University of Amsterdam
Faculty of Economics & Business
Bachelor Thesis Science - Business Studies
(Strategy)
Academic year 2013/2014
Innovations and evolution of the coffee grinding and
roasting industry in the period of 1978-2013
Student: Maria Saykina (10167420)
Supervisor: Ranjita M. Singh
4-07-2014
Table of Contents
1. Introduction ... 4
2. Literature review ... 6
2.1 Concept of innovation ... 7
2.2 Taxonomy of innovation ... 9
2.2.1 External and internal nature of innovation ... 9
2.2.2 Transilience map ... 9
2.2.3 Schumpeterian innovation activity ... 11
2.3 Technological change ... 13 2.4 Conclusion ... 15 3. Conceptual framework ... 16 4. Methodology ... 19 5. Results ... 24 5.1 Hypothesis 1 ... 24 5.2 Hypothesis 2 ... 27 5.3 Hypothesis 3 ... 28 5.4 Hypothesis 4 ... 29 5.5 Conclusion ... 30 6. Discussion ... 30 6.1 Hypothesis 1 ... 30 6.2 Hypothesis 2 ... 31 6.3 Hypothesis 3 ... 32 6.4 Hypothesis 4 ... 32 6.5 Industry-wide view ... 33
7. Limitations & Future Research ... 34
8. Managerial Implications ... 35
9. Conclusion ... 35
10. References ... 36
11. Appendix ... 38
The recent statistics of FMCG on various coffee consumption indicators shows clear trends of increasing roasted coffee popularity amongst consumers (Statista, Global Coffee Consumption). An evidence for that can be a widespread appearance of chains of coffee-restaurant (Statista, Number of establishments in the U.S. coffee and snack shops industry form 2002 to 2016). For instance, at the moment Starbucks is considered to be one of the most successful coffee companies operating in 62 countries as its total annual revenue has more than quadrupled in the last 10 years (Statista, Net income of Starbucks corporation from 2007 to 2013). However, the overall world consumption of coffee does not show an easily identifiable trend. If we refer to the statistics, we see that there were periods of quick growth in the beginning of the 1990's (78,169,000 coffee bags total in 1993) as well as declining periods as in the beginning of the 2000's (65,357,000 total in 2000; International Coffee organization statistics). Nevertheless, due to the International Coffee Agreement (ICA), which is ratified periodically, the total export and import is constantly limited by the quota agreement. This allows controlling the production and prices for coffee.
The birth of the coffee industry can be traced back to the 6th century. Since then it has truly spread across the world and at the time was the most valuable commodity. The processing procedure has not changed significantly throughout the time. It is planted, harvested, dried, milled, roasted, grinded and brewed (Coffee processing) following the same technological algorithm for decades with minor evidential technological advances. We see the evolution of coffee machines and in-house handling appliances, while the overall development of coffee grinding and roasting industry is far more complex.
Here it is necessary to stress the key importance of innovation concept in an industry. Any industry follows a certain path of technological growth since its birth (Christensen et al., 1998). This growth is a necessary process in order for an industry to prosper, to bring social benefits and profits to the producers. The innovation can refer to either product or process, thus, influencing the quality, efficiency and the cost of the first or the latter (Rosenberg, 1969). With regards to the coffee industry, the quality of grinding and roasting directly influences the final quality of coffee and, consequently, its price. Therefore, I expect that the coffee handling industry has developed parsimoniously with the coffee industry. There is little knowledge about the evolution of the grinding & roasting industry, which is why we would like to focus our research on distinguishing the development patterns. The research is theory-driven and it will contribute to the understanding of coffee industry development and development of the firms in the handling industry. The research question that evolves from the problem statement is following: what are the main patterns of evolution in the coffee
roasting and grinding industry in 1978-2013? To address this question and divide it into the sub-questions I will explore the topic of industry development and evolution patterns in the literature.
To start with the general division of innovations suggested by Malerba & Orsenigo (1996), there are two types: Schumpeterian Mark I and Schumpeterian Mark II. The first represents the widening of an industry while the latter refers to the deepening. By studying the events of innovation in the handling industry, it will be possible to identify any of the widening or deepening tendencies. In addition, to classify an industry as one of the types, Malerba and Orsenigo (1997) suggest the technological regime dimensions that correspond to each of the types. The technological regime is defined by Winter (1984) as the external condition that has an impact on the availability and the advance of relevant knowledge in an industry. The dimensions suggested by Malerba and Orsenigo (1997) correspond to the nature of innovation with regards to the: (i) concentration of innovative activity in an industry, (ii) asymmetries among firms, (iii) size of innovating activity, (iv) hierarchical stability of firms and (v) degree of entry and exit. In order to measure the dimensions Malerba and Orsenigo (1993) introduce the following criteria: opportunity for an innovation, appropriability, cumulativeness of knowledge and the complexity of knowledge base. Thus, the criteria rely on the classification and the measurement of knowledge. Identifying the levels of the four criteria will lead to the conclusion about the five dimensions of innovation. In turn, it will refer the coffee grinding and roasting industry to one of the types: Mark I or Mark II. The appropriability dimension is relaxed due to the use of patents as a method. Overall, by classifying the standalone innovation or industry development events depending on their influence and consequences I can construct the holistic picture of the industry evolution history.
The level of analysis of the research is based on the coffee-processing sector. In order to collect information about the events of technological innovation I will make use of the patents' history, forward patents' citations and relevant data. Analyzing patents is a proved method that tells us about the rate of innovation and every patent serves as evidence for innovation because it prevents a new invention from copying (Lanjouw & Shankerman, 2004). Hence, I will use patents as a proxy for innovation. This will show when and what firms were entering the market, how it was growing, and how important the patent was (i.e. innovative technology). Anderson and Tushman (1990) highlight another issue that emerges from the use of patents, which refers to the patenting activity as being an obstacle for the innovations. Thus, the use of patents should be explored. The research is qualitative and is regarded as an empirical study. I will gain the necessary insights by observing the data and
analyzing it in order to identify the patterns in the industry behavior. The period of the data to be analyzed it in the range between the years 1978 and 2013.
The paper is structured the following way. In the second section the literature review is presented where the main models and concepts are discussed in more detail. The concepts of innovation are generalized and applicable to the majority of the industries, therefore, the coffee grinding and roasting industry fits the criterion for the analysis. Afterwards the conceptual framework introduces the specific models with implications that are used in data analysis and with the help of which the research question is answered. The fourth section deals with the methodology that outlines the specific steps of the data analysis about how the data is collected, assessed, analyzed and reported. The fifth section elaborates the results of the data analysis and the sixth section serves as a follow-up with the discussion on the results. The last chapters of the thesis contain the discussion on the limitations of the research, topics for future research and the managerial implications on the issue. Finally, the last section concludes the thesis.
2. Literature review
Innovation is one of the most prominent topics in the life of a firm and a direct synonym for the concept of development and evolution. The topic has been covered by many academics and has a very broad nature of perspectives. A concept of innovation, innovation activity and industry evolution can be also viewed and explored from the financial perspective and make use of financial performance indicators such as revenue or profit. However, this thesis research abolishes their use due to the limited scope. The main focus is on the understanding of the driving forces behind the mechanism of innovation: how it occurs, how it processes and what is an outcome. This way, the thesis bases on the strategic side of an innovation. The literature review encompasses the main findings on the theory of innovation and industry development and other relative information surrounding the topic, for instance, knowledge. The goal of the literature review is to gain an insight on the theories, to see the connection between them, differences and similarities. It will give the research an academic background and provide necessary information for constructing the model that will be used further in the data analysis. The literature review starts with the definition of innovation and its basic classification. Further, more advanced models of different authors are introduced that discover an innovation activity based on diverse factors of market, industry and its environment. Such concepts as knowledge and information, transilience, product and process innovation are discussed. The section wraps up with the conclusion and subsequent direction of a research.
2.1 Concept of innovation
The theory of an industry life cycle states that the development of an industry is a dynamic process (Dosi et al., 1997). It means that the structure of an industry varies depending on the stage of development. Gort and Klepper (1982) referred to industry development as a development of the market. Therefore, they have identified following five main stages of development.
Fig.1 (Gort & Klepper, 1982, p.639)
The first stage of an industry development cycle is a commercial introduction of a revolutionary new product that has not yet existed in the market. It marks the birth of an industry. The invention of a product, however, could have happened much earlier before the commercial introduction and the producers were waiting for a right moment or customers' need in a product. The second stage in an industry development cycle is an increase in the number of new producers and entrants into an industry. Thirdly, the amount of newcomers and those who exit an industry balances out and the net entry is getting closer to 0. During the fourth stage the negative net entry persists which means that the number of entrants is less than the leavers. Lastly, the stage of the second balance occurs and the net entry again comes back to being close to 0.
The process of innovation occurs at every stage of an industry life cycle. Abernathy and Clark (1985) define innovation as a sequence of activities that involve acquisition, transfer and utilization of information. Cowan and Foray (1997) suggest a suitable theory regarding the codification of knowledge into information. According to the authors, knowledge involves into a piece of information after a process of codification which involves the use of model
building, language creation and writing of messages. Information is perceived as a good or commodity that has its cost, it can be simplified, stored, shared and understood by various agents. Information can be organized into systems, such as databases and other types of archives. This way, innovation can be regarded as an accumulation of knowledge because it mainly utilizes information in the activities mentioned above as a mean of technological advance. In its turn, accumulated information on technology refers to a selected technological problem, based on selected principles and selected material technologies. It systematically builds up into a technological paradigm (Dosi, 1982). Technological paradigm is different from the scientific paradigm that has a more general outlook and deals with the problem and a pattern or model for its inquiry, according to Dosi (1982). A technological paradigm sets a direction for the development or evolution not only internally for a firm but it also stimulates an industry innovation to a certain degree. This way, it is possible to make a connection and summarize innovation as an accumulation of knowledge that is transformed into information and aimed at a particular problem's solution.
The accumulation of knowledge in solving a selected problem leads to learning effect for a firm and an industry, overall. The main objective of innovation process is considered to be product and process innovation depending on a stage in an industry development cycle (Adner & Levinthal, 2001). The authors state that the product innovation corresponds to the period of a product enhancement and advances through multiple of innovation by various producers up until the dominant design is established. The notion of dominant design is introduced further. After the dominant design is set, the focus by the producers is made on the process innovation. The process innovation concentrates at the production process and costs of production while the dominant product design stays the same. The product innovation stage, thus, is more attributable to the early stages of industry development, for instance, during an increase in number of newcomers. This way, the overall process of innovation is viewed as a tool for a firm to strengthen its position amongst competitors and in the face of buyers or suppliers within the product innovation stage, as well as to ameliorate its internal competences during the process innovation period (Geroski & Machin, 1993). It is pursued by the circular shift of focus from product innovation to process innovation and the impact from one stage cannot have a strong effect without another (Geroski & Machin, 1993).
2.2 Taxonomy of innovation
The process of innovation is multifaceted. Depending on the criterion of an innovation's nature, different authors suggest various ways on how to classify innovation with regards to the external market forces or internal and industrial forces, the target of innovation or the
knowledge features of an innovative activity. In the following section I will refer to the three most common taxonomies and assess their differences and commonalities.
2.2.1 External and internal nature of innovation
In the case of distinction between technology-push and demand-pull innovation made by Dosi (1982), the difference is due to the generator of innovation. Technology-push innovation is described as the one that occurs due to autonomous internal changes in a firm. In particular, it can be explained by an increase in attention paid and financing to the R&D centers and departments. Technology-push activities come up with the product or process ideas before the market realizes the need. Therefore, there is an opportunity to assess an innovation a priori because the market is implicitly showing the direction for an innovation by the changes in prices and quantities of goods, which is called signaling (Dosi, 1982). However, it might be problematic and misleading in reality to detect the correct utility function of consumers and estimate the demand before the actual introduction of a product or process. The consequence of a wrongdoing in the technology-push approach can be a technological stagnation identified by Geroski and Machin (1993). It implies an absence of technological development in an industry.
A contrary approach to technology-push innovation suggested by Dosi (1982) is the demand-pull method. In this situation the external market forces stimulate innovation and the need for a product is clearly determined. A certain product can already exist in the market and consumers reveal their preferences regarding how the product can be modified. It is argued that a pure demand-pull approach can be distorted by difference in prices, budget constraints and income. Thus, the method cannot be used independently to identify the innovation opportunities. Moreover, the institutional and external environmental factors can play an insignificant role in shaping consumers' preferences and, thus, be misleading from the true demand function.
2.2.2 Transilience map
The second taxonomy of innovation discusses in the section is stated by Abernathy & Clark (1985). Here, the authors distinguish among the following four types: regular, niche creation, revolutionary and architectural innovation. The basic idea lies upon division of innovation into disruptive and improving. The first type leads to the destruction of established competences, making them obsolete. The second, in contrast, refines an already established product or process. As mentioned before, the accumulation of knowledge is the driving force behind any type of innovation activity. Abernathy & Clark (1985) name the capacity of
innovation to influence the firms established resources as transilience. The model represents a two-dimensional quadrant that sums up the four types of innovation (Fig.2).
Fig.2. Transilience map (Abernathy& Clark, 1985, p. 8)
The first dimension is the degree of transilience of capabilities of an innovation activity, i.e. whether it disrupts or improves the existing capabilities. The second dimension is the degree of transilience of a market-customer link, i.e. whether it erases, creates, or maintains the link. The first type, regular innovation, aims at improving both the competence and the market link. The second type, niche creation, as the name suggests: the competences are improved while the market link is disrupted. The revolutionary method of innovation offers a completely new competency and focuses on keeping the market-customer link. The last type, architectural innovation, is described as disrupting competence and the market-customer link.
An industry life cycle of Gort and Klepper (1982) can be associated with the above-mentioned model where the five stages of a cycle correspond to a certain type of innovation activity. For instance, the introduction of a product into the market refers to the architectural type, where competency and market link are created. The period of new producers coming into an industry points towards the niche creation type when each producer with the
established competencies tries to find a suitable market link. The regular innovation is compared to the stages of balancing out. The revolutionary period corresponds to the fourth stage of a market life cycle where competences are disrupted and the market links are refined.
2.2.3 Schumpeterian innovation activity
The classification suggested by Malerba & Orsenigo (1996) involves two types of innovation activity: Schumpeter Mark I and Schumpeter Mark II. Briefly, they can be characterized as either creative destruction (Mark I) or creative accumulation (Mark II). In particular, the distinction is based on the following parameters: rate of concentration of innovation activity among firms, degree of stability in the hierarchy of innovation firms and the rate of technological entry and exit (Breschi et al., 2000).
The first parameter, rate of concentration, indicates the number of firms in an industry involved in innovation activities or, more generally, the number of innovations. The second parameter, stability in the hierarchy, defines the degree to which a firm is able to maintain its position in innovation activity and prevent it from imitation. The last characteristic specifies the entry of new firms in an industry and the exit of firms. Therefore, the taxonomy of Malerba & Orsenigo (1996) attributes Schumpeter Mark I paradigm to score high rate of entry and exit and low on the rate of concentration of innovations and the degree of hierarchical stability. Schumpeter Mark II type, on the contrary, is described high in the rate of concentration of innovation activity and hierarchical stability and low for rate of entry and exit.
With regards to the ways in which the abovementioned factors are established, Breschi et al. (2000) mention four parameters that are also previously admitted by Malerba & Orsenigo (1996). The parameters are an opportunity for an innovation, cumulativeness of technical advances, appropriability of innovation activity and knowledge base. Opportunity parameter determines the possibility of occurrence of innovation activity in an industry, i.e. the overall conditions for performing an innovative activity and the extent to which the environment stimulates it. Cumulativeness of technological advances relates to the capacity of knowledge accumulation. For instance, those firms that are already involved in an innovation activity have higher chance to continue it due to the obtained background. Appropriability with respect to an innovation corresponds to the degree of protection awarded to an innovation innovation, i.e. a high rate of appropriability means that a firm can gain advantage from its innovation since it is protected from imitation due to the legal rights for ownership for a specific period of time. Low appropriability, in turn, means a widespread possible use of an innovation by other firms. The last parameter that helps measuring and classifying innovation
is knowledge base. This way, knowledge base can differ in its properties and relate to the generic, i.e. scientific knowledge, or to the specific, i.e. applied knowledge base. Depending on classification of knowledge the nature of technological innovation is different because knowledge has an impact on other parameters as well and is, indeed, the most important and influential from all.
The Schumpeter Mark I and Schumpeter Mark II types are described by the different degrees of the four parameters. Mark I is defined by high opportunity, low cumulativeness, low appropriability and high applied knowledge base. It can be pictured as a stage in the development of an industry with many small firms entering the market, actively innovating and innovation based on solutions for selected problems. Beesley & Hamilton (1984) call the function of small firms a 'seedbed' function. By this, they imply that widespread growth of innovation and an outcome of it is the holistic development of an industry. The Mark II type is placed on the other side of continuum and is described by low opportunity, high appropriability and cumulativeness and high generic scientific knowledge base level. An industry of type Mark II has a few dominant firms that protect their innovation and use their own experience in innovation in an industry.
In summary, according to Breschi et al. (2000), the Schumpeter Mark I is considered to be a widening design of innovation activity where the knowledge is expanded due to the presence of many innovative firms. Schumpeter Mark II assigns a deepening process for innovation activity where the knowledge is extended through continuous accumulation within several firms in an industry and newcomers suffer from obstacles to survive. The abovementioned differences among industries can be associated with a technological regime or learning regime. In the earlier section, the concept of technological paradigm was explained which directly adheres to the technological regime. It defines the mainstream of technological development in an industry. Therefore, Schumpeter Mark I and Mark II can be viewed as technological paradigms of an industry development.
2.3 Technological change
As discussed before, an industry life cycle embodies five stages that have different patterns of innovative activities, number of firms and goals of firms. Although, innovation is a part of every step of the cycle, it has its own process structure. Anderson and Tushman (1990) consolidate the main patterns of innovation in the following model.
The first pattern is a point of technological discontinuity where the new technology is introduced. At this moment the established process or product is discarded. The discontinuity can appear as an improvement or destruction of an old method. It correlates to the introduced technological paradigms Schumpeter Mark I and Schumpeter Mark II. At the same time the new designs and alternatives to the initial innovation start to appear. The period is defined as an era of ferment. During this stage, the best design is selected through competition. The competition runs not only among the new versions of technology but also between an old and innovative view on technology. The best innovation design replaces the outdated process or product. A hypothesis supported by the authors states that the competence-destroying innovation carries a longer ferment period than a competence-refining innovation. The model peaks at the point of dominant design.
The dominant design is an outcome of selection process among alternative technologies and an establishment of dominance of design in an industry. Anderson and Tushman (1990) state that the further development is based on the dominant design and its improvements. The dominant design becomes a guidepost, an industry standard. Although, the process of establishment of a dominant design is mostly autonomous in nature and, as mentioned, an outcome of natural selection process for the best design, it is not always the case. Besides the design to have the best performance, to be the most efficient, advanced and innovative, it should correspond to social and political standards (Anderson & Tushman, 1990). Hence, the setting of a dominant design is a social and political process. Anderson & Tushman (1990) and Suarez & Utterback (1995) give an example when the government finances R&D, i.e. innovation activity, in a selected industry in order to implement a certain policy or simply sets a standard for an industry. In this case, the innovation is an imposed process not autonomous.
Anderson & Tushman (1990) refer to the period following the establishment of a dominant design in an industry as an era of dominant design elaboration or an era of incremental
change. During this stage the dominant design is not changed entirely, i.e. it is not comparable to the technological breakthrough, but rather some small variations occur that have an impact on the product or process innovation. The technological breakthrough accounts for a major shift in technology while an incremental change is associated with the series of minor improvements (Anderson & Tushman, 1990). It should be noted that the patenting activity that causes an appropriability rate to be high might prevent the establishment of a dominant design. An era of dominant design elaboration and incremental change is followed by the next technological discontinuity and the cycle of innovation process continues.
In the paper Anderson & Tushman (1990) pose and support the hypothesis that refer to the cyclical model of technological change. The first supported hypothesis state that the innovation of product or process varies more during an era of ferment than during an era of incremental change. From the above-mentioned notions of the two periods, the difference can be explained by the high rate of competition during the ferment stage and more stable innovation activity after the establishment of a dominant design of product or process.
The second hypothesis acknowledges that the ferment era following a competence-destructing breakthrough lasts longer than the one following the competence-enhancing activity. It can be justified by the fact that the competence-destroying innovation involves a stronger conflict situation where old views become obsolete and the new views on technology gain prominence. Therefore, the competition is held between the producers of existing design and new design producers, as well as among the new producers with competing innovations. In its turn, the competence-enhancing innovation builds up on the previous technology and the only conflict situation is among competing designs.
The next supported hypothesis claims that the high appropriability rate of innovative activity might lead to the absence of a dominant design. According to Anderson & Tushman (1990) it happens because the competition and the positions of rivals in the market are predetermined due to industrial structure. Low appropriablity conditions, on the contrary, always result in an establishment of a dominant design in technology because the competitions is held among more firms and the imitation is possible. The spread of knowledge among the firms is greater under the low appropriability conditions than under high appropriability.
Another hypothesis affirms that the breakthrough innovation design does not become a dominant design. An important role of environmental factors such as political and governmental forces on innovation was previously highlighted. In addition, the period of
ferment introduces alternative that can be superior to the breakthrough innovations. Thus, the dominant design is not based solely on the discontinuity design.
One more supported hypothesis is connected to the difference between the eras of ferment and incremental change. It is stated that the major technological progress takes place during the ferment stage rather than the incremental change. An evidence for that is the competition during the ferment stage. The competition provokes producers to improve their design as much as possible so it has better performance than its competitors. This leads to an overall improvement of performance and diversity, and the selection of the best technological design. The last relevant hypothesis is supported only partially. It proves that the competence-refining innovative activity stems from the incumbents of an industry. The alternative hypothesis that the competence-destruction activity is associated with the new comers is not supported. Therefore, it is in tune with the Schumpeter Mark II type of innovation where established firms in an industry have a higher rate of cumulativeness of technological advances.
2.4 Conclusion
In summary, the literature review has shown the development of the theories on industry life cycle, innovation and the specific environmental factors that influence the technological progress in an industry. The indicated stages are commercial introduction, new entries, balance of entries and exit, negative net entry and another balanced stage of entries. Innovation activity appears to be incorporated into an industry life cycle at every stage. The two forces dictate appearance of innovation: technology push or demand pull. Briefly, innovation is caused either by internal or external forces. Different authors agreed that innovation activity is not purely autonomous and there are political and social mechanisms involved. Therefore, innovation is a two-sided process that takes place internally in a firm and externally in an industry. A model of Abernathy & Clark (1985) that classifies an innovation activity into regular, niche creating, revolutionary and architectural stands on two dimensions of market-customer link and competence. A firm moves around the quadrant along two dimensions to either enhancing of destructing market-customer link or competence. Because an innovation is a dynamic process a firms is positioned differently at different time periods. An important piece of theory of the model is the basic distinction between disruptive innovation and enhancing type of innovation that is further used along with other models.
The taxonomy of Malerba & Orsenigo (1996) states the division between Schumpeter Mark I and Mark II innovation activity and references to the disruptive and enhancing view. Innovation can be distinguished between product and process innovation and its application depends on the stage of an industry life cycle. Overall, innovation is a process and, thus, it
consists of steps such as breakthrough, establishment of dominant design and incremental change.
The referrals to the academic papers have suggested that patenting can be used as a proxy for the innovation activity. In addition, the patenting activity does not contain the recording of all innovation events. The implications of use of patents are discussed further. In the next section the conceptual framework is stated. It includes the model that will be used in order to analyze the data, the factors that should be considered and their functions. Moreover, the hypotheses are outlined in the conceptual framework section. The method used in this thesis to gather and analyze the data corresponds to the methods outlined in the literature review. In subsequent sections the methodology, reliability and validity of the research are discussed.
3. Conceptual framework
The conceptual framework represents the summarized findings from the literature review. The purpose of this section is to develop the model that will be further used in the data analysis and in the testing of the hypotheses. Firstly, the model of technological change of Gort and Klepper (1982) in Fig. 3 is used. The point of interest is identifying the breakthrough innovations, processes during an era of ferment, establishment of a dominant design if present, further era of incremental change and the overall cycle of this model. This way, as an industry has periods of rise and decline in innovation activity it is expected that the development of an industry is continuous and bell-shaped depending on the rate of innovation activity.
As stated by Anderson and Tushman (1990), the patenting activity can be an obstacle in the process of dominant design setting. This thesis makes use of patents as a major source of information on innovation activity in an industry of coffee grinding and roasting. Hence, it is expected that due to the patents the dominant design in an industry is not obvious and an era of ferment is rather more continuous since there is no climax.
The next step is to identify the patterns that refer an industry to either Schumpeter Mark I or Schumpeter Mark II type. For this purpose, the following parameters should be considered: concentration of innovation activity, degree of hierarchical stability and the rate of entry and exit. Because the patents imply that a firm that makes use of patents is involved in innovation and aims at protecting its innovation, the parameter of concentration of innovation activity is relaxed. This way, only two parameters are assessed: hierarchical stability and the rate of entry and exit.
In order to fully acquire the evidence in favor of one of the types, it is necessary to include the parameters, which have an affect on the previously mentioned measures: opportunity for innovation, cumulativeness of technological advances, appropriability and the knowledge base. However, the appropriability characteristic can be abolished due to the use of only patents that cannot give clear results about non-patented innovations. On the other hand, the use of patents is an evidence for Schumpeter Mark II type, therefore, it is predicted to be proved.
Another pattern that is to be identified is the power of external and internal forces on innovation. Particularly, with regards to coffee industry, it is a point of concern because of the International Coffee Agreement (ICA) between coffee producing and coffee consuming countries that is ratified periodically starting from 1962. It limits the amount of coffee to be produced and, as a result, to be exported to consumers' countries. It is expected to have a certain effect on innovation activity especially in the few years when the ICA was not signed. Yet, the predictions of an effect are unclear. On the one hand, ICA can stimulate the more effective and efficient production process and innovation in the grinding and roasting industry. On the other hand, the limits of production can lead to a consequent decrease of technological change as only the major producers will stay in an industry and benefit from the greater market share and economies of scale, which, in turn, will be another proof for the Schumpeter Mark II type.
The area of research in this thesis does not include testing whether an innovation in coffee roasting and grinding industry corresponds to the transilience map algorithm because the data on customers and market is not used as it is beyond the scope of this thesis. Some dimensions of the models are relaxed because they are either fixed and predetermined in the research, or complex to measure. For instance, I do not use the appropriability criterion from the model of Malerba & Orsenigo (1996) because the method of the research includes the use of patents and it implies high appropriability. I do not refer to the transilience map of Abernathy & Clark (1985) because of limited time, data and the scope of the research. This approach will provide more focused analysis of the relevant characteristics and will not disturb the results. The hypotheses should help me answering the research question that is stated as the following: what are the main patterns of evolution in the coffee roasting and grinding industry in 1978-2013? Based on the models and the rationale behind them, the following hypotheses are formulated.
Hypothesis 1. The patterns of innovation in the coffee roasting and grinding industry in the
years 1978-2013 follow the direction of a model of technological change of Gort & Klepper (1982).
The first hypothesis refers to the model of cyclical industry development of Gort & Klepper (1982). Because the model can be applicable to any industry, I would like to address and to test whether the coffee grinding and roasting industry in the period of 1978-2013 follows the modeled patterns of development.
Hypothesis 2. The coffee roasting and grinding industry corresponds to an industry of Mark
II type.
The second hypothesis aims at classifying the coffee grinding and roasting industry as Schumpeter Mark I or Schumpeter Mark II type suggested by Malerba & Orsenigo (1996). It is expected that the industry in the period 1978-2013 behaves as type Mark II. Because the spread of knowledge is limited due to use of patents, the appropriability criterion is high, which corresponds to Mark II. I predict the hierarchical stability to be high as well because of the relative long existence of the industry, which has influenced the positions of producers in the market and there are stable roles of market players.
The third hypothesis focuses on the dominant design of technology in the industry. The dominant design, as stated by Gort & Klepper (1982), is the period with an established product innovation, where the industry standard of a product is set. Because the industry exists for more than 30 years already, I expect that there is an innovation that can be recognized as a dominant design and it is widely referred to by other producers.
Hypothesis 3. There is a dominant design of technology in coffee grinding and roasting
industry in the period 1978-2013.
The next hypothesis deals with the impact of governmental and legal forces on the innovations in the industry. As claimed by Anderson & Tushman (1990), the government and legal authorities have power in fostering the innovations in an industry, setting the standards, and facilitating the establishment of dominant design. As the International Coffee Agreement adopted by many countries in the world regulates the production and consumption in coffee industry, I forecast that it will have an effect on the number of innovations in the industry. Gort & Klepper (1982) states that the government and legislations have more effect on the innovation activity during the second stage in an industry development cycle, where the number of producers is growing and R&D is stimulated. The hypothesis is, thus, defined as following:
Hypothesis 4. Innovations in coffee grinding and roasting industry in the period 1978-2013
are affected by the legislative power more in the second stage of industry development cycle than in other stages.
This section contributes to further data analysis and serves as a direction indicator to the discussion of the results. The next section is dedicated to the methodology of the research. The main attention is drawn to the steps of how the results are obtained and the overall validity and reliability of the study and its data sample.
4. Methodology
This thesis aims at identifying patterns of innovation activity and its classification in the coffee grinding and roasting industry. The research is conducted as a qualitative empirical study. The patterns of innovation and the answer to the research question are based on the theoretical framework that is applied to the given data set. The selected coffee grinding and roasting industry has a point of personal interest and the identified gap in the information available on the evolution in the industry. It also serves as representable example of an industry where the information can be gathered and assessed in an appropriate manner to test the innovation hypotheses.
First of all, the presence of innovation activity in an industry should be based on the specific indicators. One of the most prominent and established evidence for the innovation activity is the use of patents. They are used as a proxy for an innovation activity (Griliches et al., 1986). Therefore, each patent is considered as a standalone event of innovation. The total amount of patents assessed accounts for the total amount of innovation events, i.e. inventions, in the coffee grinding and roasting industry. Notice, however, that the total amount of innovation events and technological changes is difficult to assess due to the absence of complete reporting, i.e. incomplete information, as not every innovation is patented. Therefore, this research outlines the general direction of the innovation activity in coffee grinding and roasting industry in the selected period (1978-2013) and has a degree of deviation.
The overall design of the thesis is a qualitative research. As stated by Saunders et al. (2003) the qualitative research is the one that uses mostly non-numeric data collection techniques and refers to the interpretive philosophy. This design fits the research question, which aims at providing the evidence for the pattern relationship and does not pursue a quantitative explanation or expression for any causal relationship. The methodology selected for the data collection is the archival research. In particular, I will explore the patenting data in the period
of 1978 to 2013. Because innovation is a dynamic process, the nature of the activity can be adequately addressed when the process has been established in the past and have outcomes that can be assessed. Thus, the research is based on the past innovation events. The selected period from 1978 till 2013 is explained by, first, rise in the number of patents in coffee grinding and roasting industry in 1978 and its subsequent growths and declines in number of patents, i.e. development movements. Moreover, the International Coffee Agreement was first introduced in 1962; therefore, the selected period follows the introduction of the agreement that allows assessing the external forces in the industry. Although the patents in the industry have also been used before the year 1978 (particularly the first patent introduced dated the year of 1704 (Ukers, 1935), the main patenting activity refers to the selected period. Moreover, the amount of data gathered for the analysis allows us to draw reliable conclusions and to answer the research question. The use of patents data is a reasonable way to get information about the innovation activity. Many authors state that patents can be used as a proxy for innovation because the idea behind patenting is to prevent the inventions from being copied by competitors (Griliches et al., 1986).
The data to be analyzed is accessed through the website of United States Patent and Trademark Office. The database provides a full access to the patents the coffee grinding and roasting industries in the selected period and, in addition, gives an access to the description of the patents. The patents' description shows the detailed information about the date of its establishment, country, inventors, assignees, and backward and forward citations. Therefore, this data is further elaborated and worked with. There is no official data available on the innovations that are not protected by the patents. Moreover, the patents clearly prove that an invention can be referred to as an innovation activity. Hence, the use of patents is the only reliable and straightforward way to get the information regarding the technological changes and innovation.
The sample of the research consists of 59 patents in the period from 1978 till 2013 in the coffee and grinding industry. The 59 patents are not restricted to a specific country. As stated by Breschi et al. (2000), technological development does not vary significantly depending on the country. Therefore, the countries' specificities can be ignored. The total number of patents for the whole time of existence of the coffee grinding and roasting industry represents the population from which a sample for the research is drawn. From the abovementioned discussion, the innovation activity that is not confirmed by the presence of a patent cannot be a trustworthy identification of an innovation event.
The theoretical framework outlined in one of the previous sections is applicable to the data set. Firstly, several authors have been using the same data collection method, i.e. they have also referred to the use of patents (Anderson& Tushman, 1990; Christensen et al., 1998). Secondly, patents provide reliable information and they are the direct reference of the innovations. Thirdly, the theoretical framework makes use of several models that, together, can provide a different prospective on the data and utilize most of the available patents' descriptions and other details to answer the research question. The sampling method is representative and reliable because more than 30 patents are assessed and, in particular, the selected period represents 35 years of industry existence which is a long timeframe. Overall, the research captures the various events of innovation activity in an industry and the selected period allows enough time to explore and to see the connection among the technological changes.
The number of patents is the main measure in a process of identifying patterns and behavior of innovation activity. The measure varies depending on the hypothesis that is being tested. In order to explore the general direction of innovation in the industry, the total number of patents per year is taken. The number of patents is compared between years when identifying the stages of innovation development. The number of patents per producer is considered when testing the hypothesis on Schumpeter Mark I and Mark II classification. The number of backward citations is another measure used in the analysis. Firstly, it helps in determining the nature of the innovation event and referring it to one of the stages of the model of Anderson & Tushman (1990). Therefore, the number of patents and the number of backwards citations are the two measures that are used throughout the analysis.
The first step in analyzing the data is to identify the total number of patents per year in the period of 1978-2013. This will show the main trends and the degree of innovation activities in an industry throughout the selected timeframe. In particular, it becomes possible to distinguish the main stages of innovation: commercial introduction, new entries, balance of entries and exit, negative net entry and another balanced stage of entries. Commercial introduction will correspond to the first appearance of a patent; period of new entries is described by the rise in the amount of patents; the balance period refers to the decline in the number of patents; the negative net entry will be identified by the slowdown in the patenting activity. Lastly, the negative net entry is followed by the slight increase in the patenting activity and, thus, the number of patents. In addition, in the case of the coffee grinding and roasting industry, the patents refer to the product innovation rather than the process innovation.
The next step that is applicable to every innovation case is detecting whether it is competence-destroying or competence enhancing. The following can be assessed with the help of the backward citations. Backward citation is the referring to the previously established patent (Jaffe & Trajtenberg, 2002). The competence destroying innovation makes the previous innovation obsolete and represents a totally new process or product, and, thus, is not based on the previous experience. However, it is expected that the basic design of a product can persist, which does not refer to the dominant design, but rather an essential part of a product. This way the competence destroying technological change is assumed to have no or few backwards citations. Hence, the competence enhancing innovation is associated with the greater amount of a patent's backward citations.
Further, the producers who hold the patents will be reviewed. Combined with the amount of patents per year and the number of backwards citations, it will be possible to identify the innovation steps such as breakthrough, dominant design and an incremental change. The patents with the highest number of backward citations should be considered as the dominant design products.
The data will be reported by using representative tables that summarize and consolidate the information from the descriptions of the patents from the USPTO database for coffee grinding and roasting industry. The subsequent graphs will depict the information from the tables. These two data analyses together with the patents' descriptions are the main sources to draw the conclusions and identify the patterns. This way of analyzing and reporting data is the one that allows to fully exploring the given data. The use of several measures (number of patents and number of backward citations), the use of patents' descriptions and the theoretical framework allows for the triangulation of data. It leads to more reliable testing of the hypotheses.
The reliability, according to Saunders et al. (2003), is the consistency of findings in case if the research is replicated. Firstly, the application of patents for judging and classification of the innovation in an industry is used in the various academic researches. The chosen method provides a strong reliability as it analyzes the full data available of 59 patents. It is the total amount of patents for the period 1978-2013, therefore, the descriptive analysis and further results transfer the true behavior of the industry's innovation activity. Moreover, the data is retrieved from the official database that positively influences the strength of reliability. The usage of academic sources and references in the theoretical framework provides a reliable background for building the conceptual framework and the analysis of the data.
The internal validity, stated by Saunders et al. (2003), means that the causal relationship between the independent variable and an outcome variable is statistically established. In this thesis the independent variable is the number of patents and the number of backward citations. However, there is no outcome variable as the method is descriptive. Therefore, the emphasis is made on the drawing conclusions and making classifications by interpretation the independent variables. It can be questioned whether the internal validity of this research is strong. The support for it is the reference to multiple academic sources that provide the models that can be used as an interpretation tool. It means that the presence of a dependent variable is not necessary and cannot be established. The construct validity implies that the measures used indeed measure what they should (Saunders et al., 2003). In the case of the number of patents and backward citations that is a direct representation of the innovation activity. It establishes the direct relationship between the variable and its interpretation based on the innovation models stated in the conceptual framework. External validity evaluation, proposed by Saunders et al. (2003) refers to the practical use of the research, i.e. whether the results can be used in other settings. This research examines the patterns of innovation and technological change in the coffee grinding and roasting industry. The main tool in describing the patterns is the patents and their use. The previous researches have been running similar methods applied to another industries. Because the method of using patents does not have particular differences depending on an industry, it is possible to conduct the research in another industries. The results of this particular research can be widely used in coffee industry and market evaluations.
5. Results
This section summarizes the results obtained from the data analysis. The method used for the data analysis is outlined in the previous section. The main concept was the use of patents in the coffee grinding and roasting industry for the period 1978-2013. Each patent was analyzed at the individual level as well as combined with the whole data. The data was condensed and assessed by considering the necessary components from the patents' descriptions (Table 1, Appendix). The data analyses and this section are built based upon the sequence of the hypotheses from the Conceptual framework section. First, the data that corresponds to the analysis with regards to first hypothesis is assessed and presented. The section continues with the results on the Schumpeter Mark I and Mark II classification (Hypothesis 2) and is finalized with the data results for the third hypothesis, which identifies the presence of the power of external forces in the innovation activity. In the very end of the section the main
findings are presented once again. All of the mentioned tables and other relevant data used throughout the analysis are placed in the appendix of the thesis.
5.1 Hypothesis 1
In order to test the first hypothesis I reviewed the total number of patents per year for the period of years 1978 till 2013. Fig. 3 shows the trend line that appeared from the data (Table 2& Table 3, Appendix).
Fig. 3. Total number of patents per year
Fig.3 shows that the patenting activity is following a bell-shaped cyclical trend. The number of patents per year corresponds to the innovation events. Therefore, the trend line from the Fig.3 can be interpreted as a graph of innovating activity where the periods of rise and decline of patenting activity refer to the rise and decline in an innovating activity. Fig. 4 depicts the cumulative number of patents per year.
The total number of patents is growing throughout the whole period with minor periods of stagnation. The minimum number of total number of patents per year is 2 in 1978 and the maximum is 59 patents in 2013. There are peaks of patenting activity in certain periods and rise and declines in other times. The maximum number of patents in the industry for the total of 6 is observed in 1988 and the minimum amount of 0 appear in years 1981, 1984, 1992 1997, 1999 and 2009. The periods in between these years correspond to the periods of gradual rise and decline. The periods of rising patenting activity that can be derived from the graph are 1981-1982, 1984-1985, 1986-1988, 1990-1991, 1992-1993, 1994-1996, 1997-1998, 1999-0 1 2 3 4 5 6 7 19 78 19 80 19 82 19 84 19 86 19 88 19 90 19 92 19 94 19 96 19 98 20 00 20 02 20 04 20 06 20 08 20 10 20 12
N of patents
N of patents2000, 2002-2003, 2006-2007, 2009-2010 and 2011-2013. The decline periods are 1978-1979, 1980-1981, 1983-1984, 1988-1989, 1991-1992, 1993-1994, 1996-1997, 1998-1999, 2000-2001, 2003-2004, 2005-2006 and 2007-2009. There also exist the plateau periods of patenting activity: 1979-1980, 1982-1983, 1985-1986, 1989-1990, 2001-2002, 2004-2005 and 2010-2011. These rising and declining movements in the industry's patenting activity serve as an evidence for the overall cycle that follows the direction of the model of Gort and Klepper (1982).
Fig. 4. Cumulative N of patents per year
The plateau stage does not last for more than one year. The plateau stage in the number of patents separates the rise and the decline stages by, thus, making it not abrupt. In the plateau stage the number of patents is one year is the same as the number of patents in the subsequent year. In the model the plateau stage would correspond to the periods where net entry is close to zero. However, the same producers do not introduce the patents in two subsequent years. The five events of plateau patenting activity involved new producers entering the market.
The next step in the identification of particular stages in the technological change activity is the evaluation of the firms in the industry. Table 5 (Appendix) displays the comparison of net change in the number of patents per year versus the net change in the number of entrants in the coffee grinding and roasting industry. The number of producers of patents is used as a proxy for the number of entrants in the industry. The results exhibit the coherence and the similar patterns in rising and declining activity between two parameters. The events of decline
in the total number of patents per year correspond to the decline in net change in the entrants per year (Fig.5).
Fig. 5. Net number of entries per year
In contrast, the events of rise in the total number of patents per year are associated with the rise in the net change in entrants, i.e. there are more newcomers in the industry. The graphs depicting the total number of patents per year and the net entry per year both show same nature of the rise and decline trend. Figure 5 depicts the net entries per year where the rise, decline and plateau stages are recognizable.
With regards to the classification of periods to specific stages, the first that should be noted is that there is no commercial introduction stage within the data range. The commercial introduction, as a birth of the coffee grinding and roasting industry, dates to earlier years and the first innovations most probably were not yet patented. Further classification suggests that the increase in the number of entrants correspond to the second stage of an industry development cycle. The third stage refers to the plateau after the increase in number of entrants. The decline in the entrants' number is the fourth stage where the number of new entrants is 0. The plateau period that is following the decline is associated with the fifth stage of a cycle where the net entry is again 0. Hence, there is evidence that the coffee grinding and roasting industry is following the direction of technological development outlined in the model of Gort & Klepper (1982).
0 0,5 1 1,5 2 2,5 3 3,5 19 78 19 80 19 82 19 84 19 86 19 88 19 90 19 92 19 94 19 96 19 98 20 00 20 02 20 04 20 06 20 08 20 10 20 12
N of entrants
N of entrants5.2 Hypothesis 2
In the years where more than one patent is introduced and, especially, the years with four to six patents, the same producer introduces several patents. For instance, in the year 1988 two producers (Grindmaster Corporation, Louisville, Ky and Brevetti Gaggia S.p.A., Milan, Italy) introduce two patents each. The same pertains for the year 1993 where the same producer (Robert Krups GmbH. & Co. Kg., Solingen, Fed. Rep. of Germany) introduced four out of five patents. These examples prove the characteristic of knowledge accumulation by certain firms of Mark II type of an industry. Moreover, other producers cite those firms that hold a few patents less often. It speaks for the generally high internal accumulation of knowledge internally in a firm but also high spreading within the industry. The firms that hold one patent (e.g. Sanyo Electric Co. (2000) are cited more often. This way, the innovation concentration level is not high because the firms with existent patenting and innovation activity introduce the majority of patents but there are also many newcomers. There are nineteen standalone patents and forty that are associated with the existent producers. The data for the citations is summarized in Appendix, table 6.
The level of net entry in the industry shows the activity of firms’ entrance and exit. From the table introduced earlier, I notice that from the total of thirty-five periods, the rise in the net entry of firms occurs in twelve cases. That is considerably less than the number of decrease in net entry that account for fourteen cases. Also, the plateau periods in net entry, i.e. when there are no new entrants in the industry, accounts for nine cases. Hence, the decline in the net growth per year or stagnation in the net entry prevails over the increasing net entry. The net exit, however, cannot be assessed separately because it requires the data on specific producers and due to the limited scope of this research it is not possible to include such an informative fragment. It can be concluded that the net entry in the coffee grinding and roasting industry it still high.
The next step is identifying the ratio of competence-destroying and competence-enhancing innovations. I reviewed the amount of direct citations to other patents (Table 7, Appendix). Those patents that are not citing any other patents are considered to be competence destroying. The ones that are citing at least one other patent are seen as a competence enhancing innovation event. There are forty-five patents that reference other patents, which represents 76% of the total. Therefore, the competence enhancing innovations prevail in an industry, which is also consistent with the findings in the previous paragraph on the entry and exit activity.
The hierarchical stability in the innovation activity of firms is predominantly correlated with the level of concentration discussed above and the appropriability conditions. The appropriability conditions are high because the data set includes only patents, and the concentration level is high as the majority of patents are referred to the existing firms. From this, it can be derived that the hierarchical stability is high in the coffee grinding and roasting industry in the given period. From the data analysis, the relationship between the firms and the overall hierarchical stability can be determined. The firms that have more than one event of patenting activity tend to refer to other patents rarely and to those that have high amount of citations, which means that the patent is of high value, or to their own previous inventions. However, other producers with more than one patent commonly cite the producers with one patent. Therefore, there is a trend that coalitions of powerful firms within the patenting community in the industry maintain their high-ranked positions by referring to other strong producers. The standalone events of patenting, in this case, are remarkable for often use of backward citations. In brief, the hierarchical stability in the coffee grinding and roasting industry is being maintained. However, in the recent years more producers are entering the market and it can be expected that the hierarchy will be changing.
5.3 Hypothesis 3
In order to test the application of the model of Anderson and Tushman (1990) in this analysis, I used the industry's total number of patents, firms, and backwards and forward citations. In particular, the interest was in exploring the establishment of dominant design technologies in an industry with patents. Firstly, the total amount of backward and forward citations was analyzed for every separate patent. The patents that are never cited by others or only once are perceived as a standalone event of innovation and it does not refer to the dominant design technology. There are 27 patents out of 59 that correspond to the description of 'non-dominant'. In particular, the patents for the period 2006-2013 are never cited. It can lead to two conclusions: either the period of dominant design establishment is longer than this period or that the previously established dominant design still persists in the industry. Taking into account that one of the patents is cited 14 times that is more than others (Braun Espanola, 1982), I can refer to this patent as a dominant design. In addition, the citations of this patent occur at different years during the given period.
There are other several salient points that stand out from the data set. The patent of Braun Aktiengesellschaft (1994), Bunn-O-Matic Corporation (1988), Sanyo Electric Co. (2000) and Appliance Development Corporation (2001) are cited by others 9, 8, 8 and 7 times, respectively. Moreover, all of the above mentioned patents are cited throughout most of the year in the selected period, although, there is some crisscrossing among the citations of
patents. All the above-mentioned patents develop with an individual dynamics and the citations correspond to the specific time periods. There are a few periods where are patents are cited simultaneously. Hence, the only identifiable dominant design technology in the industry is the innovation in 1982 (Braun Espanola). As Geroski and Machin (1993) state, the process innovation follows the product innovation after the establishment of a dominant design. This way, as the innovative activity is high during the whole observed period, the innovations can be process focused. Because the dominant design is established and the dominant patent is cited regularly, the majority of innovations that are cited rarely refer to the process innovations.
5.4 Hypothesis 4
The analysis included assessing the number of patents per year and comparing it to the years of ratification of International Coffee Agreement (ICA). This way, the relationship between the innovation and the power of external forces, particularly the governmental environment, can be discovered.
The ICA was ratified in the years 1968, 1976, 1983, 1994, 2001 and 2007. The agreement deals with the limit on export and import and overall control of coffee production. The list of the exporting countries (International coffee agreement web-site) includes the countries that submit patents in the coffee grinding and roasting industry. After assessing the number of patents per year at the time when the ICA was introduced, I can state that there is no clear relationship. In the year 1983 and 2001 the number of patents rises after the adoption of the ICA. In the two other time periods, in 1994 and 2007, the number of introduced patents rises. The increasing periods account for higher net change in the number of patents than during the decline periods. Nevertheless, there is not enough evidence to support the hypothesis because these trends are not necessarily related to the agreements and I cannot interpret them in a way that brings the significant results about the relationship between two parameters.
5.5 Conclusion
Overall, the results presented in the section suggest that there is some clear data regarding the hypotheses, although, some data is controversial and requires further discussion and especially the discussion of the conditions and assumptions. The next section is dedicated to the discussion of the results where the final conclusion regarding the patterns of innovation in the industry is drawn. Further I continue with the limitations and future research suggestions, managerial implication of the thesis and finish with the final conclusion for the research.
This section wraps up with the discussion and conclusion based upon the results. The structure of the section is similar to the Results section where the data analysis is discussed according to the sequence of the hypotheses. The main emphasis is made on the interpreting the findings and making connections between the hypotheses and the theories from the literature review. In order to analyze the innovation activity in the coffee grinding and roasting industry it was also necessary to include the data on the external forces, therefore, the discussion deals with the multiple levels of overview: starting with an individual classification patterns and minor details regarding the firms themselves and finishing with the holistic picture on the industry dynamics.
6.1 Hypothesis 1
The results of the data analysis for the first hypothesis suggest that, indeed, the coffee grinding and roasting industry innovation activity follows the path outlined by the authors (Gort & Klepper, 1982). The industry shows the same dynamics, which is shown by the number of patents. The bell-shaped line is proved to be applicable to the industry. However, it was expected that during the development of the industry the base technology improves and, thus, the base level of patents would also increase and the further patented inventions will start from the higher initial level of patents. The absence of a base level might stem from a high level of competition because there are 33 producers in the industry in the dataset and all of them are active in the patenting and most of them are active during a long period where they introduce more than one patent. Moreover, it is worth noticing that most of the firms are focused only on the coffee grinding and roasting industry, therefore, they are more inclined to be active members and contributors to the industry history. The data has shown that there are a lot of new entrants in the industry throughout the whole period that are introducing patents and, therefore, competing with each other by introducing new technologies. It is coherent with the ferment stage of an industry development. Even though the separate dynamics and changes in the period can evolve in a micro-cycle, the industry as a whole related to the second stage of development in the cycle.
6.2 Hypothesis 2
The second hypothesis was aimed at figuring out to what type of technological regime the coffee grinding and roasting industry can be related. In the analysis I considered the aspects that are mentioned in the conceptual framework. It was proved initially that the level of appropriability is high because the producers in the industry use patents for their innovations. Secondly, the concentration level was medium because several producers were introducing several patents each but there are also many new firms constantly entering the industry. It is
no wonder that such a division between producers exist. This division is a natural content of an industry or a market where major players have a bigger market share compare to other smaller producers. Besides, a few firms that were identified with the largest number of backward citations are the ones that are popular worldwide, which also gives them extra power and authority in establishing the patents.
The net entry and exit was still considered to be high. At the start of an industry the net entry was growing and also growing during other periods, however, the majority of the companies have been entered at the beginning and after that were submitting patents at different times. Hence, overall there are many producers in the industry especially taking into account the timeframe. Such a behavior of firms entering the market refers to the ferment stage of an industry development cycle where the players are entering the market but already less than during the previous stage and the first evidences for the hierarchical stability start to appear. The previous step was also evidence in favor of hierarchical stability that is considered to be medium. There are a few might producers who introduce important innovations and that are commonly used by smaller producers. However, the stability in this case can be seen as subject to disruption. Some firms that started early and after that were not distinguished by active patenting, further they involve in patenting anew.
The high ration of competence-enhancing activities can be explained by a recent start of a patenting activity in the coffee grinding and roasting industry as well as the little amount of innovations compare to other industries. Some industries, in general, are considered as more primitive than others where there is not that much rooms for improvements or that these improvements are not significant and especially in relation to other industries. This way, the given industry is accumulating knowledge now as there is no solid technological base. Overall, all the factors above give a strong evidence to claim that coffee grinding and roasting industry can be related to the Schumpeter Mark I type of innovation activity. This way, the initial prediction is wrong. However, because the development of an industry is a dynamic process, the further technological changes and the movements among market players can lead to the establishment of the Mark II type of a regime.
6.3 Hypothesis 3
The third part of analysis was devoted to the identification of the dominant design of technology in the industry. The patenting activity was expected to disrupt the establishment of a dominant design because it prevents from copying and from making use of technology by the producers other than the patent owners. However, there is evidence for an establishment of a dominant design because one patent (Braun Espanola, 1982) was cited the most often.
