the racket can be
significantly reduced when high status individuals perform well with a new racket.
When the first oversize racket was introduced, it was criticized by industry experts as a
gimmick (Fiott, 1978). However, negative responses against the racket subsided when
Pam Shriver, a female professional player, surprised the tennis community by advancing
to the final of the U.S. Open in 1978 playing with the Prince oversize racket. Her success
Electronic Submission Identification Number : xxxxx
19
with a new design signaled innovation benefits about this racket because she had never
reached even the quarterfinals of women's singles championships in any Grand Slam
event before. She provided pseudo-proof that the oversize racket is not just for
"beginners or novices" (Fiott, 1978: 66) who need extra help from the equipment but
rather enhances the performance level for players at any level (Le Marche, 1986).
In short, by making professional players use their product, innovators affect the
acceptance of a new technology because professional players' endorsement provides
familiarity and legitimacy (Veryzer, 1998; Ohanian, 1991). Even in the absence of
induced usage, when players voluntarily pick their own design, the legitimization of that
design might resolve ambiguity, and thus produce a trigger toward ramp-up in its
diffusion. The following quote is from an industry expert who accords with our
argument:
“...but if the past is any guide, successful racket design concepts will emerge from the ranks of tennis players and enthusiasts. "Familiarity seems to lead to conformity in this business," Leonard said.” (From Ashley, 1993: 55)
Based on these arguments, we propose the following hypothesis:
Hypothesis The top professional tennis players’ endorsement of the racket with a focal racket design is positively associated with the likelihood of replication of the design, after controlling the effects of other variables.
Our concern with demand side conditions does not rule out the endogenous
aspects of producers in shaping the rise and fall of dominant designs. We might consider
the conventional factors, including the size of the firm (as a proxy for tangible resource
endowments), R&D intensity, and intangible assets that are complementary with any
Electronic Submission Identification Number : xxxxx
20
new product introduction. We should also heed the importance of institutional conditions,
most notably the role of regulation in legitimizing certain innovations.
IV. METHOD
IV. 1. Empirical Models
Available data allow us to test the above hypothesis and are expected to help in
resolving the central issue of the paper: we need to consider both producer and consumer
aspects since consumers matter too in the rise and fall of dominant designs. To explore
the hypothesis, Poisson regression was first considered. While Poisson regression is
widely considered as the base model for the analysis of count data, adjustments are often
required because assumptions of the model are not always satisfied. In particular, when
the assumption that the mean and the variance are equal (E(y) = Var(y) = λ) is violated,
the following problems arise even if the coefficients are not biased (Allison, 1999):
underestimation of the standard errors, overestimation of the chi-square statistics, and
inefficiency of the maximum likelihood estimation. The phenomenon is often referred to
overdispersion. In this study, the variance (0.243) exceeds the mean (0.06) so that we
need to make the necessary adjustments. The overdispersion problem is amended by
including a gamma-distributed error term in the equation (Cameron and Trivedi, 1986;
Hausman, Hall, and Griliches, 1984):
log λi = β0 + β1χi1 + β2χi2 + …+ βkχik + σεi. [1]
Here, the dependent variable Yi is assumed to have a Poisson distribution with expected
value λi, conditional on εi, and the unconditional distribution of yi has a negative
binomial distribution (Allison, 1999).
Electronic Submission Identification Number : xxxxx
21
While the negative binomial takes into account the overdispersion problem, it
does not address the correlation among observations (Barron, 1992). In this study, each
year has multiple observations depending on the racket design considered. For example,
the year of 1990 appears in the dataset multiple times for the counts of new rackets with
oversize design, widebody design, and longbody design, to just name a few. Since some
of the variables such as the market and institutional characteristics are correlated among
these observations in the same year, the correlation should be accounted for in the
estimation. Due to the assumption of independence among observations, the maximum
likelihood estimation is not appropriate for this study. Instead, we employ the method of
generalized estimating equations (GEE, hereafter) which addresses the correlation among
observations (Allison, 1999; Diggle, Liang, and Zeger, 1994). In the maximum
likelihood estimation, the iteratively reweighted least squares algorithm is used where
observations are treated statistically independent. It means that a weight matrix W has
off-diagonal elements equal to 0 and diagonal elements equal to pi(1 – pi), where pi is the
predicted probability from the previous iteration. In the GEE procedure, however, off-
diagonal elements of the matrix W are allowed to have non-zero values, which are
functions of the correlations among the observations. The GEE estimation can be
obtained by SAS 6.12 through the so called GENMOD procedure with REPEATED
statement (Allison, 1999).
The model employed to analyze adoption of new design had the general
specification of the following:
Yi (t) = αi (t) + β1Χi,1(t – 1) + β2Χi,2 + β3Χi,3 (t – 1) + σεI [2]
Electronic Submission Identification Number : xxxxx
22
where Yi refers to the number of rackets adopting new design, and Χ1, Χ2, and
Χ3, are vectors of variables, description of which follows below.
Χ1 denotes a vector of variables representing the environmental conditions at time
(t – 1). First, we include the size of the tennis playing population because it determines
the carrying capacity of the environment. Second, the number of firms in the market, the
number of rackets in the market, and their square terms are included. Furthermore, major
regulatory changes are included as dummy variables since it is well documented that they
have crucial impacts on the nature of innovative activities (Garud and Rappa, 1994). In
the tennis racket industry, the International Tennis Federation’s ruling on equipments
such as rackets and balls is the major legal barrier for new innovations. Therefore, we
create dummy variables for the years of 1980 and 1997 in which ITF made major
changes in the section 4 (Rackets) in Rules of Tennis.
X2 is a matrix of variables representing the characteristics of the innovating firm.
Previous studies in economics and sociology have primarily focused on the
characteristics of the firm that is more (less) likely to innovate. In order to see whether
the firm’s effort to actively persuade consumers to favor its innovation is robust even
after we take the firm characteristics into consideration, we include the following in the
model: (1) whether the innovating firm is a new entrant or an incumbent firm, (2)
whether the firm has other related businesses, and (3) whether the firm is a U.S. or a
foreign entity.
X3 includes the number of top professional players who endorse and use a tennis
racket with a new technology of interest, the quality index of the racket with a technology
of interest, the amount of advertising in popular magazines, and their interaction terms.
Electronic Submission Identification Number : xxxxx
23
As we discussed above, when a top professional player uses a racket with a new
technology, she not only helps consumers form a tentative hypothesis regarding the
benefit of the new technology, but also enhances its legitimacy. However, it can be
argued that a new technology is replicated due to its technical superiority. In order to
rule out this argument, we include a measure of the quality of a tennis racket with an
innovative technology.
IV. 2. Data
Information on the introduction of new tennis rackets is available from popular
magazines such as World Tennis and Tennis. Two reasons make them an attractive
source of information. First, manufacturers run advertisements when they introduce new
tennis rackets that contain detailed information about unique features of the racket.
Second, the magazines have sections in which experts review the racket currently
available in the market. From various issues of these magazines, we identified 1,292 new
tennis rackets during the period between 1965 and 1995.
The Sporting Goods Directory published by the Sporting Goods Dealers
Association contains the name of the company and its products. This information
enables us to construct variables on firm characteristics such as age and environmental
characteristics like the number of firms. Dummy variables for regulatory changes were
constructed by referring to various issues of Rules of Tennis published by the
International Tennis Federation. The market size is measured by the tennis playing
population and collected from the publication by Sports Business Research Network.
Electronic Submission Identification Number : xxxxx
24
IV. 3. Measurement of the Variables
Table 1 provides the description of the variables used in this study. In the
following paragraphs we will first describe the dependent variable followed by the
independent variables. The independent variables include attributes of the market, the
manufacturers, the product design and the product. A very important category involves
the quality of the racket.
Dependent Variable
In our hypothesis, the dependent variable was racket design adoption by
competing firms. Therefore, every single new tennis racket introduced to the U.S. market
was the subject of our study. The question was whether the racket embodied a certain
racket design. The measurement of racket design adoption was straightforward because
the specification of new rackets was readily available in various sources. The dependent
variable was coded 1 when the racket embodied focal racket design and 0 otherwise.1
Eight racket designs were considered in the analysis: (1) oversized head, (2) extra length,
(3) tension adjust device, (4) ergonomic design, (5) three directional string design, (6)
convex throat design, (7) fan string pattern, and (8) widebody design.2
Independent Variables
Professional Endorsement Whether a new racket was used by top professional players
was derived from various issues in popular tennis magazines such as World Tennis,
Tennis, and Tennis Industry. We define top professional players as those who were
included in the top 10 players of the year list published by the Association of Tennis
Professionals (ATP) for male players and the Women’s Tennis Association (WTA) for
female players, respectively. In addition to the top 10 players both in men’s and
Electronic Submission Identification Number : xxxxx
25
women’s annual rankings, we also included those players who reached semi-finals at four
Grand Slam tournaments (Wimbledon, U.S. Open, French Open, and Australian Open)
because they attract a significant amount of attention from the tennis community even
though they were not top ten players. This is legitimate because the ranking of players is
only a proxy for those players who are highly visible in the tennis community.
Advertising As a proxy for advertising spending, we counted the number of advertising
pages in Tennis. We collected the information from August and December issues of the
magazine each year. Those issues were chosen because it was expected for racket
manufacturers to concentrate their promotion efforts in these two months: August issues
run the U.S. Open special coverage and December issues include an annual review of the
industry. When a full one page advertisement was dedicated for one tennis racket, we
counted it as 1. When a multiple number of rackets appeared in a one page
advertisement, we gave each racket (1/the number of rackets on the page). When a
manufacturer ran a two page advertisement for one racket, we counted two for the racket.
By the same token, 0.5 was given to the case in which an advertisement only occupied a
half page. When racket design adoption was the main concern, we summed the number
of advertising pages promoting a specific racket design. For example, not only Prince,
the original innovator of the oversize racket, but also followers promoted their oversize
rackets in advertising. Since the advertising by followers contributed to the increase in
familiarity and acceptability of the oversize design, we collapsed each manufacturer’s
advertising count into one for the overall oversize design. We noticed that the amount of
advertising varied across years. In order to control this variation, we divided the count of
advertising for each racket and design by the total number of advertising pages per year.
Electronic Submission Identification Number : xxxxx
26
Design Quality Quality of racket design poses a critical and challenging task to this
study. It would be ideal if the underlying technological merit of individual racket design
could be objectively measured regardless of its application. However, such information
is not available. Instead, we collected data on performance of rackets, rather than
designs, from a series of racket reviews published in World Tennis. It is believed that
this measure could be used as a substitute for individual design quality because the
contribution of a particular design to racket quality should be reflected on overall racket
quality. If a design truly improves playing characteristics of the racket, overall quality of
the rackets with the focal design should be higher than that of those without the focal
design. Therefore, we compared quality of the racket with focal design to that of rackets
without it, and used this relative measure of racket quality as a quality measure of the
racket design.
World Tennis published a series of articles on racket quality based on extensive
tests over fifteen years. The tests consisted of two very different parts: one measured by
scientific instruments in the lab and the other measured by players on the tennis court.
The lab test was conducted by three scientists at MIT. They measured three dimensions
of the tennis racket that are believed to critically influence its quality.3 The numbers
reported in the magazine were the average numbers given by multiple testers, who vary
in gender and playing style. In addition to the three dimensions, the item called
“playability” measured overall quality of the racket.
While these quality measures were used for individual rackets when individual
racket sales were concerned, we constructed the following variable when new design
adoption was of interest. When the success of racket design is an issue, what matters is
Electronic Submission Identification Number : xxxxx
27
how good a racket with focal design could be in comparison to quality of rackets without
the design. In other words, since what we need to measure is the frontier that could be
achieved with a particular racket design, we chose the racket with the highest quality
index from a group of rackets with focal design as the number representing racket design
quality that could be accomplished at the time. Furthermore, we standardized the quality
index with the mean and variance of quality of rackets without focal design. Therefore,
best quality index for racket design i was calculated as the following:4
).(.
)(
2~,
2~,,,
−≠
−≠−=
ttij
ttijtiti QualityDevStd
QualityAverageyBestQualityIndexBestQualit [3]
where j=1, 2,…, i, …, k racket designs.
The results reported in the following are based on the comparison with the rackets
without focal design introduced in the past two years from the year of interest since three
years, rather than one, is considered reasonable as a window of competition rather than
one year.5
V. RESULTS
Table 4 reports the results on new racket introduction data during the period from
1981 to 1992. The effects of environmental characteristics and innovator characteristics
on the dependent variable were first estimated (Models 1 and 2). Of particular interest is
whether the innovator is a new entrant: a new racket design developed by an incumbent
firm is more likely to be replicated.
[ Insert Table 4 about here ]
In addition to the variables included in Models 1 and 2, we added design related
variables (Models 3 and 4). First, when interaction was not considered (Model 3), only
Electronic Submission Identification Number : xxxxx
28
professional endorsement had positive effects on the design replication: neither design
quality nor advertising was significant. However, when an interaction term was included
(Model 4), all three main effect variables, i.e., professional endorsement, design quality,
and advertising, turned out to be significant. In addition, the interaction between
professional endorsement and design quality was also significant. The interaction
between professional endorsement and advertising was not reported in Model 4 since it
was not significant and did not change the effect of any other variable.
VI. IMPLICATION
This study has an implication for the selection environment within evolutionary
perspectives. Facing the anomalies of an arguably inferior technology’s dominance in
the market (e.g, David, 1985), researchers have attempted to uncover the selection
environment, which has been largely treated as a black box. While some attribute the
seemingly contradictory outcome to “historically small events” (Arthur, 1989) in the
presence of positive feedback effects, others emphasize more systematic issues such as
the role of the institutional environment (e.g., Garud and Rappa, 1994).
These studies, however, keep silent on one of the key elements of the selection
environment: consumers. Recent studies begin to recognize the importance of consumers
as a critical component of the selection environment and attempt to bring demand side
issues into discussion. Cowan, Cowan, and Swann (1997) argue that the market as we
know it properly functions only when consumption as well as production side evolve
together. What really separates these studies from traditional approaches is that
consumers are not assumed to have fixed preferences. They lament on the assumption of
exogenously given preference which has precluded a serious discussion of how consumer
Electronic Submission Identification Number : xxxxx
29
attitudes change over time. This perspective corresponds to the recent trend in decision
analysis and marketing. There, preferences (or goals) are treated as being constructed by
making choices rather than choices which are made on the basis of existing preferences
(e.g., March, 1978; Devetag, 1999; Payne, Bettman, and Johnson, 1992; Carpenter and
Nakamoto, 1989).
One venue that firms can take to actively change consumer preferences in their
favor is advertising. Galbraith (1958, 1988) argued that the role of advertising becomes
very important in the industrialized economy since the basic needs which inspired the
traditional economic thinking are already fulfilled. In fact, Schumpeter (1934) also noted
the constructive nature of preferences. In the discussion of the entrepreneur, he said:
“Yet innovations in the economic system do not as a rule take place in such a way that first new wants arise spontaneously in consumers and then the productive apparatus swings round through their pressure. We do not deny the presence of this nexus. It is, however, the producer who as a rule initiates economic change, and consumers are educated by him if necessary; they are taught, as it were, to want new things…” (p. 65)
This study shows that, in addition to good quality, producers also have to carefully
consider how to deliver their innovative products to consumers. Professional
endorsement is one way to significantly increase consumers’ confidence about the novel
products. Its effects cascades into high levels of replication of the design and exploitation
of the conceptual basis of the idea behind the innovation. Advertising constitutes another
promotion strategy concept for according consumers a greater role in the legitimization
of new racket designs.
VII. CONCLUSION
This paper suggests us to consider demand factors such as star endorsement, and
opinion- leader behavior as a key factor in accounting for the emergence of dominant
designs.
Electronic Submission Identification Number : xxxxx
30
Compared to the existing literature on technological evolution, this study,
therefore, takes a more balanced view in that it brings unsuccessful innovations under
scrutiny. By doing so, the study not only presents technological evolution in the tennis
racket industry closer to the reality, but also provides a setting in which we can
illuminate some issues that improve our understanding on the process of technological
development. This study describes the cases where some racket designs became extinct
even though their quality was at least as good as successful ones. Empirical analyses
suggest that, in addition to innovation quality, such promotional strategies as professional
endorsement significantly improves the chance of an innovation’s success.
It is surprising to note that only one side of the market, the producer side, has
been extensively explored in previous studies. Figure 4 reiterates the need to jointly
consider both supply and demand conditions, and to view producers and consumers in
interaction with each other. That interaction conditions the success and failure of
innovations and the route that technological trajectories undertake.
[ Insert Figure 4 about here ]
While it is not explicitly stated, it is assumed in previous studies that the benefit
of an innovation is self-evident to consumers and they know what they want. So the
main target for the firm is to discover what they want and satisfy it with a good product.
However, as emphasized by Galbraith (1958), Schumpeter (1939), and March (1978),
innovators should exert a significant amount of effort to persuade consumers by
providing relevant information through various means, because the success of an
innovation emerges out of consensus among producers and consumers. The punch line of
Electronic Submission Identification Number : xxxxx
31
this paper, therefore, is an invitation to adopt a more fully specified model that
incorporates also consumer considerations in the rise and fall of dominant designs.
It should be pointed out that the tennis racket industry and the sporting goods
industry, in general, have some peculiarities that are at variance with other industries.
The high visibility of professional players as a way to highlight consumer aspects and
technological innovation provides an ideal setting for this study. In addition, since the
conservative nature of the sport makes it more difficult for entrepreneurs to introduce
new ideas, the importance of entrepreneurial skills other than technological ones is well
highlighted. However, the industry seems very vulnerable to the ebb and flow of fads in
products rather than to substantive changes in technology. While recent studies suggest
that the endorsement by and the relationship with prominent players in the industry have
significant effects on the fate of patents and projects (e.g., Podolny and Stuart, 1995), we
also need to find out the limits under which endorsement effects effectively work.
Electronic Submission Identification Number : xxxxx
32
References
Allison, P. 1999. Logistic Regression Using the SAS System: Theory and Application. Cary, North Carolina: SAS Institute Inc.
Anderson, P., and Tushman, M. L. 1990. Technological discontinuities and dominant designs: A cyclical model of technological change. Administrative Science Quarterly, vol. 35, pp. 604 – 633.
Arthur, W. B. 1989. Competing technologies, increasing returns, and lock-in by historical events. Economic Journal, vol. 99, pp. 116 – 131.
Ashley, S. 1993. High-tech rackets hold court. Mechanical Engineering, August, pp. 80 – 81.
________. 1995. A new racket shakes up tennis. Mechanical Engineering, August, pp. 80 – 81.
Barron, D. N., and Hannan, M. T. 1991. Autocorrelation and density dependence in organizational founding rates. Sociological Methods and Research, vol. 20, pp. 218 – 241.
Bobrow, D. G., and Norman, D. A. 1975. Some principles of memory schemata. In Representation and Understanding: Studies in Cognitive Science, edited by Bobrow, D. G., and Collins, A. New York, NY: McGraw-Hill.
Bower, G. H., and Hilgard, E. R. 1981. Theories of Learning. Englewood Cliffs, NJ: Prentice-Hall.
Brody, H. 1979. Physics of the tennis racket. American Journal of Physics, vol. 47, pp. 482 – 487.
Burt, R. 1992. Structural Holes. Cambridge, MA: Harvard University Press. Cameron, A. C., and Trivedi, P. K. 1986. Econometric models based on count data:
Comparisons and applications of some estimators and tests. Journal of Applied Econometrics, vol. 1, pp. 29 – 53.
Carpenter, G., and Nakamoto, K. 1989. Consumer preference formation and pioneering advantage. Journal of Marketing Research, vol. 26, pp. 285 – 298.
Chow, G. C. 1967. Technological change and the demand for computers. American Economic Review, vol. 57, pp. 1117 – 1130.
Collis, D., Bane, P. W., and Bradley, S. 1997. Winners and losers: Industry structure in the converging world of telecommunications, computing, and entertainment. In Competing in the Age of Digital Convergence edited by David B. Yoffie, pp. 159 – 200. Boston, MA: Harvard Business School Press.
Cowan, R. 1991. Nuclear power reactors: A study in technological lock-in. Journal of Economic History, vol. 5, pp. 541 – 567.
Cowan, R., Cowan, W., and Swann, P. 1997. A model of demand with interactions among consumers. International Journal of Industrial Organization, vol. 15, pp. 711 – 732.
David, P. 1985. Clio and the economics of QWERTY. American Economic Review, vol. 75, pp. 332 – 337.
Electronic Submission Identification Number : xxxxx
33
Devetag, M. 1999. From utilities to mental models: A critical survey on decision rules and cognition in consumer choice. Industrial and Corporate Change, vol. 8, pp. 289 – 351.
Diggle, P., Liang, K.-Y., and Zeger, S. 1994. Analysis of Longitudinal Data. New York, NY: Oxford University Press.
Dosi, G. 1982. Technological paradigms and technological trajectories. Research Policy, vol. 11, pp. 147 – 162.
Ellsberg, D. 1961. Risk, ambiguity, and the Savage axioms. Quarterly Journal of Economics, vol. 75, pp. 643 – 669.
Fiott, S. 1978. Tennis Equipment. Boston, MA: Tennis Research Group. Galbraith, J. K. 1958. The Affluent Society. Boston, MA: Houghton Mifflin Company. ____________. 1988. Economics and advertising: Exercise in denial. Advertising Age,
November 9, pp. 81 – 84. Garud, R., and Rappa, M. 1994. A socio-cognitive model of technology evolution: The
case of cochlear implants. Organization Science, vol. 5, pp. 344 – 362. Gavetti, G and Levinthal, D.A.. 2000. Gavetti, G., Levinthal, D. (1999), “Looking
Forward and Looking Backward: Cognitive and Experiential Search”, Administrative Science Quarterly (forthcoming).
Greenstein, S., and Khanna, T. 1997. What does industry convergence mean? In Competing in the Age of Digital Convergence edited by David B. Yoffie, pp. 201 – 226. Boston, MA: Harvard Business School Press.
Grilliches Z. 1957. Hybrid corn: An exploration in the economics of technological change. Econometrica, vol. 25, pp. 501 – 522.
Hausman, J. Hall, B., and Griliches, Z. 1984. Econometric models for count data with an application to the patents – R&D relationship. Econometrica, vol. 52, pp. 909 – 938.
Hoch, S. J., and Ha, Y-W. 1986. Consumer learning: Advertising and the ambiguity of product experience. Journal of Consumer Research, vol. 13, pp. 221 – 233.
Kahneman, D., and Tversky, A. 1979. Prospect theory: An analysis of decision under risk. Econometrica, vol. 47, pp. 263 – 291.
Le Marche, R. 1986. How the Prince patent has changed the game. Tennis, vol. 21, April, pp. 106 – 116.
Mansfield, E. 1961. Technical change and the rate of imitation. Econometrica, vol. 29, pp. 741 – 766.
March, J. G. 1978. Bounded rationality, ambiguity, and the engineering of choice. Bell Journal of Economics, vol. 9, pp. 587 – 608.
Mitchell, W. 1989. Whether and when?: Probability and timing of incumbents’ entry into emerging industrial subfields. Administrative Science Quarterly, vol. 34, pp. 208 – 230.
Nelson, P. 1970. Advertising as information. Journal of Political Economy, vol. X, pp. 729 – 754.
Nelson, P. 1974. Information and consumer behavior. Journal of Political Economy, vol. X, pp. 311 – 329.
Nelson, R. R. and Winter, S. G. 1982. An Evolutionary Theory of Economic Change. Cambridge, MA: The Belknap Press of Harvard University Press.
Electronic Submission Identification Number : xxxxx
34
Ohanian, R. 1991. The impact of celebrity spokespersons’ perceived image on consumers’ intention to purchase. Journal of Advertising Research, vol. X pp. 46 – 54.
Payne, J., Betman, J., and Johnson, E. 1992. Behavioral decision research: A constructive processing perspective. Annual Review of Psychology, vol. 43, pp. 87 – 131.
Pinch, T., and Bijker, W. 1987. The social construction of facts and artifacts: Or how the sociology of science and the sociology of technology might benefit each other. In The Social Construction of Technological Systems edited by W. Bijker, T. Hughes, and T. Pinch, pp. 17 – 50. Cambridge, MA: MIT Press.
Podolny, J., and Stuart, T. 1995. A role-based ecology of technological change. American Journal of Sociology, vol. 100, pp. 1224 – 1260.
Porac, J. 1998. How do we know that a market exist?: Identities and equivocality in market sensemaking. Working paper, University of Illinois at Urbana Champaign.
Robinson, J. 1933. The Economics of Imperfect Competition. London: Macmillan. Rogers, E. 1995. Diffusion of Innovations. 4th ed. New York, NY: Free Press. Sahal, D. 1981. Patterns of Technological Innovation. Reading, MA: Addison-Wesley. Samuelson, W., and Zeckhauser, R. 1988. Status quo bias in decision making. Journal of
Risk and Uncertainty, vol. 1, pp. 7 – 59. Schmalensee, R. 1978. A model of advertising and product quality. Journal of Political
Economy, vol. 86, pp. 485 – 503. Schumpeter, J. A. 1934. The Theory of Economic Development. Cambridge, MA:
Harvard University Press. _____________. 1939. Business cycles: A theoretical, historical, and statistical analysis
of the capitalist process. New York, NY: McGraw-Hill Book Company, Inc. Simon, H. A. 1955. A behavioral model of rational choice. Quarterly Journal of
Economics, vol. 69, pp. 99 – 118. Sutton, J. 1991. Sunk Costs and Market Structure: Price Competition, Advertising, and
the Evolution of Concentration. Cambridge, MA: MIT Press. Teece, D. J. 1986. Profiting from technological innovation: Implications for integration,
collaboration, licensing and public policy. Research Policy, vol. 15, pp. 285 – 305.
Thaler, R. 1980. Toward a positive theory of consumer choice. Journal of Economic Behavior and Organization, vol. 1, pp. 39 – 60.
Tripsas, M. 1997. Surviving radical technological change through dynamic capability: Evidence from the typesetter industry. Industrial and Corporate Change, vol. 6, pp. 341 – 377.
Tushman, M. L., and Anderson, P. 1986. Technological discontinuities and organizational environments. Administrative Science Quarterly, vol. 31, pp. 439 – 465.
Tversky, A., and Kahneman, D. 1974. Judgment under uncertainty: Heuristics and biases. Science, vol. 185, pp. 1124 – 1131.
_________________________. 1986. Rational choice and the framing of decisions. Journal of Business, vol. 59, pp. s251 – s278.
Urban, G., and von Hippel, E. 1988. Lead user analyses for the development of new industrial products. Management Science, vol. 34, pp. 569 – 582.
Electronic Submission Identification Number : xxxxx
35
Utterback, J. 1995. Mastering the Dynamics of Innovation. Boston, MA: Harvard Business School Press.
Utterback, J., and Abernathy, W. J. 1975. A dynamic model of process and product innovation. Omega, vol. 3, pp. 639 – 656.
Utterback, J., and Suarez, F. 1993. Innovation, competition, and industry structure. Research Policy, vol. 22, pp. 1 – 21.
Veryzer, R. W. 1998. Key factors affecting customer evaluation of discontinuous new products. Journal of Product Innovation Management, vol. 15, pp. 136 – 150.
Electronic Submission Identification Number : xxxxx
36
[ Table 1 ] Description of Variables
Variables
Description
Environment Characteristics
- Size of Tennis Playing Population - Number of Racket Manufacturers - Number of Rackets - Number of Rackets with Focal Design - Regulatory Change in 1981
: The number of people playing tennis regularly : The number of tennis racket manufacturers : The number of tennis rackets available in the market : The number of tennis rackets with focal design (i.e., oversize) : Dummy variable for change in rules on tennis rackets by the International Tennis Federation in 1981
Firm Characteristics
- New Entrant - Diversified - U.S. Ownership
: Whether a firm is a new entrant or an incumbent (1=Entrant; 0=Incumbent) : Whether a firm has related businesses (1=Diversified; 0=Single Business) : Whether a firm is an American or a foreign entity (1=American; 0=Foreign)
Design Characteristics
- Professional Endorsement - Design Quality (Playability) - Advertising
: The number of top professional players (top ten male and female players and semi-finalists in Grand Slam tournaments) who use rackets with focal design : Design quality measures (Overall racket performance measure by playtest panels) : The amount of advertising appearing in Tennis magazine
[ Table 2 ] Results of Experiment with Three Rackets
27 Inch Wood Racket 27 Inch Graphite-Composite Racket
29 ¼ Inch Graphite-Composite Racket
Average Speed of Serves 122 m.p.h. 124 m.p.h. 126 m.p.h.
Accuracy of Serves 60% 52% 80%
Source: Tennis, March 1997.
Electronic Submission Identification Number : xxxxx
37
[ Table 3 ] Correlation among Different Measures of Racket Attributes
Variables Lab Test Male Testers Female Testers
Mean S.D. 1 2 3 4 5 6 7 8 9 10
Lab Test
1. Stiffness 5.31 1.57
2. Stability 6.88 1.82 0.2**
3. Power 7.30 1.87 0.2* 0.6**
Male Testers
4. Stiffness 6.75 1.05 0.4** 0.1 0.1*
5. Stability 6.99 1.04 0.3** 0.4** 0.3** 0.6**
6. Power 7.11 1.00 0.2** 0.5** 0.5** 0.3** 0.6**
7. Playability 6.81 1.05 0.1* 0.2** 0.2** 0.4** 0.6** 0.6**
Female Testers
8. Stiffness 6.45 1.28 0.3** -0.1 -0.2** 0.4** 0.3** 0.2** 0.2*
9. Stability 6.64 1.25 0.3** 0.3** 0.1 0.3** 0.4** 0.3** 0.3** 0.6**
10. Power 7.35 1.01 0.3** 0.1 0.1* 0.3** 0.3** 0.3** 0.2** 0.3** 0.5**
11. Playability 7.00 1.17 0.1** 0.1* 0.1* 0.3** 0.3** 0.2** 0.4** 0.2** 0.6** 0.6**
* p < 0.05, ** p < 0.01
Electronic Submission Identification Number : xxxxx
38
[ Table 4 ] Poisson Regression of Racket Design Replication‡ ( from 1981 to 1992 )
Variable Model 1 Model 2 Model 3 Model 4
Intercept 29.714***
(6.081) 31.248***
(6.437) 24.501***
(8.596) 52.023***
(11.655) Environment Characteristics
Tennis Playing Population/1000 0.000
(0.000) 0.000
(0.000) 0.000**
(0.000) 0.000**
(0.000) Number of Racket Manufacturers -2.188***
(0.485) -2.323***
(0.537) -1.988***
(0.760) -3.770***
(0.992) Number of Racket Manufacturers 2 0.032***
(0.008) 0.034***
(0.008) 0.030**
(0.012) 0.063***
(0.017) Number of Rackets 0.032***
(0.008) 0.034***
(0.008) 0.036***
(0.013) 0.074***
(0.020) Number of Rackets 2 -0.000***
(0.000) -0.000***
(0.000) -0.000***
(0.000) -0.000***
(0.000) Regulatory Change in 1981 1.177***
(0.201) 1.151***
(0.204) 0.931***
(0.230) 1.163***
(0.225) Innovator Characteristics New Entrant (1=Yes; 0=No) − -1.251**
(0.500) -0.963*
(0.511) -1.966***
(0.440) Diversified (1=Yes; 0=No) − 0.668
(1.318) 0.408
(1.149) -0.176
(1.542) Domestic Firm (1=Yes; 0=No) − -0.884
(1.182) -0.794 (0.786)
-0.466 (1.181)
Design Characteristics
Professional Endorsement − − 0.169*** (0.027)
1.108*** (0.184)
Design Quality − − 0.494
(0.583) 1.970***
(0.633) Advertising − − 0.950
(1.315) 4.678**
(2.147) Professional Endorsement x Design Quality − − − -0.698***
(0.145) Design 1 (Oversize) 5.635***
(0.998) 7.776***
(1.459) 5.152*
(2.686) 2.717
(2.864) Design 2 (Longbody) 1.099
(1.247) 1.986
(1.427) 1.257
(1.751) -0.850 (1.878)
Design 3 (Ergonomic) 1.878* (1.008)
4.674* (2.514)
3.042 (3.349)
0.122 (3.995)
Design 4 (Three String) 1.135 (1.123)
3.903 (2.454)
2.803 (3.101)
1.268 (3.851)
Design 5 (Convex Throat) 2.998***
(1.109) 4.828***
(1.627) 3.442
(2.952) 0.481
(3.220) Design 6 (Fan String) 0.474
(1.534) 1.448
(1.692) 1.284
(1.719) -0.408 (1.846)
Design 7 (Widebody) 4.301***
(1.161) 5.173***
(1.240) 4.061*
(2.275) 0.563
(2.626) Design 8 (Tension Adjustable) 0.000
(0.000) 0.000
(0.000) 0.000
(0.000) 0.000
(0.000)
Log Likelihood 681.87 684.96 692.02 699.54
* p < 0.10; ** p < 0.05; *** p < 0.01 ‡ Playability is used for design quality. The numbers in parentheses are empirical standard errors.
Electronic Submission Identification Number : xxxxx
39
[ Figure 1 ] Number of Racket-related Patents
[ Figure 2 ] Change in Dominant Designs
0
20
40
60
80
100
120
1954 1956 1958 1960 1962 1964 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998
Year
Num
ber o
f Pat
ents
Issu
ed
Keyword Search by "Tennis Racket" or"Tennis Racquet"Class 273/73+
100%
59%
47%34%
39%
36%
13%
17%
32%
12%
77%
37%
61%
94%
60%
38%
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
1956 - 1965 1966 - 1970 1971 - 1975 1976 - 1980 1981 - 1985 1986 - 1990 1991 - 1995 1996 - 1998
Period
New
Rac
ket I
ntro
duct
ion 6. comp - over - wide - LONG
5. comp - over - WIDE - regular4. comp - OVER - regular - regular3. COMP - small - regular - regular2. METAL - small - regular - regular1. WOOD - small - regular - regular Miscellaneous
1
2
3
4
5
6
material - head - width - length
Electronic Submission Identification Number : xxxxx
40
[ Figure 3 ] A few examples of radical innovations
Electronic Submission Identification Number : xxxxx
41
[ Figure 4 ] Product Innovations as a Function of Producer and Consumer Conduct
Advertising, Endorsements,
Sales Reviews, Star
Producers
Consumers
Electronic Submission Identification Number : xxxxx
43
1 It should be noted that a new tennis racket with a focal racket design introduced by the firm who developed the design was excluded from the subset for the racket design since our main interest was the adoption of a racket design by competing firms. For example, an oversize racket introduced by Prince, the pioneer of this design, was eliminated from the oversize subset. 2 These designs were selected based on content analysis of racket reviews in popular magazines. We also conducted sensitivity analyses by including additional designs on the list. The results did not change the basic premise of our study. 3 The first dimension was ‘stiffness.’ It was the measure of the amount that racket frame bends during impact. While it was measured in pounds/inch in the lab, the numbers reported in the magazine was a relative scale, of which higher number means the racket was stiffer.
The second dimension was ‘stability,’ which measured the ability of a racket to resist the twisting motion with off-center hits. The unit of measurement was inch-pounds/degrees. As in the case of stiffness, the numbers reported in stability was also a relative scale. For example, one represented an absolute torsional stiffness of 10*(inch-pounds/degree) which was very flexible, while ten represented an absolute torsional stiffness of 100*(inch-pounds/degree) which was very stiff.
The last dimension was ‘power zone,’ which was the measure of the racket face area where the power exceeded a specified level of power. A racket with a higher number had a larger sweet spot. The total of 500 rackets were tested during the period from 1981 and 1995. While the test started in 1981, the sample included representative rackets introduced earlier than 1981. 4 We also tested the hypothesis with alternative measures of design quality: (1) an average of three attributes measured in the lab, (2) an average of three attributes measured on the court, (3) an index based on regression analysis of playability on three attributes measured in the lab, (4) an index based on regression analysis of playability on three attributes measured on the court, (5) an index based on factor analysis of three attributes measured in the lab, and (6) an index based on factor analysis of three attributes measured on the court. While the tables are not reported in the paper due to the limit on space, the main results are not qualitatively different from those in Table 4. 5 The results were not qualitatively different from those in the following even when we ran the same models based on one year comparison.