|
Article Excerpt
Innovation is one of the most important issues in business research today. It has been studied in many independent research traditions. Our understanding and study of innovation can benefit from an integrative review of these research traditions. In so doing, we identify 16 topics relevant to marketing science, which we classify under five research fields:
* Consumer response to innovation, including attempts to measure consumer innovativeness, models of new product growth, and recent ideas on network externalities;
* Organizations and innovation, which are increasingly important as product development becomes more complex and tools more effective but demanding;
* Market entry strategies, which includes recent research on technology revolution, extensive marketing science research on strategies for entry, and issues of portfolio management;
* Prescriptive techniques for product development processes, which have been transformed through global pressures, increasingly accurate customer input, Web-based communication for dispersed and global product design, and new tools for dealing with complexity over time and across product lines;
* Defending against market entry and capturing the rewards of innovating, which includes extensive marketing science research on strategies of defense, managing through metrics, and rewards to entrants.
For each topic, we summarize key concepts and highlight research challenges. For prescriptive research topics, we also review current thinking and applications. For descriptive topics, we review key findings.
Key words: innovation; new products; consumer innovativeness; diffusion models; network externalities; strategic entry; defensive strategy; ideation; rewards to entrants; metrics
History: This paper was received October 6, 2004, and was with the authors 4 months for 1 revision; processed by Leigh McAlister.
Introduction
Innovation, the process of bringing new products and services to market, is one of the most important issues in business research today. Innovation is responsible for raising the quality and lowering the prices of products and services that have dramatically improved consumers' lives. By finding new solutions to problems, innovation destroys existing markets, transforms old ones, or creates new ones. It can bring down giant incumbents while propelling small outsiders into dominant positions. Without innovation, incumbents slowly lose both sales and profitability as competitors innovate past them. Innovation provides an important basis by which world economies compete in the global marketplace.
Innovation is a broad topic, and a variety of disciplines address various aspects of innovation, including marketing, quality management, operations management, technology management, organizational behavior, product development, strategic management, and economics. Research on innovation has proceeded in many academic fields with incomplete links across those fields. For example, research on market pioneering typically does not connect with that on diffusion of innovations or the creative design of new products.
Overall, marketing is well positioned to participate in the understanding and management of innovation within firms and markets, because a primary goal of innovation is to develop new or modified products for enhanced profitability. A necessary component of profitability is revenue, and revenue depends on satisfying customer needs better (or more efficiently) than competitors can satisfy those needs. Research in marketing is intrinsically customer and competitor focused, and thus well situated to study how a firm might better guide innovation to meet its profitability goals successfully.
To encourage and facilitate further research on innovation in marketing, we seek to collect, explore, and evaluate research on innovation. Key goals of this paper are to provide a structure for thinking about innovation across the fields, highlight important streams of research on innovation, suggest interrelationships, and provide a taxonomy of related topics. Table 1 identifies five broad fields of innovation and various subfields within each of them. We hope this attempted integration will stimulate fertilization and interaction across fields and promote productive new research. This review attempts to summarize key ideas, highlight problems that are on the cusp of being addressed, and suggest questions for future research.
In the interests of space and relevance to marketing, our review is relatively focused. It does not include research on the antecedents of product development success (see Henard and Szymanski 2001 and Montoya-Weiss and Calantone 1994 for metaanalyses reviewing this research), the role of behavioral decision theory to inform product development (Simonson 1993, Thaler 1985), marketing's integration with other functional areas (Griffin and Hauser 1996), innovation metrics (Griffin and Page 1993, 1996; Hauser 1998), or the engineering aspects of product development (Ulrich and Eppinger 2000). Readers interested in an in-depth record of the extant literature can find an extended bibliography on www.msi.org, mitsloan.mit.edu/vc, and the Marketing Science Web site (http://mktsci.pubs.informs.org/).
Successful innovation rests on first understanding customer needs and then developing products that meet those needs. Our review of the literature, therefore, starts with our understanding of customers and their response to and acceptance of innovation. Because we are interested in how firms profit from innovation, the article then reviews organizational issues associated with successfully innovating and with how organizations adopt innovations. Customer understanding and the organizational context are underpinnings to innovating successfully. They must be in place before proceeding. The next three sections of the article then follow the flow of innovation: from first setting strategy in preparation for initiating development, through the prescriptions in the literature for moving the idea from conception and into the market, and ending with the rewards that accrue to innovators and defending against others entering.
The subsequent sections review each of the research topics within their corresponding research fields. When the research area is prescriptive, we attempt to summarize what can be accomplished and where the greatest challenges exist. When the research area is descriptive, we attempt to summarize the knowledge available today, the important gaps in that knowledge, and how that knowledge might lead to prescriptions.
Consumer Response to Innovations
"I don't want to invent anything that nobody will buy." Thomas Alva Edison
The success of innovations depends ultimately on consumers accepting them. Successful innovation rests on first understanding customer needs and then developing products that meet those needs. Our review of the literature starts with understanding customers. Research in many disciplines, but especially in marketing, has long sought to describe, explain, and predict how consumers (or customers (1)) and markets respond to innovation. A vast body of research has developed on the behavioral and decision aspects of this quest (Gatignon and Robertson 1985, 1991) and on the dynamics by which new products diffuse through a population (Rogers 2003).
Within this vast domain, we identify three subfields that have been particularly well researched or offer the most promise for managerial applications and future research: consumer innovativeness, models of new-product growth, and network externalities. Research on consumer innovativeness describes the mental, behavioral, and demographic characteristics associated with consumer willingness to adopt innovations. This research investigates adoption at the individual level. Models of new-product growth help firms understand and manage new products over their life-cycles. The diffusion literature focuses on understanding adoption at the aggregate level. Research on network externalities tries to understand the prevalence and effects of positive (or negative) feedback loops between consumers' adoption of a product and the product's value. This research focuses on understanding the relationship between individual-level adoption and patterns of aggregate adoption.
Consumer Innovativeness
Consumer innovativeness is the propensity of consumers to adopt new products. As Hirschman (1980, p. 283) suggested, "Few concepts in the behavioral sciences have as much immediate relevance to consumer behavior as innovativeness." Research on consumer innovativeness focuses on the characteristics that differentiate how fast or eagerly consumers adopt new products. We classify this research as focusing on the measurement of innovativeness, its relatedness to other constructs, and innovativeness variance across cultures.
Measurement. If innovativeness is a valid predictor for new-product adoption, then measures of innovativeness should identify those consumers most likely to adopt new products so that firms can target marketing efforts and improve forecasts. Over decades, researchers have developed and proposed numerous scales that differ in their theoretical premise, internal structure, and purpose (e.g., Midgeley and Dowling 1987). There has been no attempt to synthesize research or findings across all these different scales, although Roehrich (2004) has reviewed and classified them into two groups: (1) life innovativeness scales and (2) adoptive innovativeness scales.
The life innovativeness scales focus on the propensity to innovate at a general behavioral level. They describe attraction to any kind of newness and not to the adoption of specific new products. Kirton's (1976, 1989) innovators-adaptors inventory (KAI) is the most popular in this set of scales. However, because it taps innovativeness in general, its predictive validity tends to be low (Roehrich 2004).
The adoptive innovativeness scales focus specifically on the adoption of new products. Examples of these scales are Raju (1980), Goldsmith and Hofacker (1991), and Baumgartner and Steenkamp 1996). Raju's (1980) scale has good internal consistency, but Baumgartner and Steenkamp (1996) criticize it for its structure. Goldsmith's and Hofacker's scale (1991) measures domain-specific innovativeness, but Roehrich (2004) questions its discriminant validity. Baumgartner and Steenkamp (1996) developed a scale to measure consumers' tendency toward exploratory acquisition of products (rather than innovativeness per se). Exploratory acquisition is similar to innovativeness expressed in information seeking.
Despite extensive research, progress in this area has been hindered by a lack of consensus about a most appropriate scale. Actually, researchers have not yet agreed about a single definition of innovativeness. Current definitions vary from an innate openness to new ideas and behavior, to propensity to adopt new products, to actual adoption and usage of new products.
Relatedness to Other Constructs. Many researchers have used the measures of innovativeness to study its relationship to other constructs. Im et al. (2003), Midgeley and Dowling (1993), and Venkatraman (1991) explored the relationship between innovativeness and demographics. Foxall (1988, 1995); Foxall and Goldsmith (1988); Goldsmith et al. (1995); Manning et al. (1995); and Midgeley and Dowling (1993) studied the relationship between innovativeness and the adoption of innovations. Steenkamp et al. (1999) and Hirschman (1980) researched the relationship between innovativeness and other related constructs. While some studies have shown that innovators are better educated, wealthier, more mobile, and younger, other studies have failed to validate these findings (Rogers 2003, Gatignon and Robertson 1991). Another stream of research uses innovativeness measures combined with other observable characteristics such as marketing strategy, marketing communication, and category characteristics to predict actual trial probability for a new product (Steenkamp and Katrijn 2003).
This research is promising because it connects consumer innovativeness with observable characteristics. It could benefit from a synthesis with earlier models of pretest market analyses, such as Claycamp and Liddy (1969). In practice, many pretest market analyses often merge laboratory measures with "norms" based on past experience. The primary limitation of this literature is the lack of consensus on measures, scales, and methods of research. However, the adoption of new products by consumers is crucial to new product success. It is important to understand what drives consumers' propensity to adopt new products.
Variation Across Cultures. Currently there is a small but important effort to study the innovativeness of consumers across diverse cultures and countries. For example, Steenkamp et al. (1999) studied 3,000 consumers across 11 countries of the European Union. Tellis et al. (2004) studied over 4,000 consumers across 15 major countries of the Americas, Europe, Asia, and Australia. They find that innovativeness differs systematically across countries, although innovators also show certain demographic commonalities. Such analyses can throw light on optimal strategies for global entry. By using the same instrument across cultures, researchers can partly bypass the problem of choosing the appropriate scale. However, to obtain valid results, researchers need to ensure that the instrument is properly translated, back translated, and retranslated. They also need to control for cultural biases in responsiveness, such as reticence among east Asians or exuberance among southern Europeans.
Research Challenges. The key challenge is the need for a consensus among researchers on measures, scales, and methods of inquiry. This research would be facilitated with a deeper underlying theory that includes individual characteristics as well as the individual's relationship to the social network (e.g., Allen 1986, Souder 1987, Van den Bulte and Lilien 2001). Specific research opportunities include:
* Developing parsimonious, unified scales for consumer innovativeness that encompass the strengths of existing scales while avoiding their weaknesses;
* Using such a scale to study how or whether innovativeness varies across product category, geography, or culture;
* Identifying within-country differences in innovation that might be due to ethnic, cultural, demographic, or historical factors;
* Linking individual-level theories of innovativeness with social networks;
* Assessing the ability of innovativeness to predict the adoption of specific new products and, in particular, a synthesis with the prescriptive models of pretest market analyses;
* Incorporating measures of individual consumer innovativeness into models of new-product growth (reviewed in the next section).
Growth of New Products
Consumer innovativeness critically affects the adoption of new products and their subsequent growth. While the research on consumer innovativeness focuses on adoption at the individual level, the new product diffusion literature focuses on adoption at the aggregate level. The aggregate growth of new products has enjoyed intensive study in marketing over the last 35 years, beginning with Bass (1969) and now totaling over 700 estimates of the parameters of diffusion or applications of the model (Bass 2004, Van den Bulte and Stremersch 2004).
The Bass model expresses the adoption of a new product as a function of spontaneous innovation of consumers (due to unmeasured external influence) and cumulative adoptions to date (due to unmeasured word of mouth). The basic model is estimated using three parameters, which have been interpreted as the innovation rate (or coefficient of external influence), the imitation rate (or coefficient of internal influence), and the market potential. The ratio of these coefficients defines the shape of the sales curve and the speed of diffusion; their typical sizes are responsible for the commonly observed S-shape of new product sales for most consumer durables (Van den Bulte and Stremersch 2004).
The Bass model has had great appeal and widespread use because it is simple, generally fits data well, enables intuitive interpretations of the three parameters, and performs better than many more complex models. At the same time, the model has some limitations that subsequent research sought to address. First, the original model did not include explanatory variables, such as marketing-mix variables, that firms use to influence the imitation rate or total market potential. When included, these variables complicate specification and estimation. Second, the model's parameters are highly sensitive to the inclusion of new data points. Parameter estimates based on six years of data may be very different than estimates using eight years of data. Third, the original estimation by multiple regression suffered from multicollinearity. Fourth, estimating the model requires knowing two key turning points in early sales (takeoff and slowdown); however, once these events have occurred, the model's value is primarily descriptive or retrospective, rather than predictive.
A vast body of research has explored solutions to these and other problems. Examples of subsequent research include modeling:
* Dependence of the three key parameters on relevant endogenous and marketing or exogenous variables (e.g., Horsky and Simon 1983, Kalish and Lilien 1986, Kalish 1985);
* Improvements in estimation analytics, including maximum likelihood estimation (Schmittlein and Mahajan 1982), nonlinear least squares (Jain and Rao 1990, Srinivasan and Mason 1986), Bayesian estimation (Sultan et al. 1990), hierarchical Bayesian estimation (Lenk and Rao 1990, Talukdar et al. 2002), augmented Kalman filter (Xie et al. 1997), and genetic algorithms (Venkatesan et al. 2004);
* Dependence of diffusion on related innovations (e.g., Bayus 1987, Peterson and Mahajan 1978);
* Successive generations of innovation (e.g., Bass and Bass 2004, Norton and Bass 1987);
* Adopter categories (e.g., Mahajan et al. 1990);
* Variation of parameters across countries and their explanation by sociological, economic, and cultural factors (e.g., Gatignon et al. 1989, Putsis et al. 1997, Roberts et al. 2004, Takada and Jain 1991, Talukdar et al. 2002, Van den Bulte and Stremersch 2004);
* Stages in the adoption process (e.g., Kalish 1985, Midgeley 1976);
* Supply restrictions (e.g., Ho et al. 2002, Jain et al. 1991);
* Continuous-time Markov models (Hauser and Wernerfelt (1982a, b);
* Repeat and replacement purchases (Lilien et al. 1981, Mahajan et al. 1984);
* Retailer adoption (e.g., Bronnenberg and Mela 2004) and spatial diffusion (Garber et al. 2004);
* Processes for interpersonal communication (e.g., cellular automata, Garber et al. 2004, Goldenberg et al. 2002);
* Cross-market communication (Goldenberg et al. 2002).
Detailed reviews of this area are available (Mahajan et al. 1990, Chandrasekaran and Tellis 2005). Rogers (2003) positions this research stream in a broader review of research on the diffusion of innovations. Sultan et al. (1990) and Van den Bulte and Stremersch (2004) provided meta-analytic estimates of model parameters. Mahajan et al. (1995) provided a summary of the empirical generalization of the research. These reviews suggest an emerging consensus on the following points:
* A plot of sales over time in the early years of the product life-cycle is generally S-shaped unless there is cross-market communication, in which case there may be a slump in sales.
* The S-shaped curve could emerge from social contagion among consumers or due to increasing affordability among a heterogeneous population of consumers.
* The S-shaped curve seems to hold for successive generations of the product.
* The coefficient of innovation is relatively stable and averages about 0.03.
* The coefficient of imitation varies substantially across contexts, with an average of about 0.4.
* The ratio of the coefficients of imitation to innovation is increasing over calendar time, indicating a faster rate of diffusion of new products.
Although the extant literature on the growth of new products is enormous, recent research in the area suggests new directions. First, there are some product categories for which a different pattern of adoption applies. For example, when weekly movie sales are plotted against time, the shape of the curve seems to decline exponentially, with a peak in one of the first few weeks (e.g., Eliashberg and Shugan 1997, Sawhney and Eliashberg 1996). This pattern holds for national and international sales (e.g., Elberse and Eliashberg 2003) and for theater and video sales (e.g., Lehmann and Weinberg 2000). A model based on the Erlang 2 distribution seems to fit weekly sales of movies better than the Bass model, suggesting that additional forces may be affecting movie sales differentially, such as initial marketing efforts, the impact of the distribution chain (movie theaters), or repeat viewing.
Second, the Bass curve seems to be punctuated by two distinct turning points--takeoff and slowdown--as illustrated in Figure 1 (Agarwal and Bayus 2002, Foster et al. 2004, Golder and Tellis 1997, Kohli et al. 1999, Stremersch and Tellis 2004, Tellis et al. 2003). Takeoff is the sudden spurt in sales that follows the period of initial low sales after introduction. Slowdown is a sudden leveling in sales that follows a period of rapid growth. Slowdown frequently is followed by what has been called a saddle, trough, or chasm (Goldenberg et al. 2002, Golder and Tellis 2004, Moore 1991). The above empirical studies over multiple categories of consumer durables suggest the following potential generalizations:
* New consumer durables have long periods of low growth before takeoff, steep growth after takeoff, and erratic growth after slowdown.
* The time to takeoff currently averages six years, the growth stage about eight years, and trough about five years.
* These patterns, especially time to takeoff, vary systematically and dramatically by country.
* New products take off and grow much faster in recent decades than in earlier ones.
* New electronic products have a much shorter time to takeoff and faster growth rate than other household durables.
Research Challenges. Despite substantial research, many challenges remain for future research, including:
* Exploring the generalizability of the S-shaped curve, the turning points, and the declining exponential growth curves across categories;
* Developing an integrated model to predict the turning points in the S-shaped curve, such as compound hazard models, multivariate regime-switching models, or time-series models with structural breaks;
* Exploring competing theories for the S-shaped curve and the turning points, such as social contagion, heterogeneity in proximity (crossing the chasm), heterogeneity in income (affordability), informational cascades, or network externalities (see below);
* Comparing the patterns and dynamics of new-product growth across countries, cultures, and ethnic groups;
* Determining whether and how network effects influence diffusion (see the next section).
[FIGURE 1 OMITTED]
Network Externalities
Consumer acceptance of new products and their subsequent growth can be affected greatly by network externalities. Network externalities refer to an increase in the value of a product to a user based on either the number of users of the same product (direct network externality) or the availability of related products (indirect network externality). For example, fax machines exhibit a direct network externality because the value of each node (fax machine) increases with more users who can receive or send faxes. DVD players exhibit an indirect network externality because the value of each DVD player increases as more DVD titles for the player become available. More titles will become available if there are more DVD players. Similar indirect network externalities exist for HDTV sets (available programming), alternative-fuel vehicles (refueling stations), and computer hardware platforms (software programs).
Many economists have studied whether firms become monopolies or grow and stay dominant in markets due merely to network externalities (e.g., Church and Gandal 1992, 1993; Farrell and Saloner 1985, 1986; Katz and Shapiro 1985, 1986, 1992, 1994). Based on this line of research, regulators have argued that Microsoft holds monopoly power in the operating system market, in part, because of network externalities: The Windows operating system and Office products are more attractive to customers because so many other customers own and use them.
Another premise that some economists have postulated is the existence of path dependence--early dominance of a market (due to early entry or some favorable event) might lead to the inability of subsequent superior products from ever becoming successful (Arthur 1989, Krugman 1994). A classic example cited in favor of this theory is the success of the QWERTY keyboard over the Dvorak keyboard, to which some researchers attribute performance superiority.
A...
|
