Wednesday, December 4, 2019

Survival Strategy for Startup Business free essay sample

We appreciate comments of the following people on a much earlier version of this paper: Jay Barney, Gaurab Bhardwaj, Oliver Chatain, Raj Echambadi, Glenn Hoetker, Steven Klepper, MB Sarkar, Anju Seth, Charles Williams and participants at seminars/conference presentations given at University of Illinois at Urbana Champaign, Purdue University, University of Toronto, the 1st ACAC conference, the 2003 Academy of Management and Strategic Management Society meetings and the 2nd West Coast Symposium on Entrepreneurship. More recently, the comments of participants at the Penn State and Texas AM Marketing Research Camps have been valuable. We especially appreciate the comments of the journal associate editor and reviewers, as well as the financial support of the Ewing Marion Kauffman Foundation. All remaining errors are ours. Product Strategies and Firm Survival in Technologically Dynamic Industries ABSTRACT Studying the US personal computer industry from its inception in 1974 through 1994, we address the following questions. What product strategies increase the survival chances of entrants into new, technologically dynamic industries? Does the effectiveness of these product strategies differ by pre-entry experience? Does the effectiveness of these product strategies differ by when firms enter a new industry? Consistent with the published literature, we find that diversifying entrants have an initial survival advantage over entrepreneurial startups. We will write a custom essay sample on Survival Strategy for Startup Business or any similar topic specifically for you Do Not WasteYour Time HIRE WRITER Only 13.90 / page But, we find the reverse for later entrants: startups that enter later in the industry have a survival advantage over the later entering diversifying entrants. We explain this finding in terms of the firms’ product strategies, pre-entry experience, and entry timing. Importantly, our research is very revealing over the existing literature—the effects of pre-entry experience on firm survival disappear when controls for product strategy are included in the analysis. Our findings highlight that it is crucial to study what firms do after they enter a new industry in order to more completely understand their ultimate performance. 1. Introduction What product strategies increase the survival chances of entrants into new, technologically dynamic industries? Does the effectiveness of these product strategies differ by pre-entry experience? Does the effectiveness of these product strategies differ by when firms enter a new industry? Providing answers to these important questions has long been of interest to researchers in the economics, marketing, management, and strategy disciplines. Unfortunately, a complete understanding of why some entrants into new industries ultimately fail is still lacking. Studies of organization mortality tend to fall into one of three main streams of inquiry: one stresses the importance of environmental and industry-level factors, a second emphasizes the pre-entry experience of entrants, and the third considers firms’ post-entry activities. Within the first stream, organizational ecologists argue that corporate demographics matter (e. g. , see the reviews in Carroll and Hannan 2000; Carroll and Khessina 2005). In particular, a number of empirical studies demonstrate that firm tenure and size in the new industry, as well as competitive density, are important explanatory factors related to survival. In a review of the second research stream, Helfat and Lieberman (2002) conclude that diversifying entrants have access to relevant resources that bestows a survival advantage over entrepreneurial startup firms with no pre-entry experience. Indeed, several empirical studies confirm that diversifying entrants with prior experience have higher survival rates than entrepreneurial startups (e. g. , Mitchell 1991; Carroll et al. 1996; Klepper and Simons 2000; 2005; Klepper 2002). Because the evolution of new and emerging industries crucially depends on innovation and new product introductions (e. g. , Gort and Klepper 1982; Agarwal and Bayus 2002), research in the third stream tends to focus on the relationship between product strategy, variety, and firm survival (e. g. , Romanelli 1989; Christensen et al. 1 998; Dowell and Swaminathan 2000; Sorenson 2000; Barnett and Freeman 2001; Jones 2003). As a whole, this body of research has increased our knowledge of the factors related to firm survival in new industries. But, this understanding has come from isolated analyses within each research stream. Only a few studies consider the joint effects of corporate demographics and pre-entry experience (e. g. , Carroll et al. 1996; Tea garden, et al. 2000), and even fewer attempt to integrate findings across the pre-entry experience and post-entry product strategy literatures (e. g. , Fosfuri and Giarratana 2004). In this paper, all three factors are considered. We empirically study the relationship between firm survival and the product strategies employed by diversifying entrants and entrepreneurial startups, while controlling for key corporate demographic effects. We examine these effects in a technologically dynamic setting, i. e. , a new industry characterized by the simultaneous availability of successive generations of improved product technologies. Our emphasis is on the potential conditioning effects of pre-entry experience and entry time on the relationship between product strategies and firm survival. Importantly, our research approach is very revealing over the existing literature—the effects of pre-entry experience on firm survival disappear when controls for product strategy are included in the analysis. This finding highlights that it is crucial to study what firms do once they enter a new industry in order to understand any performance outcomes. The empirical setting for our research is the US personal computer industry from its inception in 1974 through 1994. The personal computer industry has been one of the most innovative sectors of the economy and one of its most competitive. This industry is a rich and dynamic context in which to study product strategies and firm survival (e. g. , see the discussions in Langlois 1992; Steffens 1994; Bayus 1998). Entrants into this new industry 2 included diversifying entrants (e. g. , IBM, Epson America, Tandy/Radio Shack), as well as entrepreneurial startups (e. g. , Apple, Dell, Eagle Computers). Distinct from empirical studies that consider less technologically-dynamic industries like automobiles or tires (e. g. , Carroll et al. 996; Klepper 2002), prominent features of the personal computer industry are the availability of multiple, overlapping product technologies at any point in time, rapidly advancing technology, frequent new product introductions, ease of firm entry and exit, and the inability of any single firm to establish a long-term competitive advantage. Consistent with the published literature, we find that diversifying entrants have an initial survival advantage over entrepreneurial startups. But, we find the reverse for later entrants: startups that enter later in the industry have a survival advantage over the later entering diversifying entrants. To explain this result, we empirically demonstrate that the product strategies related to higher survival rates differ by pre-entry experience and entry time. In the early years of a new industry before product standards are set, typically there are several alternative product technologies from which entrants can choose. Many entrants will not initially select the product technology that will eventually become the standard, and thus they will have high risks of failure. Among these early entrants however, diversifying entrants with greater resources are better able than startups to migrate to the product standard when it becomes known. As a result, the early diversifying entrants have higher survival rates than the early entering startups. Once the technological trajectory is established however, survival depends on introducing products with the latest technology. By virtue of their lower sales targets, startups can grow by market expansion (i. e. , introducing new products based on the most recent technology). Diversifying entrants, on the other hand, usually have higher sales requirements and thus attempt to grow via market penetration (i. e. , introducing â€Å"popular† products that are typically based on â€Å"older† product 3 technology). Given the importance of staying close to the technology frontier in the later stages of industry evolution, later entering startups have higher survival rates than later entering diversifying entrants. 2. The Personal Computer Industry The traditional viewpoint in industrial organization is that the evolution and shakeout of new industries follows the product life cycle pattern: an initial period of intense competition, significant entry and exit of firms, and fragmented market shares is eventually followed by a shakeout in which the number of firms dramatically falls, leading to higher industry concentration (e. g. , Gort and Klepper 1982). This pattern is consistent with the technology management literature that maintains there is a shift over the product life cycle from product to process innovation as a dominant design emerges (e. g. , Utterback 1994; Christensen, et al. 1998). Under these industry conditions, firms with the lowest costs grow to be bigger and the firms with the lowest costs are those with pre-entry experience and those that enter early (Klepper 1996). Empirical research demonstrates that pre-entry experience, time of entry, and exploitation of scale economies are crucial determinants of firm survival in traditional shakeout industries (e. . , Carroll, et al. 1996; Klepper 2002; Klepper and Simons 2000; 2005). Importantly, research in this stream generally downplays the role of post-entry product strategies. While the industry we study adheres to this general product life cycle pattern, the technologically dynamic environment of the personal computer industry is quite different. Our information on the US personal computer market comes from International Data Corporation’s (IDC) Processor Installation Census 1 . Details of the data are discussed in a later A personal computer is defined as a general-purpose, single-user machine that is microprocessor based and can be programmed in a high-level language. In our study, personal computers include all desktop, tower, notebook, and laptop computers (excluding workstations) selling for less than $15,000. As noted by Lawless 4 section. As shown in Figure 1, the personal computer industry has witnessed rapid growth since its inception in 1974. Personal computer sales grew from a few thousand units in 1975 to over 18 million by 1994. Figure 1 also shows that the number of competitors in this industry steadily grew between 1974 and its peak of almost 250 firms in 1989. Since 1983 there have been over 100 competing firms in this industry in any given year. Not surprisingly, the proliferation of advanced technology has encouraged frequent new product introductions. Moreover, significant entry and exit occurs in this industry throughout the time period of our study (see Figure 2). [insert Figures 1 and 2 about here] Both hardware and software technology improved substantially over this twenty-one year period (e. . , Curry and Kenney 1999; Evans et al. 1999). Figure 3 shows unit sales associated with each successive microprocessor 2 technology generation (2nd generation technology became available in 1979, 3rd generation in 1982, 4th generation in 1985, 5th generation in 1989, and 6th generation in 1993) 3 . Each new microprocessor is associated with increased processing speed, enabling the development and use of more sophisticated operating systems, graphics, and application packages. As such, each new microprocessor entails high associated switching costs between generations (e. . , Anderson 1995; Wade and Anderson (1996), IDC is the oldest among the various companies that tracks the computer industry and is widely respected as having an accurate picture of the activity in this industry. 2As discussed in Steffans (1994) and Anderson (1995), the most parsimonious way to describe the technology generations of personal computers is to compare their microprocessors or CPUs (central processing unit). The CPU is the brain of the computer since it contains the arithmetic and logic component, as well as the core memory and control unit for the computer. Thus, CPU design determines the computer’s overall power and performance. follow the common convention of distinguishing technology generations as follows (e. g. , see The PC Tech Guide 2004): 1st generation (8-bit CPUs, including Zilog’s Z80, Mostek’s 6502, Intel’s 8080), 2nd generation (e. g. , Intel’s 8088 and 8086, NEC’s V20-40), 3rd generation (e. g. , Intel’s 286, Motorola’s 68000 and 68010), 4th generation (e. g. , AMD and Intel’s 386, Motorola’s 68020), 5th generation (e. g. , AMD and Intel’s 486), 6th generation (e. g. , Intel’s Pentium). 3We 5 1995). The curves in Figure 3 make it clear that sales of personal computers with older technology were dominant for a number of years after more advanced technology became available (e. g. , even though more advanced 2nd generation technology was available in 1979, sales of personal computers with 2nd generation technology did not surpass sales of products with 1st generation technology until 1985) 4 . [insert Figure 3 about here] In the early years of this new industry, firms had a wide choice of microprocessors from manufacturers like Motorola, Intel, Mostek, Zilog, RCA, Texas Instruments, Rockwell, National Semiconductor, and Signetics. Although Intel’s x86 CPU architecture was available in 1979, it was not until IBM introduced their PC 5150 with the Intel 8088 in late 1981 that Intel became the dominant CPU design for personal computers (e. g. , Steffens 1994; Anderson 1995; Teagarden, et al. 1999). By 1988, the Intel x86 architecture had become the industry standard as personal computer sales with an Intel x86 CPU represented over 50% of the market (Steffens 1994). 3. The Conceptual Framework and Hypotheses Organizational research emphasizes that entrant heterogeneity is an important factor affecting subsequent firm performance. In particular, pre-entry experience is an important source of heterogeneity since founding conditions that imprint on an organization can have long-lasting effects (Stinchcombe 1965). A nascent industry has very little industry specific stock of knowledge (e. g. , Gort and Klepper 1982), and thus has a malleable institutional environment. Diversifying entrants generally possess a wide range of resources and capabilities than can be leveraged into a new industry, including capital, organizational 4Based n the IDC data, the most â€Å"popular† microprocessors in terms of sales were the following: 8080 (19741976), Z80 and 6502 (1977-1982), 6502 (1983-1984), 8088 (1985-1986), 286 (1987-1989), 386 (1990-1992), 486 (1993-1994). 6 structure, technical and market knowledge, specialized skills, and experience from related activities (e. g. , see the review in Helfat and Lieberman 2002). Since these firms bring relevant experiences to help structure the uncertain marketplace, diversifyi ng entrants are expected to have a survival advantage during the early years of industry evolution (e. g. , Klepper and Simons 2000; Klepper 2002). Moreover, the resource endowments of diversifying entrants enable them to leverage or develop collateral assets that help build market infrastructure and create customer demand in the emerging market (e. g. , Teece 1986, Tripsas 1997). In the absence of industry specific knowledge and legitimacy among consumers, the endowment and reputation effects of diversifying entrants during the early years acts as a surrogate mechanism to tip the balance in their favor. Consistent with these studies, we expect that diversifying entrants will have an initial survival advantage over entrepreneurial startups. At the same time, the dynamic changes that characterize industry evolution are documented across rich bodies of literature in technology management (e. g. , Utterback 1994), organizational ecology (e. g. , Carroll and Hannan 2000), and evolutionary economics (e. g. , Gort and Klepper 1982). Evolution introduces a dynamic element into selection processes since the customer demand facing firms changes as the industry transforms. Existing research generally argues on theoretical grounds that organizational inertia can make it difficult for diversifying entrants to fully exploit any new opportunities as the industry changes, i. . , if the pace of change in the industry is faster than the pace of change within the organization, the initial match of resources and capabilities with the diversifying entrants will erode over time (e. g. , Baum, et al. 1994). It is generally presumed that their relative lack of flexibility (e. g. , Hannan and Freeman 1984), potential incompatibility of complem entary assets (e. g. , Teece 1986; Tripsas 1997), and internal politics (e. g. , Pfeffer and Salancik 1978) 7 may render the diversifying entrants incapable of changing as quickly as required by the environment. Thus, this line of reasoning centers on a firm’s flexibility in adapting to dynamic industry conditions and how this flexibility may differ by pre-entry experience. A few studies in the pre-entry experience literature incorporate dynamic effects by examining how entry timing and firm age conditions the impact of prior experience on firm survival. Teagarden, et al. (2000) and Klepper (2002), for instance, find that the relative survival advantage of diversifying entrants is greater for firms that enter early rather than later, and Carroll, et al. 1996) find some support for organizational inertia (i. e. , young diversifying entrants have higher survival rates than startups, but this advantage diminishes as firms age). Consistent with this research, we expect that any initial survival advantage of diversifying entrants over entrepreneurial startups will diminish with entry time 5 . Although prior empirical research considers how entry time and firm age moderates the relationship between pre-entry experience and firm survival, it does not directly consider the underlying mechanism for the proposed effects. In contrast to these theoretical rguments, we offer an alternative framework in this section based on the product strategies employed at different times by diversifying entrants and entrepreneurial startups. We propose that survival depends on the product strategies implemented by firms and that the success of these strategies is closely linked to pre-entry experience and entry timing. Unlike Klepper (2002), we argue that the effectiveness of these product strategies is heterogeneous across firms’ pre-entry experience and the effectiveness of these strategies varies as the new industry evolves (i. e. the efficacy of a specific product strategy not only depends on who implements it, but also when it is employed). the course of our empirical analyses of the personal computer industry, we found no evidence that firm age moderates the relationship betw een pre-entry experience and firm survival. 5In 8 3. 1 Product Strategies and Firm Survival The first commercialized forms of an innovation are typically primitive in nature (e. g. , the first personal computer was â€Å"a box with blinking lights†). Competition in the early years of a new industry is primarily on the basis of continued product improvements (e. g. Gort and Klepper 1982; Agarwal and Bayus 2002). As a result, product variety increases as firms experiment with different designs, technologies, and product combinations (e. g. , Tushman and Anderson 1986). This variation is associated with high uncertainty about which technology will eventually become the product standard (e. g. , David and Greenstein 1990; Gabel 1991; Schilling 1998) or dominant design (e. g. , Tushman and Anderson 1986; Utterback 1994). In most cases, a single product architecture is widely accepted as the industry standard. Even when multiple technologies persist, the increasing returns ssociate d with network effects and technology lock-in suggests that survival is intimately related to whether or not a firm joins the bandwagon of firms, customers, and suppliers supporting a particular product standard (e. g. , Wade 1995; Schilling 1998). Thus, we have the following hypothesis. H1: Firms that offer products incorporating the technology standard have higher survival rates than firms that do not. A prominent characteristic of technologically dynamic industries is the simultaneous availability of successive generations of improved product technologies (e. g. , see Figure 3). In this kind of setting, the firm’s challenge is to manage its product offerings across the different product technologies available in every year and to plan for the newer technologies that can be used to continually refresh its product line. Firms applying practices, routines, and knowledge across product generations can gain a competitive advantage over firms that do not (e. g. , Burgelman 1994; Iansiti and Clark 1994). In the context of the early US bicycle industry, Dowell and Swaminathan (2000) empirically find that firms offering products in 9 overlapping (successive) technology generations had lower mortality rates. In technologically dynamic industries, firms that do not offer any products with the latest technology risk having an obsolete set of offerings. hypothesis. H2: Firms that offer products with the most recent technology have higher survival rates than firms that do not. 3. 2 The Conditioning Effects of Entry Time and Pre-Entry Experience In the early years of a new industry before the product standard is widely established, firms generally offer products based on competing, incompatible designs. As new and improved products and technologies become available, firms must adapt their offerings to avoid obsolescence. Often this means that firm survival depends on switching to a product technology that the firm did not originally develop. Diversifying entrants have an advantage over entrepreneurial startups in this situation since their prior experience in new product development better prepares them to respond to technological changes (e. g. , Meyer and Roberts 1986). Startups tend to focus on a more narrow technological area (Meyer and Roberts 1986) and thus become locked-in to their initial product designs because they lack the resources, knowledge, and experience to either change or modify them (e. g. Tushman and Anderson 1986; Tegarden, et al. 2000). Diversifying entrants are also less likely to be overconfident about their original product technology choices, and therefore more willing than startups to change to a more promising alternative (Busenitz and Barney 1997). Together, these arguments suggest that diversifying entrants are more adaptable than startups during the early years of a new industry when uncertainty about the eventual product standard is high. Thus, it is not as important for diversifying entrants to initially enter with the technology standard since they can later migrate once it is known. Based on this line of reasoning, along with H1, we propose the following hypothesis. 10 Thus, we propose the following H3: Among early entrants that do not enter with products incorporating the technology standard, diversifying entrants have higher survival rates than entrepreneurial startups. Once a product standard is accepted, competition shifts from alternative technological designs to market growth (e. g. , Brush, et al. 2000; Mishina, et al. 2004). Because diversifying entrants are generally larger than entrepreneurial startups, they have greater sales requirements to meet their corporate growth targets (Penrose 1959). Thus, we expect that diversifying entrants will be most interested in pursuing product strategies that have the greatest potential to generate high sales levels. Unlike diversifying entrants, however, startups usually have a narrow resource base, lack of capital, and limited technical and marketing experience that is directly relevant to the new industry. While diversifying entrants tend to act like generalists, startups have characteristics in common with specialists (e. g. , Carroll, et al. 2002; Khessina and Carroll 2004; Sorenson, et al. 2005). As a result, entrepreneurial startups are more likely than diversifying entrants to be successful focusing on smaller niches rather than adopting a strategy targeting a wider set of customers. By adopting a niche strategy, startups can also circumvent direct competition with the larger diversifying entrants that target the larger market segments. The later market conditions in a technologically dynamic industry suggest that startups will be more interested than diversifying entrants in introducing products with the latest technology when it first becomes available. The reason for this is hat sales for the products with the most recent technology are typically much lower than sales of products with older technology (see Figure 3). Consistent with their existing organizational routines and human capital resources, diversifying entrants tend to pursue a market penetration strategy whereby they offer products based on â€Å"familiar† technology to a greater number of customers (e. g. , Mishina, et al. 2004). Entrepreneurial startups, on the other hand, tend to 11 adopt a market expansion strategy in which they introduce products with the latest technology as it becomes available to a more narrow market segment. Together, these arguments suggest that the later entering startups are more nimble than diversifying entrants since they incorporate the most recent technology into their products. Thus, it is less important for startups to initially enter with the most recent technology since they can later expand their product line. This line of reasoning, together with H2, leads to the following hypothesis. H4: Among later entrants that do not enter with products using the most recent technology, entrepreneurial startups have higher survival rates than diversifying entrants. 4. Data and Variable Definitions The population of US personal computer manufacturers we study is based on a census listing from IDC of all domestic firms and foreign subsidiaries that built such products in the US during 1974-1994 6 . Annual firm-level data were constructed from detailed product-level information in the IDC database. The resulting data set includes 3,083 firm-year observations for 624 personal computer manufacturers (78% of these firms exited before 1994). Summary descriptive statistics of our variables is in the Appendix. 4. 1 Firm Survival Like most studies in this research stream (e. . , Carroll, et al. 1996; Klepper 2002), our analyses are conditional on firms having already made their entry decision (i. e. , all firms in our sample entered the personal computer industry at some point). Thus, the dependent variable we analyze is the timing of firm exit from the personal computer industry. A firm is considered to have exited in year t if its unit sales for years t+1 to 1994 were zero; otherwise, 6This information is only available through 1994 since IDC changed its data collection procedure to a more aggregate format in 1995. 2 the firm exit date was coded to be a right censored observation. As noted by Stern and Henderson (2004), the personal computer industry is predominated by exits of singlebusiness entities; the few multi-business corporations (e. g. , Tektronix) that exited were treated as failures. In this industry, acquisitions were infrequent and when they did occur, the acquired firm continued to operate as a distinct entity from the parent (e. g. , even though ATT acquired NCR in 1991, NCR was left intact; see Swanson 2002; Stern and Henderson 2004). 4. Pre-Entry Experience To compile information on pre-entry experience, we primarily referred to the annual volumes of the Thomas Register of American Manufacturers. The Thomas Register, which dates back to 1906, is a national buying guide that has been used to study firm activity in th e evolution of markets (e. g. , Gort and Klepper 1986; Klepper 2002a; Agarwal and Bayus 2002). In describing various sources of business information, Lavin (1992, p. 129) states that â€Å"the Thomas Register is a comprehensive, detailed guide to the full range of products manufactured in the United States. Covering only manufacturing companies, it strives for a complete representation within that scope. † Pre-entry experience (if any) was determined by manually matching each firm in the IDC database with its corresponding information in the Thomas Register. More specifically, as in Agarwal, et al. (2002), if a firm was listed in the index volumes of the Thomas Register for the year preceding its entry into personal computers, it was classified as a diversifying entrant. The resulting classifications were also confirmed using other data sources such as Lexis/Nexis and the International Directory of Company Histories. Personal computer firms that did not appear in these sources before their inclusion in the IDC database were classified as being a new start-up (e. g. , Apple, Compaq, Dell, Acer). As is typically the case in new 13 industries (e. g. , Carroll, et al. 1996; Helfat and Lieberman 2002), the majority of entrants in the personal computer industry are entrepreneurial start-ups with no prior industry experience (almost 75% of the firms in our sample are startups). We define the variable Startup to be 1 if the firm is classified as having no pre-entry experience, and 0 if the firm is a diversifying entrant. . 3 Entry Timing and Corporate Demographics Firm entry timing plays a prominent role in several studies examining the relationship between pre-entry experience and firm survival. For example, Teagarden, et al. (2000) and Klepper (2002) find that the relative survival advantage of diversifying entrants if greater for firms that enter early rather than late. Much research has also consider ed the relationship between firm survival and the timing of its entry into the new industry (e. g. , see the reviews in VanderWerf and Mahon 1997; Lieberman and Montgomery 1998). In new, technologically dynamic industries, early entrants generally have higher survival rates than later entrants (e. g. , Christensen, et al. 1998; Sorenson 2000). Based on the product-level information from IDC, a firm’s Entry Time into the personal computer industry is defined to be the year in which the firm first sold a personal computer (less 1973). Following the well-established organizational ecology literature, we also include several firm and industry controls in our analyses (e. g. , Carroll and Hannan 2000; Carroll and Khessina 2005). Because smaller firms typically have higher hazards of exit due to their capability and resource constraints, we include a variable Firm Size (measured as the firm’s personal computer unit sales in the prior year divided by 10,000). Firm tenure in the new industry is also an important explanatory variable, so we include Firm Age (measured as the number of years the firm participated in this industry) and its square (to capture any nonlinear effects). 14 The theory of density dependence is based on the two contrasting effects of legitimization and competition. Firm survival is presumed to be a U-shaped function of firm density: as the number of firms in a new industry initially increases, the hazard rates of all firms decline due to the legitimacy signal associated with more firms engaged in the same new industry; but at higher levels of firm density, resources become thin, the competitive effects intensify and hazard rates increase. Of course, if the competitive effects in a new industry dominate, firm survival is simply an increasing function of firm density. Density dependence theory also predicts that competition at the time of a new firm’s founding is positively related to the hazard of exit (e. g. , resource scarcity and high selection pressures are associated with more competition). To account for these effects, we include Density (measured as the number of firms in the industry in the prior year), its square, and Density at Founding (measured as the number of firms in the industry in the year prior to the firm’s entry year). 4. 4 Product Strategies Two key aspects of a firm’s product strategy are important for testing the hypotheses in Section 3. First, information on whether or not a firm offers products incorporating the technology standard during its tenure in the personal computer industry is needed to test H1 and H3. To capture the effects of a product standard in the personal computer industry, we define Offer Intel x86 to be 1 if the firm ever introduced a personal computer with a microprocessor involving the Intel x86 architecture, and 0 otherwise. Similarly, we also define Not Entering with Intel x86 to be 1 if the firm first enters the personal computer industry with a product not incorporating an Intel x86 microprocessor, and 0 otherwise. Second, information on whether or not a firm offers products with the most recent technology is required to test H2 and H4. We define Offer Most Recent Product Technology to be 15 1 if the firm ever introduced a personal computer using the most recent microprocessor technology at the time, and 0 otherwise. Similarly, we also define Not Entering with the Most Recent Product Technology to be 1 if the firm does not initially enter the personal computer industry with the most recent technology, and 0 otherwise. See Figure 3 for the dynamically changing definition of â€Å"most recent technology. † 5. Estimation Methodology Because firms could exit at any point during the year, the actual underlying hazard rates are continuous time. The IDC data were only updated annually, however, and thus the year of exit can be determined but not the exact month or day. Therefore, discrete time survival models are most appropriate for our empirical study. Following Jenkins (2005), the survivor function at the beginning of the tth interval is: S( t ? 1) = Pr(T gt; t ? 1) = 1 ? F ( t ? 1) (1) Here, the length of survival in the new industry is a realization of a continuous random variable T, and the failure function is F ( t ) = Pr(T ? ) . Let us assume that the underlying continuous time model is summarized by the hazard rate ? ( t , X ) , where t is firm age and X is a vector of independent variables (some of which may be time varying). The survivor function at the end of the tth interval is: S ( t , X ) = exp[ ? ? ? (? , X ) d? ] 0 t (2) We will also assume that the hazard rate satisfies the p roportional hazard specification: ? ( t , X ) = ? 0 ( t ) e ? X = ? 0 ( t ) ? (3) Together, (2) and (3) imply that: S ( t , X ) = exp[ ? ? ? 0 (? ) ? d? ] = exp[ ? ? 0 (? ) d? ] = exp[ H t ] 0 0 t t (4) 16 Here, Ht is the integrated baseline hazard evaluated at the end of the interval, and thus the baseline survivor function at age t is S 0 ( t ) = exp( ? H t ) . The discrete time hazard function (i. e. , the probability of exit in interval t, conditional on surviving up to the beginning of interval t) is then: h( t , X ) = 1 ? S( t , X ) = 1 ? exp[ ? ( H t ? 1 ? H t )] S( t ? 1, X ) (5) This further implies that: log(1 ? h( t , X )) = ? ( H t ? 1 ? H t ) (6) and thus: log( ? log[1 ? h( t , X )]) = ? X + log( H t ? H t ? 1 ) (7) Similarly, the discrete time baseline hazard for the tth interval is: 1 ? h 0 ( t ) = exp( H t ? 1 ? H t ) (8) and hence: log( ? log[1 ? h 0 ( t )]) = log( H t ? H t ? 1 ) = ? ( t ) (9) In our analyses, we use a flexible, non-monotonic quadratic functional form for ? (t). Together, (7) and (9) give the discrete time (interval) hazard rate function we employ: log( ? log[1 ? h( t , X )]) = ? X + t + t 2 (10) Here, the log(-log(†¢)) transformation is known as the complementary log-log transformation and the discrete time proportional hazards model in (10) is referred to as a cloglog model. For estimation purposes, we use the cloglog procedure implemented in STATA 9. 0 (along with the Huber/White robust variance correction for repeated observations within firms). The extensive robustness checks we performed are discussed in a later section. 17 6. Estimation Results Estimation results for the corporate demographic variables, as well as entry time and pre-entry experience, are in Table 1. Table 2 contains the hazard model estimation results for the product strategies, and Table 3 reports the results for the conditioning effects of entry time and pre-entry experience. We discuss each in turn. [insert Table 1 about here] Across all our analyses, the results for the corporate demographic variables are as expected. The personal computer industry is generally characterized by a strong competitive environment (the linear Density term is positive and significantly larger than Density2 or Density at Founding). The significant coefficient estimates for the linear and quadratic Firm Age terms imply that firms are subject to a liability of obsolescence, i. e. in dynamically changing industries, firms’ initially successful alignment with its founding environment erodes with the passage of time due to structural inertia and the inability to make necessary adjustments (Barron, et al. 1994; Carroll and Hannan 2000; Carroll and Khessina 2005). This is also in line with the notion that in markets facing continuous technological change, vintage effects associated with older technology offset the benefits of experience, and that the inertia tendencies of old er firms overshadow any learning by doing effects (e. g. Bahk and Gort 1993; Jovanovi and Nyarko 1996). As firms age, they have to navigate more technology transitions and are thus subjected to higher risks of failure. This age effect, however, is counter-balanced by firm size: Firm Size is negatively related to exit and significant, i. e. , firms that achieve and maintain a high level of sales in the new industry tend to have higher survival rates. As expected, the estimated coefficient for Entry Time is positive and significant in Table 1. This implies that early entrants in the personal computer industry generally have 8 higher survival rates than later entrants. Although pre-entry experience in Table 1, Model 3 (without the Entry Time interaction term) is insignificant, the more complete results in Model 4 are as expected. The highly significant and positive estimate for Startup, combined with the negative interaction involving pre-entry experience and Entry Time, indicates that diversifying entrants have an initial survival advantage over startups in the personal computer industry, but this advantage diminishes for later entrants. Moreover, the results in Model 3 suggest that the interaction effects between pre-entry experience and entry time need to be included in the analysis to identify the true underlying effects. [insert Table 2 about here] Strong support for H1 and H2 comes from the results in Table 2. The negative and highly significant coefficient estimates for Offer Intel x86 and Offer Most Recent Product Technology indicate that these product strategies are associated with higher survival rates. In other words, firms offering products with the (eventual) technology standard and products with the latest available technology have higher survival rates than firms not pursuing these strategies 7 . Importantly, the estimates in Model 3 also demonstrate that the effects of preentry experience disappear once controls for product strategies are included. These results are very revealing over the existing literature—indicator variables for pre-entry experience may only be proxying for the real underlying product strategies directly related to firm survival. [insert Table 3 about here] Separate statistical analyses available from the authors indicate that this conclusion does not vary across firms by pre-entry experience or entry cohort. 19 To parsimoniously test H3 and H4, we split our sample into â€Å"early† and â€Å"late† entrants based on Entry Time. We use 1985 as the dividing year for two reasons 8 : (1) from the results in Table 1, the survival advantage of diversifying entrants over startups reverses around year twelve (which translates into 1973+12=1985), and (2) the start of the fourth generation product technology (32-bit technology) occurs in 1985. A set of mutually exclusive dummy variables involving their product strategy and pre-entry experience were constructed to directly test for any differences between entrepreneurial startups and diversifying entrants (since the results in Table 2 indicate that firms implementing these product strategies have higher survival rates, we define the reference category to be firms, startup or diversifying entrant, that did implement the product strategy at entry). The results in Table 3, Model 1 (and Model 5) strongly support H3. The positive and highly significant coefficient estimate of Startup Not Entering with Intel x86 indicates that the survival of entrepreneurial startups in the early years of the personal computer industry depends on whether they enter the industry with the (eventual) product standard. In this case, startups not entering with the technology standard have higher exit hazards than diversifying entrants not entering with the standard. We provide evidence for discriminant validity of this finding by demonstrating that this same result does not hold for later entrants in Model 2 (and Model 6). We argue that this result is due to the firms’ product strategies involving the emerging technology standard. Before 1985, two-thirds of the entrants did not enter with a personal computer incorporating the Intel x86 CPU. Of these entrants, 59% of the diversifying entrants later migrated to the Intel x86 architecture while only 34% of the startups did so (this difference is statistically significant at the 0. 05 level). After 1984, not surprisingly the vast majority of firms entered with a product involving the Intel standard 8Similar results are obtained for other reasonable cut-points. 20 95% of the startups and 87% of the diversifying entrants). Interestingly, 6% of the later entering diversifying entrants never introduced a personal computer incorporating the Intel standard; not surprisingly, these firms had high hazards of exit (e. g. , see Table 3, Model 2). The results in Table 3, Model 4 (and Model 6) strongly support H4. The positive and significant coefficient estima te of Diversifying Entrant Not Entering with the Most Recent Product Technology indicates that the survival of later entering diversifying entrants depends on whether they enter the personal computer industry with the latest available product technology. In this case, diversifying entrants not entering with the latest available product technology standard have higher exit hazards than startups not entering with this technology. Again, we provide evidence of discriminate validity for this finding by showing that this same result does not hold for early entrants in Model 3 (and Model 5). We believe that this result is due to firms’ product technology strategies. After 1984, entrepreneurial startups are the firms most likely to stay close to the technology frontier (e. g. 76% of startups introduced at least one personal computer with the most recent technology, whereas only 67% of diversifying entrants did so). On the other hand, diversifying entrants were more likely than startups to introduce personal computers with the latest technology before 1985. This behavior is entirely consistent with the higher growth requirement of diversifying entrants—unlike in the later stages of this industry, sales of the latest generation products are very close to those of the â€Å"older† technology products in the years before industry sales took off (see Figure 3). . Robustness Checks Several robustness checks were undertaken to confirm the empirical results reported in the previous section. These analyses are briefly discussed here. First, other discrete time survival formulations like the logistic hazard model gave very similar results to those 21 presented in Tables 1, 2 and 3. Second, we considered the possibility of unobserved heterogeneity among firms by fitting frailty models with normal and gamma parameter distributions (Jenkins 2005); no evidence of unobserved heterogeneity was found. Third, we considered whether our main results are strongly influenced by the activities of a small number of key players in the personal computer industry. Discrete time hazard models estimated without IBM, Apple, Dell, and HP (e. g. , as in Bresnahan, et al. 1997) were very similar to those already discussed and thus our conclusions remain unchanged. Fourth, we were able to confirm our results and conclusions when market share was the performance measure rather than firm survival. Random-effects panel regression models of market share that parallel the models in Tables 1, 2, and 3 gave similar results to those already discussed. Finally, the sensitivity of our reported results to the type of pre-entry experience was explored. Following Steffens (1994), the Thomas Register information on the primary line of business was used to classify the pre-entry experience of firms into technical experience and market experience. Firms with prior technical experience include those in related productmarkets (e. g. , mainframe or minicomputers, video games, typewriters, business machines) and/or technology-markets (e. g. , microprocessors or semiconductors). Examples of firms in our data set with prior technical experience include IBM, Hewlett Packard, and Texas Instruments. Firms with prior market experience include those with knowledge of the potential customers for personal computers (e. g. , retailers, consultants, manufacturers of peripherals). Prominent examples include Tandy/Radio Shack, NCR and IBM. Few firms had both technical and marketing experience. While firms with technical or marketing experience initially had higher survival rates than entrepreneurial startups, not surprisingly firms with technical experience had lower hazards of exit than firms with marketing experience. Separate hazard models were also estimated with diversifying entrants restricted 2 to only be firms with prior technical experience or firms with prior marketing experience. Again, our general results are consistent with those already presented, and thus our major conclusions are unaffected. 8. Implications and Conclusions We started with three questions that guided our research. What product strategies increase the survival chances of entrants into new, technologically dynamic industries? Does the effectiveness of these product strategies differ by pre-entry experience? Does the effectiveness of these product strategies differ by when firms enter a new industry? We find that successful product strategies in the personal computer industry involve migrating to the eventual technology standard and staying close to the advancing technology frontier. Moreover, we find that the effectiveness of these product strategies depends on who implements it (pre-entry experience) and when it is employed (entry timing). In particular, diversifying entrants are better able than entrepreneurial startups to migrate to the technology standard when it becomes known, and consequently, they enjoy higher survival rates in the early years of this new industry. For later entrants however, startups are more likely than diversifying entrants to introduce products with the latest available technology, and thus they tend to have higher survival rates in the later years. In agreement with the existing literature, we find that diversifying entrants have an initial survival advantage over entrepreneurial startups. But, we find the opposite for later entrants: startups that enter later in the industry have a survival advantage over the later entering diversifying entrants. To explain this result, we empirically demonstrate that survival depends on the product strategies implemented by firms and that the success of these strategies is conditioned by pre-entry experience and entry timing. Importantly, we empirically demonstrate that the effects of pre-entry experience on firm survival vanish 23 when controls for product strategy are included in the analysis. Our findings highlight that it is crucial to study what firms do after they enter a new industry in order to more completely understand their ultimate performance. Unlike the findings for the television industry reported by Klepper and Simons (2000), our results suggest that a â€Å"dominance by birthright† does not exist in the personal computer industry. In other words, early entering diversifying entrants do not always have a survival advantage over other entrants. As already noted, our finding that firm survival is significantly related to firms’ product strategies after they enter a new industry indicates the important role of post-entry activities. In addition, our estimates of the corporate emographic effects in the personal computer industry suggest that the survival rates of later entrants can surpass those of the early entering diversifying entrants. Although firm size tempers the effects, the positive coefficient estimates for firm age (liability of obsolescence) and entry time indicate that the exit hazards of young, later entrants can be lower than those of old, early entrants (under some conditions). These results su pport our general approach of studying the product strategies of later entrants as well as early entrants. Our study has the usual set of limitations. In particular, studies of other industries need to be undertaken before our results for the personal computer industry can be generalized. Although new datasets may be needed, such efforts will move us closer to a more complete theory of firm behavior and survival dynamics. Following prior research, we used a single dummy variable for pre-entry experience to examine effects of firm heterogeneity at time of entry; future research could include time-varying and continuous measures of experience within the focal industry as well as in other (diversified) industries. We also used static dummy variables for the product strategies employed by firms; future research could further explore firm heterogeneity in implementing common product 24 strategies (e. g. , product line length, mix of advanced technology and more popular products), as well as the timing of implementation (e. g. , is it advantageous to anticipate the technology standard? ). Like most research in this stream, we lack appropriate data to study the entry selection question; future research is clearly needed to link the firm’s entry decision to their eventual performance (e. g. , as in Klepper and Simons 2000). While our study focuses on pre-entry experience embodied in organizations, additional research is also needed on how pre-entry experience possessed by managerial teams may affect firm performance. Recent research emphasizes the spin-out phenomenon (i. e. , entrepreneurial start-ups with pre-entry experience due to the prior employment of its founders with an incumbent firm; e. g. , Agarwal, et al. 2004). Spin-outs seem to have superior performance relative to all other entrants. Importantly, experienced managers of entrepreneurial startups may mitigate the advantage of diversifying entrants. While we lack systematic data on founding teams of entrepreneurial start-ups, the personal computer industry includes cases of successful spin-out firms such as Compaq and Gateway, along with several entrepreneurial start-ups that did not benefit from pre-entry experience through founder affiliation who also performed well (e. g. , Apple, Dell). Extending our study in this direction should lead to an even greater understanding of the role of product strategies in technologically dynamic industries. 25 References Agarwal, R. , A. Franco, R. Echambadi, and M. B. Sarkar (2004), â€Å"Knowledge Transfer Through Inheritance: Spinout Generation, Development and Survival,† Academy of Management Journal, 47 (4), 501-522. Agarwal, R.. and B. L. Bayus (2002), â€Å"The Market Evolution and Take-off of New Product Innovations,† Management Science, 48(5), 1024-41. Anderson, P. (1995), â€Å"Personal Computer Manufacturers,† in G. Carroll and M. Hannan (eds. ), Organizations in Industry: Strategy, Structure, and Selection, New York: Oxford University Press, 37-58. Bahk, B. and M. Gort (1993), â€Å"Decomposing Learning by Doing in New Plants,† Journal of Political Economy, 101 (4), 561-83. Barnett, W. and W. Freeman (2001), â€Å"Too Much of a Good Thing? Product Proliferation and Organizational Failure,† Organization Science, 12(3), 539-558. Barron, D. , E. West, and M. Hannan (1994), â€Å"A Time to Grow and a Time to Die: Growth and Mortality of Credit Unions in New York, 1914-1990,† American Journal of Sociology, 100, 381-421. Baum, J. , H. Korn and S. Kotha (1994), â€Å"Dominant Designs and Population Dynamics in Telecommunications Services: Founding and Failure of Facsimile Service Organizations 19651992,† Social Science Research, 24, 97-135. Bayus, B. L. (1998), â€Å"An Analysis of Product Lifetimes n a Technologically Dynamic Industry,† Management Science, 44 (June), 763-775. Bresnahan, T. , S. Stern and M. Trajtenberg (1997), â€Å"Market Segmentation and the Sources of Rents from Innovation: Personal Computers in the Late 1980s,† RAND Journal of Economics, 28, S17-S44. Brush, T. , P. Bromiley, and M. Hendrickx (2000), â₠¬Å"The Free Cash Flow Hypothesis for Sales Growth and Firm Performance,† Strategic Management Journal, 21, 455-472. Burgelman, R. (1994), â€Å"Fading Memories: A Process Theory of Strategic Business Exit in Dynamic Environments,† Administrative Science Quarterly, 39, 24-56. Busenitz, L. nd J. Barney (1997), â€Å"Differences Between Entrepreneurs and Managers in Large Organizations: Biases and Heuristics in Strategic Decision Making,† Journal of Business Venturing, 12(1), 9-30. Carroll, G. , L. Bigelow, M. Seidel, and L. Tsai (1996), â€Å"The Fates of De Novo and De Alio Producers in the American Automobile Industry: 1885-1981,† Strategic Management Journal, 17, 117-137. Carroll, G. and M. Hannan (2000), The Demography of Corporations and Industries, Princeton, NJ: Princeton University Press. Carroll, G. and O. Khessina (2005), â€Å"Organizational and Corporate Demography,† in D. Poston and M. Micklin (eds. ), Handbook of Population, Berlin, Germany: Springer, 451-478. Christensen, C. , F. Suarez, and J. Utterback (1998), â€Å"Strategies for Survival in Fast-Changing Industries,† Management Science, 44(12), S207-S220. Curry, J. and M. Kenney (1999), â€Å"Beating the Clock: Corporate Responses to Rapid Change in the PC Industry,† California Management Review, 42(1), 8-36. David, P. and S. Greenstein (1990), â€Å" The Economics of Compatibility Standards: An Introduction to Recent Research,† Economics of Innovation and New Technology, 1(1-2), 3-41. 26 Dowell, G. and A. Swaminathan (2000), â€Å"Racing and Back-Pedalling into the Future: New Product Introduction and Organizational Mortality in the US Bicycle Industry: 1880-1918,† Organization Studies, 21(2), 405-431. Evans, D. , A. Nichols, and B. Reddy (1999), â€Å"The Rise and Fall of Leaders in Personal Computer Software,† NERA working paper. Fosfuri, A. and M. Giarratana (2004), â€Å"Product Strategies and Startups’ Survival in Turbulent Industries: Evidence from the Security Software Industry,† Universidad Carlos III de Madrid working paper. Gabel, L. (1991), Competitive Strategies for Product Standards, London, UK: McGraw-Hill. Gort, M. and S. Klepper (1982), â€Å"Time Paths in the Diffusion of Product Innovations,† Economic Journal, 92(3), 630-653. Hannan, M. and J. Freeman (1984), â€Å"Structural Inertia and Organizational Change,† American Sociological Review, 49, 149-164. Helfat, C. and M. Lieberman (2002), â€Å"The Birth of Capabilities: Market Entry and the Importance of Pre-History,† Industrial and Corporate Change, 11(4), 725-760. Iansiti, M. and K. Clark (1994), â€Å"Integration and Dynamic Capability: Evidence From Product Development in Automobiles and Mainframe Computers,† Industrial and Corporate Change, 4, 557605. Jenkins, S. 2005), â€Å"Survival Analysis,† University of Essex Lecture Notes, July 18, 2005, www. iser. essex. ac. uk/teaching/degree/stephenj/ec968/pdfs/ec968lnotesv6. pdf (accessed 1/25/06). Jones, N. (2003), â€Å"Competing After Radical Technological Change: The Significance of Product Line Management Strategy,† Strategic Management Journal, 24, 1265-1287. Jovanovic, B. and Y. Nyarko (1996), â€Å"Learning by Doing and the Choice of Technology,† Econometrica, 64 (6), 1299-1310. Khessina, O. and G. Carroll (2004), â€Å"Product Dynamics of De Novo and De Alio Firms in the Worldwide Optical Disk Drive Industry, 1983-1999,† Georgetown University working paper. Klepper, S. (1996), â€Å"Entry, Exit, Growth, and Innovation Over the Product Life Cycle,† American Economic Review, 86(3), 562-583. Klepper, S. (2002), â€Å"Firm Survival and the Evolution of Oligopoly,† Rand Journal of Economics, 33(1), 3761. Klepper, S. and K. Simons (2000), â€Å"Dominance by Birthright: Entry of Prior Radio Producers and Competitive Ramifications in the U. S. Television Receiver Industry,† Strategic Management Journal, 21(10-11), 997-1016. Klepper, S. and K. Simons (2005), â€Å"Industry Shakeouts and Technological Change,† International Journal of Industrial Organization, 23, 23-43. Langlois, R. (1992), â€Å"External Economies and Economic Progress: The Case of the Microcomputer Industry,† Business History Review, 66 (Spring), 1-50. Lavin, M. (1992), Business Information: How to Find It, How to Use It. Phoenix: Oryz Press. Lieberman, M. and D. Montgomery (1998), â€Å"First-Mover (Dis)Advantages: Retrospective and Link with the Resource-Based View,† Strategic Management Journal, 19 (12), 1111-1125. Mishna, Y. , T. Pollock, and J. Porac (2004), â€Å"Are More Resources Always Better for Growth? Resources Stickiness in Market and Product Expansion,† Strategic Management Journal, 25, 1179-1197. 27 Mitchell, W. (1991), â€Å"Dual Clocks: Entry Order Influences on Incumbent and Newcomer Market Share and Survival When Specialized Assets Retain Their Value,† Strategic Management Journal, 12(2), 85-100. Meyer, M. and E. Roberts (1986), â€Å"New Product Strategy in Small Technology-Based Firms: A Pilot Study,† Management Science, 32(7), 806-821. Penrose, E. (1959), The Theory of the Growth of the Firm, Oxford: Oxford University Press. Pfeffer, J. and G. R. Salancik (1978), External Control of Organizations, New York: Harper and Row. Romanelli, E. 1989), â€Å"Environments and Strategies of Organization Start-Up: Effects on Early Survival,† Administrative Science Quarterly, 34(3), 369-387. Schilling, M. (1998), â€Å"Technological Lockout: An Integrative Model of the Economic and Strategic Factors Driving Technology Success and Failure,† Academy of Management Review, 23(2), 267-284. Sorenso n, O. (2000), â€Å"Letting the Market Work For You: An Evolutionary Perspective on Product Strategy,† Strategic Management Journal, 21(2), 577-592. Sorenson, O. , S. McEvily, C. Ren, and R. Roy (2005), â€Å"Niche Width Revisited: Organizational Scope, Behavior, and Performance,† London Business School working paper. Steffens, J. (1994), New Games: Strategic Competition in the PC Revolution, New York: Pergamon Press. Stern, I. and A. Henderson (2004), â€Å"Within-Business Diversification in Technology-Intensive Industries,† Strategic Management Journal, 25, 487-505. Stinchcombe, A. (1965), â€Å"Social Structure and Organizations,† in J. G. March (ed. ), Handbook of Organizations, Chicago: Rand McNally, 153-193. Swanson, A. (2002), Form Coherence and the Fates of De Alio and De Novo Organizations in the U. S. Digital Computer Industry: 1951-1994, Ph. D. Dissertation, Stanford University. Teagarden, L. D. Hatfield, and A. Echols (1999), â€Å"Doomed From the Start: What is the Value of Selecting a Future Dominant Design? † Strategic Management Journal, 20, 495-518. Teagarden, L. , A. Echols and D. Hatfield (2000), â€Å"The Value of Patience and Start-Up Firms: A ReExamination of Entry Timing for Emerging Markets, Entrepreneurship Theory and Practice, 24(4), 41-48. Teece, D. (1986), â€Å"Profiting from Technological Innovation: Implications for Integration, Collaboration, Licensing, and Public Policy,† Research Policy, (15), 285-305. The PC Tech Guide (2004), â€Å"Historical Perspective,† http://www. pctechguide. com/02procs_Historical_perspective. htm, accessed 2/20/06. Tripsas, M. (1997), â€Å"Unraveling the Process of Creative Destruction: Complementary Assets and Incumbent Survival in the Typesetter Industry,† Strategic Management Journal, 18 (Special Issue), 119-142. Tushman, M. and P. Anderson (1986), â€Å"Technological Discontinuities and Organizational Environments,† Administrative Science Quarterly, 31(3), 439-465. Utterback, J. (1994), Mastering the Dynamics of Innovation, Boston: Harvard Business School Press. VanderWerf, P. and J. Mahon (1997), â€Å"Meta-Analysis of the Impact of Research Methods on Findings of First-Mover Advantage,† Management Science, 43 (November), 1510-1519.

No comments:

Post a Comment

Note: Only a member of this blog may post a comment.