Digital Marketplaces Unleashed
Page 31
Second, the development and diffusion of new technologies takes time and requires a multitude of incremental and cumulative innovations. But these processes are not smooth: turbulence and ultimately disruption occurs by impulses, which are typically industry‐specific [7].
Third, however, that there is no such thing as “technological innovation”, but diverse and variegated forms of innovation in different industries according to the specific nature and characteristics of the relevant technologies and markets.
Thus, it should come as no surprise that many representations of the innovative process have been proposed, sometimes complementary to and sometimes conflicting with each other. Against this background, the concept of disruptive innovation was introduced by Clayton Christensen in 1995 [8] and received quickly an enormous popularity. This idea adds significant nuances and try at the same time to generalize older insights about “creative destruction”. A widely used and workable definition can be found in Wikipedia: an innovation that creates a new market and value network and eventually disrupts an existing market and value/support network, displacing established market leaders and alliances. This definition highlights the main substantial features of the concept. First, the adjective “disruptive” is applied to innovations rather than to technologies, because few technologies are intrinsically disruptive (or sustaining); rather, the disruptive character of the innovation is linked to the business model that the new technology enables. Second, emphasis is attributed to the disruption of the value/support network of a company or an industry, that is to say to the set of relationships with customers and suppliers, rather than to the ability of extant market leaders to absorb and master the new required technological capabilities. Thus, “…, disruptive innovations were technologically straightforward, consisting of off‐the‐shelf components put together in a product architecture that was often simpler than prior approaches. They offered less of what customers in established markets wanted and so could rarely be initially employed there. They offered a different package of attributes valued only in emerging markets remote from, and unimportant to, the mainstream.” (Christensen 1997, p. 15). In particular, disruption is more likely to occur starting from market niches where customers do not need the full performance valued by customers at the high end of the market and/or in new markets previously unserved by the products supplied by existing incumbents Hence, third, disruptive innovations tend to be introduced by outsiders and entrepreneurs, because market leaders downplay the potential threat of these innovations, considering them less profitable than current products and absorbing resources from the current businesses.
Unsurprisingly, the notion of disruptive innovation triggered controversies and debates. Some criticism claims that there is nothing inherently new in the concept or that it is only a refinement or a particular case of broader processes of creative destruction. Others have pointed that the evidence on which the theory is based is actually made out of a few, debatable case studies (For reviews, see [9] and [10–13]).
In what follows, this article will not review the details of this debate. Rather, some more basic issues are discussed about the intensity and forms of and the strategies and reactions of incumbents to the threats presented by new technologies. Different explanations are available but they often lack generality and in most instances theories rely on a quite narrow set of specific cases of particular firms, products and industries.
In effect, our knowledge of how creative destruction/disruptive innovation1 occurs is still limited: it is possible to cite many examples of disruption, where industry leaders were actually swiped away by new competitors (Kodak in digital photography, Nokia for a while in cellular phones, etc.) as well as many cases where industry leaders were able to maintain or even increase their dominance (Fuji in digital photography, big pharma companies vs. new biotechnology firms, etc.). Sweeping generalizations are hard to make in this context, much depending on the industry as well as on the specific characteristics of individual firms. To be sure, the destruction of dominant positions – especially by new entrants – is a much less frequent and in any case a much more nuanced phenomenon than is often assumed.
Indeed, some very basic questions remain open and difficult to resolve, due both to limitations in the data and sometimes in less than robust conceptual clarity about the specific phenomena that are analyzed.
For example: (i)How frequent and how strong is actually “creative destruction”, and how long does it take? How often do we observe dramatic changes in industry leadership? Is disruption a continuous, systematic process? Or is it a punctuation over the history of any one industry or product?
(ii)Where does creative destruction come from and where does it occur? Are always (or most of the time) new firms that introduce disruptive innovations?
(iii)How does disruption occur? Through direct competition and head‐on attack on the products of industry leaders or indirectly, via the introduction of products, processes or business models in different market segments that progressively weaken dominant positions?
(iv)When and how industry leaders able to maintain their leadership despite the appearance of potentially disruptive innovations?
22.2 The Aggregate Background
To begin with, it might be useful to recall a few important and robust aggregate results that have been emerging from empirical studies made possible from the growing availability of data at the firm level for sufficiently long periods of time. These results help in providing a broader perspective to the analysis of creative destruction/disruptive innovation (see [14] for a survey).
First, relatively high rates of entry of new firms are seen in virtually all industries, even those marked by high capital intensity and other apparent barriers to entry. Further, and contrary to what standard economic textbooks would suggest, rates of entry do not appear to be particularly sensitive to the average rate of profit in an industry. And in most industries there is considerable exit as well as entry. Indeed, exit and entry rates tend to be strongly correlated. Both entry and exit tend to be significantly higher in new industries, and to decline somewhat as the industry matures. However, even relatively mature industries often are marked by continuing entry and exit.
Second, the vast majority of entrants are small firms, and most of them exit the industry within a few years: 20–40% of entrants die in the first two years and only 40–50% survive beyond the seventh year in a given cohort. Survivors grow faster but more erratically than incumbents and they reach average levels of productivity only gradually and slowly over time (around a decade). Only a few outliers in an entry cohort are able to attain superior performances, but, especially in the presence of significant technological and market discontinuities, they sometimes displace the incumbents and become the new industry leaders. Even in relatively mature industries one often observes persistent turbulence and churning in the profile of industrial evolution, due not only to continuous entry and exit flows but also to changes in the incumbents’ market shares.
Third, even in mature industries there tends to be persistent heterogeneity among firms regarding any available measure of firms’ traits and performance: size, age, productivity, profitability, innovativeness, etc. (For an overview, see [15]). The distributions of these variables tend to be highly asymmetric, and often display fat tails in their rates of change. What is even more interesting though, is that heterogeneity is persistent: more efficient firms at time t have a high probability to be highly efficient also at time t+T, and the same applies for size, profitability, and (more controversially) innovation. The degree of persistence tends to decline the longer the time span considered. However, this tendency is weak and thus heterogeneity decays slowly and it is still present in the limit.
Fourth, positive relationships are typically found among these variables: more
efficient firms tend to be also more innovative and profitable and to gain market shares as time goes by. The magnitude of these relationships, however, is extremely variable across samples and across industries. Thus, most studies find only weak or no relationship at all between productivity and profitability on the one hand, and growth on the other. Firms’ expansion appears to be independent from size, possibly with smaller companies exhibiting higher but more variable growth rates. And in general, firms’ growth remains very hard to explain. While some studies describe it as driven by small, idiosyncratic, and independently distributed shocks – and therefore as essentially erratic – others find highly complex underlying structures. If anything, the evidence would seem to suggest that firms grow and decline by relatively lumpy jumps which cannot be accounted by the accumulation of small, “atom‐less”, independent shocks. Rather “big” episodes of expansion and contraction are relatively frequent. (For an overview, see [14].)
Sixth, further important results are offered by studies which decompose aggregate (sectoral or economy‐wide) productivity growth, separating (i) idiosyncratic changes in firm/plant productivity levels – the so called within component – that broadly captures improvements occurring within incumbent firms; (ii) changes in average productivity due to reallocation of output or employment shares across firms – the between component – that imperfectly measures the impact of market selection in shifting resources to the more efficient firms; and (iii) the contribution thereof due to entry into and exit from the market. Summarizing heroically, most studies do indeed find further evidence of a steady process of creative destruction involving significant rates of input and output reallocation even within 4‐digit industries. Again these studies confirm that the process is accompanied by a good deal of “churning” with relatively high flows of entry and exit. However, the most interesting finding is that the largest contribution to productivity growth comes by far from the within component, that is to say from the learning processes of existing firms. The role of the between component – market selection – is much smaller and in some cases it has even a negative sign. Last, as already mentioned above, the contribution of net entry is highly variable, possibly with exit rather than entry showing a larger impact.
These findings suggest that heterogeneous processes of learning and selection drive industry dynamics, generating a puzzling coexistence of remarkable stability and drastic change. Moreover, continuous change and turbulence and permanent differences among firms coexist with the emergence of remarkably stable structures at higher levels of aggregation. However, the strength, speed and directions of these processes vary significantly across sectors and countries.
22.3 Sectoral Patterns of Innovation and Industrial Dynamics
Indeed, while some of these aggregate properties of the processes of industrial evolution are common to most industries, still fundamental differences appear across sectors. Thus, for example, [16] found that while innovative firms are likely to be rather small in industrial machinery, big companies prevail in chemicals, metal working, aerospace and electrical equipment, and many “science‐based” sectors (such as electronics and pharmaceuticals) tend to display a bimodal distribution with high rates of innovativeness associated to small and very large firms.
Analyses have increasingly emphasized the relevance of various factors that impact the patterns of innovation and industrial dynamics. To begin with, it is now acknowledged that technology often develops according to its own internal logic, following trajectories that are only partially responsive to market signals [17]. Moreover, there is no such thing as “technology in general” but rather an array of different technologies, with different properties and characteristics, yielding different patterns of technological advance [18]. Technologies differ in terms of opportunities for innovation, and in terms of the degree of appropriability of its benefits. Including measures of these variables in the analysis (either statistical or qualitative) almost always improves results. Typically technological change proceeds cumulatively: creative accumulation rather than creative destruction is the norm in many industries and over relatively long periods of time. Yet, in some technologies and industries – pharmaceuticals being a clear example – it is harder to use cumulated knowledge to develop new products and processes. This difference has implications for the evolution of industry structure. In some industries, largely public or semi‐public organizations produce much of the relevant knowledge base on which innovation depends, which is in principle available to everybody who has the requisite scientific and technological absorptive capabilities. In other cases, technological advances do not rely much on publicly available knowledge, but on private and firm‐specific know‐how and expertise. Clearly, innovation can arise in and impact on very different industry structures.
The well‐known taxonomy by Keith Pavitt [18] was a first and still invaluable attempt at mapping ‘industry types’ and industry dynamics. Pavitt taxonomy comprises four groups of sectors, namely: (i)‘supplier dominated’, sectors whose innovative opportunities mostly come through the acquisition of new pieces of machinery and new intermediate inputs (textile, clothing, metal products belong to this category);
(ii)‘specialized suppliers’, including producers of industrial machinery and equipment;
(iii)‘scale intensive’ sectors, wherein the sheer scale of production influence the ability to exploit innovative opportunities partly endogenously generated and partly stemming from science based inputs;
(iv)‘science based’ industries, whose innovative opportunities co‐evolve, especially in the early stage of their life with advances in pure and applied sciences (microelectronics, informatics, drugs and bioengineering are good examples).
Other, rather complementary, taxonomic exercises have focused primarily on some characteristics of the innovation process, distinguishing between a ‘Schumpeter Mark I’ and a ‘Schumpeter Mark II’ regime, dramatizing the difference between the views of innovative activities from Schumpeter (1911) and Schumpeter (1942): see [19–21].
As mentioned previously, Schumpeter himself distinguished two (extreme) patterns of innovation. In the first one, as theorized in The Theory of Economic Development (1911) and often labeled as Schumpeter Mark I [22], innovation is created by the bold efforts of new entrepreneurs, who are able and lucky enough to displace incumbents, only to be challenged themselves by imitative entrants. At the other extreme, as described in Capitalism, Socialism and Democracy (1942) and often referred to as Schumpeter Mark II, the main sources of innovation are instead large corporations, which accumulate difficult‐to‐imitate knowledge in specific domains, and are therefore able to gain long‐lasting and self‐reproducing technological advantages (and economic leadership). Following this intuition, the notion has been developed that innovation and market structure evolve according to different technological regimes [23]. distinguished between science‐based vs. cumulative regimes [24]. Further developed this concept by modeling the different evolution of industries under an “entrepreneurial” as opposed to a “routinized” regime [25] and [20]. provided further empirical evidence concerning the relationships between the properties of technologies, the patterns of innovation, and market structure.
More specifically, a technological regime may be defined by the combination of some fundamental properties of the relevant technology, namely the degree of opportunities for innovating; the degree of appropriability, i. e. the ease and the instruments by which innovators are able to appropriate the economic benefits stemming from innovation; the degree of cumulativeness of innovation, i. e. the extent to which innovators today enjoy cognitive and/or economic advantages vis‐a‐vis competitors that make it more likely to innovate again the future.
Thus, the Mark I regime is characterized by high opportunities, low appr
opriability conditions and low cumulativeness: innovations are therefore carried out to a good extent by innovative entrants who continuously challenge incumbents. Market structure is highly unstable, with leadership changing frequently. Examples might be biotechnology, mechanical engineering, furniture, etc.. Conversely, at the other extreme, under the Mark II regime innovative activities are much more cumulative and imitation is difficult. Innovation is therefore undertaken to a greater extent by a few incumbents which turn out to be ‘serial innovators’: chemical engineering, semiconductors, aerospace, etc..
The structure of the demand side of the market – the demand regime – plays also an important role in shaping the patterns of industrial dynamics. In particular, when the aggregate market is actually composed by a large number of (actual and potential) almost independent niches, it is more difficult for any one firm to build a dominant and persistent leadership in the aggregate market: pharmaceuticals is a classic example (see [21]).