Understanding the open innovation trends: An exploratory analysis of breadth and depth decisions

Bernal, Pilar; Salazar, Idana; Vargas, Pilar Article Understanding the open innovation trends: An exploratory analysis of breadth and depth decisions Administrative Sciences Provided in Cooperation with: MDPI – Multidisciplinary Digital Publishing Institute, Basel Suggested Citation: Bernal, Pilar; Salazar, Idana; Vargas, Pilar (2019) : Understanding the open innovation trends: An exploratory analysis of breadth and depth decisions, Administrative Sciences, ISSN 2076-3387, MDPI, Basel, Vol. 9, Iss. 4, pp. 1-15, https://doi.org/10.3390/admsci9040073 This Version is available at: https://hdl.handle.net/10419/239970 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by/4.0/ administrative sciences Article Understanding the Open Innovation Trends: An Exploratory Analysis of Breadth and Depth Decisions Pilar Bernal 1,*, Idana Salazar 2and Pilar Vargas 2 1Departamento de Dirección y Organización de Empresas, University of Zaragoza, 50005 Zaragoza, Spain 2Departamento de Economía y Empresa, University of La Rioja, 26006 Logroño, Spain; [email protected] (I.S.); pilar[email protected] (P.V.) *Correspondence: [email protected] Received: 3 June 2019; Accepted: 15 September 2019; Published: 20 September 2019   Abstract: The study of firms’ decisions on open innovation has recently attracted the attention of scholars studying the process that firms follow from closed to open models. Extant research has acknowledged that firms tend toward open innovation models and has identified the optimum levels of breadth and depth of openness toward which firms should tend. Surprisingly, there is little evidence on how firms move toward open innovation and whether they follow scholars’ recommendations. In this paper, we investigate the adoption of the open innovation model, studying firms’ decisions on breadth and depth and switching behaviours over time. This paper provides a discussion of firms’ degree of openness and how firms structure and reassess their decisions on open innovation over time. This framework was applied to the Panel of Technological Innovation database that includes data on Spanish innovating firms for the period 2005–2013. Keywords: open innovation; breadth; depth; innovation strategies 1. Introduction As competition intensifies, the advantages that firms gain from the sole use of internal research and development (R&D) investments are not sufficient. This fact forces firms to increasingly open up their innovation processes by combining internal and external knowledge (Dahlander and Gann 2010;Foss et al. 2011;Ferraris et al. 2018). Although innovation was traditionally based on internal and exclusive sources, now firms are adopting an open innovation model (Huizingh 2011), based on “the use of purposive inflows and outflows of knowledge to accelerate internal innovation, and expand the markets for external use of innovation, respectively” (Chesbrough 2003, p. 2). As a result, the notion of open innovation has emerged as an underlying theme in research on innovation (Heimstädt and Reischauer 2019). One of the main assumptions of this stream of research is to consider innovation as a continuum between closed forms and opened forms of innovation, instead of a dichotomous decision. Along this continuum of innovation, firms could range from closed to open, covering different degrees of openness (Dahlander and Gann 2010;Huizingh 2011). Based on this idea, one of the topics that have attracted the attention of scholars is the conceptualization of the degree of openness in terms of firms’ breadth and depth in external knowledge search. On the one hand, the breadth of openness specifies to what extent firms access different external knowledge sources, including customers, suppliers, competitors and universities (Bahemia and Squire 2010;Leiponen and Helfat 2010). On the other hand, the depth of openness refers to how deeply or intensively firms draw from these different external knowledge sources (Laursen and Salter 2006; Drechsler and Natter 2012). Adm. Sci. 2019,9, 73; doi:10.3390/admsci9040073 www.mdpi.com/journal/admsci Adm. Sci. 2019,9, 73 2 of 15 Over the last years, this line of research has focused on providing some insights into both components, particularly through the analysis of their impact on performance (Bahemia and Squire 2010;Inauen and Schenker-Wicki 2011;Parida et al. 2012;Ferraris et al. 2017). From the literature, it is clear that breadth and depth of openness provide potential advantages, such as uncertainty reduction, resource variety and higher technological opportunities (Laursen and Salter 2006;Huizingh 2011), which may allow firms to improve their innovation performance. However, scholars also find that there is a negative side in drawing broadly and/or deeply on external sources. On the one hand, obtaining information from too many and different external knowledge sources, to the extent that each of them involves different characteristics, requirements and rules (Laursen and Salter 2006;Dittrich and Duysters 2007), may lead firms to invest considerable efforts in managing the complexity of the information they obtain. On the other hand, a very deep search may produce redundant information (Katila and Ahuja 2002), leading to overlapping of firms’ knowledge bases (Vanhaverbeke 2006). All these ideas may suggest that there is a point at which breadth and depth of openness become disadvantageous. In keeping with this idea, a number of recent contributions have tried to determine the optimum level of each of these two components, arguing that there should be an increasing trend on firms’ open innovation decisions toward this specific level (Laursen and Salter 2006;Leiponen and Helfat 2010;Salge et al. 2013). Against this background, it is surprising that there is a limited understanding of whether firms have been influenced by scholars’ recommendations on breadth and depth optimum levels and the process that they follow in order to achieve these specific levels. Until now, scholars have analysed firms’ breadth and depth decisions without verifying whether the optimum levels are being reached and without giving consideration to how firms organize their open innovation decisions over a time span (Drechsler and Natter 2012). That fact has prevented us from checking whether there exist a trend toward the optimum levels found by the literature and, therefore, it is not entirely clear to what extent firms take and modify their practices to adhere to scholars’ guidelines about the optimum level of breadth and depth. Based on the lack of studies that may corroborate the relevance of previous findings offered by the open innovation literature, it seems that an analysis of both dimensions over a long time span is particularly needed. To the best of our knowledge, only three studies devoted their attention to the analysis of open innovation adoption over time (Batterink 2009;Poot et al. 2009;Cricelli et al. 2016). However, these studies do not pay attention on breadth and depth decisions and, therefore, do not determine whether firms reach the optimum levels proposed by the literature and how they do it. To address this gap, in this paper we examine, over an extended period of time, the patterns of breadth and depth to find any kind of evidence of a tendency in firms’ decisions about each of these components of open innovation. Although there is a sense that firms are opening their boundaries to external sources, there is little evidence that confirms that the levels of breadth and depth are increasing, or that describes the specific level to which firms tend. We analyse the patterns of breadth and depth by comparing firms’ decisions about the use of external knowledge sources in each year of our sample, and secondly by determining how firms change their breadth and depth decisions over time. We address this issue by using the Panel of Technological Innovation database (PITEC), which contains information about the innovative activity of Spanish firms. This database is especially useful for the purposes of this work, for two main reasons. First, it provides information about the type of partner from which firms are obtaining knowledge. With this information in mind, and following the logic of previous papers (see, for example, Belderbos et al. 2012), we are able to study firms’ breadth and depth of openness. Second, the data provided by PITEC, in contrast to other innovation surveys, is administered on a yearly basis. Hence, the dataset has a longitudinal dimension, providing information about innovation variables from 2003 to 2013. That fact makes the data very suitable to analyse the changes that firms’ breadth and depth decisions show over time. Our contribution to the literature is twofold. First, we are able to analyse, over an extended period of time, the firm’s tendency of breadth and depth. We do this by taking into account the sources of information according to their origin and over time. Until now, firms’ decisions on breadth and depth Adm. Sci. 2019,9, 73 3 of 15 have not been examined over a time span, leading us to study firms’ isolated decisions about the adoption of open innovation models, a fact that has prevented us from engaging in a discussion about open innovation integrally. Second, we investigate how firms’ breadth and depth decisions change over time. This represents a significant advance over the analysis of breadth and depth of openness, which has been abstracted from studying the changes that firms’ decisions on both components show over time. 2. Theoretical Framework The open innovation model describes the recent trend in which firms look for new information and technologies for innovation outside their boundaries (West and Bogers 2017). The main idea behind this model is that the benefits that firms obtain through the isolated use of internal resources are decreasing, forcing firms to draw on knowledge from external sources (Dolfsma and van der Eijk 2017;Santoro et al. 2019). The origins of this idea can be found in those papers that study how firms tend to establish ties with external parties, such as customers, suppliers, competitors and universities (Hamel 1991;Atuahene-Gima 1995;Santoro 2000;Battistella and Nonino 2012). However, according to a seminal work by Chesbrough (2003), who established the basis for the open innovation model, it is argued that firms’ innovation processes can and should be based on a combination of internal and external ideas and mechanisms to create value. In other words, the most important step of the open innovation model is to consider the innovation process as an open system (Laursen and Salter 2006). Building on Chesbrough’s proposal and taking as a starting point its broad nature, scholars have tried to deepen the open innovation model by defining openness in different ways (West and Bogers 2017;Bogers et al. 2017). For instance, a stream of the literature has focused on two dimensions, inbound and outbound, to define the activities that firms can develop when opening their boundaries (Gassmann and Enkel 2004;Cheng and Huizingh 2010;Bianchi et al. 2011;Rangus et al. 2016). Inbound open innovation refers to the process through which firms internally use the knowledge obtained from external sources, while outbound open innovation consists of the external use of the knowledge generated internally (Dahlander and Gann 2010;Santoro et al. 2019). In contrast, another line of research has defined the degree of openness of firms’ external search processes, taking into account two different components: breadth and depth (Laursen and Salter 2006). Recent papers on open innovation have highlighted the importance of distinguishing between these two components since they capture two different external search strategies that show how firms organize themselves when deciding to open up their boundaries (Bahemia and Squire 2010). Since searching for external knowledge is the open innovation pattern most frequently used, an analysis of the different strategies through which firms search for new external ideas may constitute a key factor in exploring firms’ movement toward open innovation (Laursen and Salter 2006). Therefore, it seems that to analyse firms’ behaviour, taking into account both components may allow us to deepen and to integrally address the phenomenon of open innovation. 2.1. Breadth and Depth of Openness: Beneficial and Detrimental Effects The literature has defined breadth as “the number of external sources or search channels that firms rely upon in their innovative activities” (Laursen and Salter 2006, p. 4). When studying the breadth of openness, several studies have emphasized its benefits. Specifically, scholars have argued that the most successful firms are those that search more broadly for knowledge, instead of looking for information narrowly (Ahuja and Lampert 2001;Ahuja and Katila 2004). The point is that, since innovation results are unclear due to turbulence and technological change, using various sources of knowledge in innovation processes can help firms in increasing the probability of maximizing their innovation outcomes. In the same vein, Tether (2002) argued that the higher cost derived from R&D and the evolution of technology make firms try to capture valuable resources by searching widely. Some researchers, such as Love et al. (2014), have demonstrated the positive effect on innovation performance that firms can obtain by increasing the different types of external sources that they use. Adm. Sci. 2019,9, 73 4 of 15 The underlying idea is that to search widely leads organizations to find a variety of resources that allows them to generate new combinations of knowledge and to access a rich range of technological opportunities (Laursen and Salter 2006). Nevertheless, some studies, also recognizing the benefits of breadth, underline that there should be a limit when accessing different knowledge domains (Katila and Ahuja 2002). It has been argued that high levels of breadth can generate increasing cost of handling the integration of knowledge, even overcoming the benefits that it generates. On this basis, there is a stream of research recommending that firms concentrate their efforts on a limited number of external sources. Accordingly, Laursen and Salter (2006) have argued that to understand the norms and routines of a wide number of external knowledge sources, firms must dedicate considerable efforts, especially if they have to identify the ideas that they would like to absorb. In line with this idea, Laursen and Salter (2006) also pointed out that not all ideas come at the correct time to be exploited. Therefore, firms can be wasting energy having access to too many sources. In the same way, due to the high levels of breadth, firms cannot pay the required attention to some ideas that are really relevant. In a nutshell, some scholars have considered that, although breadth is positive, there is also a negative effect that should be taken into consideration. Based on these arguments, the literature has acknowledged that the relationship between breadth and innovation performance may show a non linear effect. Taking this idea as starting point, researchers have gone a step further by determining the optimum level of breadth to which firms should tend. Specifically, Laursen and Salter (2006) found that firms maximize their performance using eleven sources of knowledge of the sixteen that they have available. That is, the optimum point can be found in the use of 68.75% of the sources available. Likewise, Leiponen and Helfat (2010) considered that the maximum return is obtained when using eight sources of a total of twelve, that is, in the use of 66.66% of the external knowledge sources. Meanwhile, Salge et al. (2013) indicated that, in order to maximize firms’ results, the optimum level of breadth could be found in the use of six sources of external knowledge of a total of thirteen, that is, when using 46.15% of the external knowledge sources available. For its part, depth is defined by the literature as “the extent to which firms draw deeply from the different external sources” (Laursen and Salter 2006, pp. 4–5). As in the case of breadth search, scholars have emphasized that drawing deeply from external sources is positive but can also have negative consequences. Maintaining a pattern of interaction with the external environment over time is understood to be positive since it allows firms to better understand the behaviour, habits and rules of the external knowledge sources. That fact allows firms to easily identify the valuable resources and to recombine them in different ways. In a similar vein, Katila and Ahuja (2002) argued that using the same resources repeatedly reduces the probability of error and enhances innovation. This, in turn, leads firms to learn how to access the information they need and to integrate it into their innovation process. Drawing knowledge heavily from external sources could also make firms become familiar with the knowledge they exchange, better identifying the activities that they develop. In this way, firms improve their efficiency, avoiding unnecessary steps (Eisenhardt and Tabrizi 1995;Katila and Ahuja 2002). Accordingly, scholars such as Laursen and Salter (2006) have demonstrated that firms focusing on a depth search will be more innovative, because they build long-lasting and secure relationships with their external environment. However, drawing deeply on external ideas may require firms to invest substantial time and resources, a fact that negatively affects innovation performance (Laursen and Salter 2006;Lakhani et al. 2012). In addition, scholars have argued that using the same knowledge over time can result in overlapping. To ensure that the same knowledge will generate new ideas, firms must invest significant time and attention, which might not be worth such effort. To use the same knowledge during a long period of time can also generate rigidity problems (Katila and Ahuja 2002). Firms will try to solve problems based on their previous responses. However, because a strategy allowed a firm to get solutions does not mean that this approach should always be used. This can make firms focus on a strategy and then invest resources in it, even though it may not be appropriate (Laursen and Salter Adm. Sci. 2019,9, 73 5 of 15 2006). For all the reasons presented above, scholars have found a non linear relationship between depth and innovation performance. As with breadth, scholars have tried to determine the optimum level of depth that allows firms to maximize their performance. Laursen and Salter (2006) concluded that firms should draw deeply on three sources of external knowledge, of the sixteen that they have available, in order to maximize their results. That is, the optimum point can be found in the deep use of 18.75% of the external sources available. In such a context, in which the non linear effects of breadth and depth are generally accepted and the optimum levels of breadth and depth appear to be a much-debated topic, it is remarkable that there is little evidence on how the conclusions on these issues have influenced firms’ trend on breadth and depth decisions. In other words, most of the literature has abstracted from analysing whether the confluence of positive and negative effects obtained and the optimum level of breadth and depth that scholars have found have conditioned firms’ pattern of behaviour, which beforehand is unclear. Therefore, analysing to what extent firms’ intensity on breadth and depth has changed over time seems to be necessary in order to avoid confusion around the adoption of open innovation strategies and to confirm the validity of the open innovation model (Chesbrough et al. 2006;Bianchi et al. 2011;West and Bogers 2017). 2.2. Breadth and Depth of Openness: Evidence on Their Adoption Over the last years, researchers have consistently argued that firms are replacing their closed innovation models by open ones (Huston and Sakkab 2006;Bianchi et al. 2011). In reaching this conclusion, scholars have used different methods of analysis and have focused on specific industries, periods of analysis and countries. In the seminal study of open innovation, Chesbrough (2003) used case studies to document how the business model of companies belonging to high-technology industries, such as IBM and Intel, has changed from closed to open forms. Subsequent research has confirmed this insight, also using case studies but applying them to other sectors. For instance, Dodgson et al. (2006) and Huston and Sakkab (2006) studied the specific case of Procter and Gamble, documenting how this firm develops open innovation practices. Gassmann (2006) conducted a case study of 15 companies in Germany, Liechtenstein, Switzerland, and The Netherlands that belong to different industries. Chesbrough and Crowther (2006) focused on the specific case of 12 companies that operate in “industries outside high technology that are early adopters of the concept.” By studying firms’ practices, they confirmed that open innovation practices are also present in more traditional and slow-growing industries, a fact that confirms the validity of the Open Innovation model. Regarding the financial service industry, Fasnacht (2009), by studying the case of 18 firms belonging to this sector, identified a shift from closed to open models. While case studies are common when analysing open innovation adoption trends, some researchers have longitudinal data to analyse firms’ behaviour. Such is the case of Van de Vrande et al. (2009), who used data from 605 Dutch companies to document the adoption of open innovation in small and medium-sized enterprises (SMEs), confirming the firms’ trend toward the use of external knowledge sources. Salmi et al. (2008) focused on 59 Finnish firms to conclude that open innovation practices are present although they are still low. Poot et al. (2009) incorporated a longitudinal perspective in the analysis of the adoption of open innovation practices using the Dutch Community Innovation Survey. More recently, Bianchi et al. (2011) explored firms’ intensity on open innovation in a sample of 20 companies belonging to the biopharmaceutical industry, by distinguishing between inbound and outbound dimensions. Besides these limited contributions, scholars have not systematically studied the adoption of open innovation models by taking into account firms’ degree of openness. Researchers studied firms’ trend toward open innovation but without a detailed specification on firms’ decisions, although this would allow us to better understand the intensity of the adoption of open innovation models. As Chiaroni et Adm. Sci. 2019,9, 73 6 of 15 al. (2011) have argued, “understanding the anatomy of the process from closed to open innovation requires identification of the dimensions along which change occurs.” In this sense, the analysis of firms’ breadth and depth of openness can help us understand the specific actions that open innovation firms have taken. This paper contributes to fill this gap by analysing how firms’ breadth and depth evolve over time. 3. Data and Descriptives 3.1. Sample Our empirical analysis is based on the Technological Innovation Panel database (PITEC). The database has been developed by the National Institute of Statistics (INE), with the support of the Spanish Foundation for Science and Technology (FECYT) and the Spanish Foundation for Technological Innovation (COTEC). The structure of PITEC is based on the Community Innovation Survey (CIS) framework, which in turn follows the guidelines of the Oslo Manual (OECD 2005). 1 PITEC is a panel data survey that provides annual information about the innovative behaviour of a large sample of Spanish firms from 2003 and has been previously used for several purposes (Fariñas and López 2007;Molero and Garcia 2008;De Marchi 2012). PITEC has two main advantages for this study. First and most important, the PITEC questionnaire (Appendix A), in contrast to the majority of work on open innovation, is administered on a yearly basis. This implies that our analysis will provide information about the innovative activity of Spanish firms for an extended period of time. Second, the survey contains information about firms’ decisions on knowledge sourcing. Specifically, the data include the type of information sources to which firms have access, which is necessary to analyse breadth and depth. From these data we used information for the period 2005 to 2013. We selected our final sample by following three steps. First, we restricted our sample to firms engaging in innovative activities (Laursen and Salter 2006). 2 Second, we focused our analysis on manufacturing and service firms. Third, we excluded those firms with no information for the main variables, those that have suffered problems associated with mergers and acquisitions, and those that are public or newly created. This means that we were left with 57,984 observations. Table 1shows the distribution of the firms depending on size. 3 As can be seen, about 40% of companies have between 10 and 49 employees (small enterprises), closely followed by medium enterprises (around 32% of the sample). Table 1. Firms by size. N % Micro-enterprises 5761 9.94 Small enterprises 23,129 39.89 Medium enterprises 18,358 31.66 Large enterprises 10,736 18.51 Total 57,984 100.00 1 The data set, the questionnaire, and the description of each variable is available at: http://icono.fecyt.es/PITEC/Paginas/por_ que.aspx. In order to prevent firms from being identified, some variables are anonymised. L ó pez (2011) shows that the expected biases due to this anonymisation are small through the comparison of regressions that use original and harmonized data alternatively. 2 Innovators are those firms that have developed product or process innovation and those that have attempted it and have failed in doing so. 3 Classification made according to the criteria established by the European Commission Regulation (CE) No 800/2008 of 6 August 2008 (DOUE L214/3 of 9 August 2008), which defines the requirements for three categories of companies: microenterprises, comprising those that employ fewer than 10 workers; small businesses which includes those that employ 10 to 49 workers; and medium enterprises, comprising those that employ between 50 and 249 workers. Adm. Sci. 2019,9, 73 7 of 15 Similarly, Table 2shows the distribution of the firms according to their main activity. Table 2. Firms by activity. N % Manufacturing 36,468 62.89 Services 21,516 37.11 Total 57,984 100.00 3.2. Variables Breadth and depth are the variables in this study. Both variables are constructed as a combination of eleven sources of knowledge for innovation. These information sources are: firms that are part of the same group, suppliers, customers, competitors, consulting companies, universities, the public sector, technological centers, conferences, journals and professional associations. Regarding breadth, and in line with Laursen and Salter (2006), each of the eleven sources is coded as a binary variable that takes the value 0 if the firm has indicated that this source of information is not used, and 1 if the source is used. After that, the eleven sources are added up and, by doing this, the variable takes the value of 0 when no knowledge sources are used and 11 when all sources are used. Therefore, firms that use a higher number of sources are more open in terms of breadth. To construct our depth variable, and also in line with Laursen and Salter (2006), each of the eleven sources takes the value 1 if the firm indicates that it uses the source to a high degree and 0 in the case of low, medium, or no use of the given source. In a similar way as in the case of breadth, the eleven sources are added up so the variable takes the value of 0 when no knowledge sources are used to a high degree and 11 when all sources are used to a high degree. Again, firms that obtain higher values are those that are more open in terms of depth. 4. Results In order to analyse in some detail the patterns of firms’ decisions about the use of external knowledge sources, we compare the distribution of firms in different degrees of breadth and depth across each year of our sample. Descriptive data for observations of firms’ breadth decisions are shown in Table 3. Table 3. Proportions of sample in each level of breadth of openness by time period. Breadth 2005–2013 (No) 2005 2006 2007 2008 2009 2010 2011 2012 2013 0 4.03 (2339) 0.03 0.35 0.24 4.12 8.81 8.76 5.81 5.07 5.26 1 8.06 (4671) 8.07 9.98 10.79 8.56 6.42 7.29 6.76 6.89 6.25 2 5.34 (3099) 5.70 5.9 6.21 5.44 4.90 4.11 4.8 5.34 5.41 3 5.22 (3025) 6.28 5.88 5.47 4.90 4.69 4.67 4.85 4.97 4.8 4 5.95 (3449) 7.28 6.63 6.35 5.38 5.33 5.39 5.36 5.63 5.73 5 7.32 (4243) 8.97 7.75 7.39 7.34 7.07 6.87 6.38 6.38 6.99 6 7.61 (4414) 8.5 7.91 8.15 8.22 7.58 6.82 6.94 6.96 6.63 7 8.92 (5172) 10.34 9.51 9.42 8.98 8.75 8.35 8.42 8.24 7.26 8 9.98 (5784) 10.73 10.66 10.43 10.34 9.68 9.77 9.79 8.76 8.72 9 7.25 (4206) 7.44 7.29 7.17 6.92 7.22 6.85 7.07 7.68 7.88 10 6.95 (4028) 6.31 6.32 6.54 7.12 7.48 7.35 7.83 6.93 6.97 11 23.38 (13,554) 20.35 21.83 21.83 22.68 22.07 23.77 25.99 27.14 28.09 As can be seen in the table, and focusing on the results of the entire period, firms tend to get medium and high levels of breadth. Around 23% of firms in our sample use the eleven sources of innovation that are available to them, approximately 8% have access to six sources, and only 4% indicate that zero knowledge sources are used. Table 4shows the descriptive data for firms’ depth decisions. Adm. Sci. 2019,9, 73 8 of 15 Table 4. Proportions of sample in each level of depth of openness by time period. Depth 2005–2013 (No) 2005 2006 2007 2008 2009 2010 2011 2012 2013 0 23.35 (13,542) 20.69 22.09 23.32 25.15 26.49 27.25 22.27 21.22 0.66 1 30.80 (17,861) 32.94 32.33 30.93 29.94 28.43 28.99 31.58 31.15 30.76 2 20.73 (12,020) 21.65 21.57 21.03 20.48 20.76 19.30 20.31 20.08 20.99 3 12.18 (7061) 12.57 11.85 12.11 12.27 11.91 11.63 12.01 12.88 12.61 4 6.16 (3571) 5.95 6.14 6.34 5.72 5.78 5.9 6.47 6.6 6.95 5 3.34 (1939) 3.13 3.1 3.06 3.24 3.25 3.57 3.68 3.54 3.87 6 1.66 (961) 1.62 1.5 1.58 1.44 1.67 1.59 1.73 2.28 1.73 7 0.88 (513) 0.69 0.69 0.82 0.91 0.87 0.77 1.17 1.12 1.14 8 0.46 (267) 0.43 0.37 0.47 0.43 0.43 0.46 0.38 0.58 0.67 9 0.21 (122) 0.12 0.17 0.17 0.27 0.25 0.29 0.18 0.21 0.25 10 0.07 (38) 0.01 0.03 0.03 0.03 0.09 0.06 0.09 0.14 0.19 11 0.15 (89) 0.2 0.15 0.17 0.13 0.07 0.18 0.13 0.19 0.17 Unlike what happens in the case of breadth, firms tend to have low levels of depth. Approximately 30% of firms deeply focus on one source of knowledge, and 23% indicate that no knowledge sources are used to a high degree. About 2% of firms have deep access to six sources of knowledge, and only about 0.15% of firms intensively draw from the eleven sources of knowledge for which they are asked. Tables 3and 4also show the tendency of firms’ breadth and depth decisions over time. As can be seen in Table 3, there is slight evidence of an increase of the percentage of firms focusing on eleven sources of knowledge, and a decrease in the case of firms using one or zero sources of knowledge. Regarding the changes in firms’ depth decisions, Table 4shows that, during the period of analysis, there is a modest increase of the percentage of firms in each category of depth, followed by a moderate decrease. In addition to the above, we investigate how firms’ breadth and depth decisions change over time using a transition matrix (see Tables 5and 6) that presents some data that will guide us in explaining the evolution of both constructs from 2005 to 2013. This matrix shows the switching behaviour of firms between different levels of breadth and depth, allowing us to examine whether there is any systematic change in firms’ decisions on open innovation over time. Table 5. Transition matrix of breadth of openness. End Category Starting Category 0 1 2 3 4 5 6 7 8 9 10 11 0 30.93 13.77 6.91 5.13 5.95 5.73 4.34 5.13 4.95 2.95 2.48 11.73 1 2.76 41.08 7.9 5.38 5.43 5.75 4.72 5.51 4.56 2.41 2.17 12.32 2 2.01 11.05 40 6.9 6.27 6.31 4.42 3.94 4.34 2.37 1.89 10.49 3 1.61 8.28 7.87 42.5 7.13 6.34 5.31 4.2 4.41 2.39 2.1 7.87 4 1.62 6.61 5.43 7.26 44.36 8.81 5.39 4.78 4.28 2.26 1.91 7.3 5 1.59 5.49 3.38 5.17 8.03 46.26 7.51 5.83 4.56 2.34 2.28 7.57 6 1.18 3.54 2.44 3.18 5.1 8.61 50.49 8.17 5.87 2.47 2.17 6.77 7 0.91 3.06 2.01 2.78 3.15 5 8.47 51.91 8.99 4.18 2.73 6.82 8 0.82 2.72 2.02 2.02 2.02 3.52 5.03 9.69 53.79 6.16 3.44 8.76 9 0.64 2.16 1.46 1.37 1.49 2.52 3.19 5.97 10.79 52.59 6.92 10.9 10 0.43 2.19 1.33 1.04 1.47 1.38 2.28 3.8 5.99 9.37 52.51 18.21 11 1 3.36 2.01 2.03 1.79 2.38 2.3 3.06 4.42 4.03 6.1 67.53 Adm. Sci. 2019,9, 73 15 of 15 L ó pez, Alberto. 2011. The effect of microaggregation on regression results: An application to Spanish innovation data. The Empirical Economics Letters 10. Love, James H., Stephen Roper, and Priit Vahter. 2014. Dynamic complementarities in innovation strategies. Research Policy 43: 1774–84. [CrossRef] Molero, Jose, and Antonio Garcia. 2008. The innovative activity of foreign subsidiaries in the Spanish Innovation System: An evaluation of their impact from a sectoral taxonomy approach. Technovation 28: 739–57. [CrossRef] OECD. 2005. Oslo Manual, Guidelines for Collecting and Interpreting Innovation, 3rd ed. Paris: OECD. Parida, Vinit, Mats Westerberg, and Johan Frishammar. 2012. Inbound open innovation activities in high-tech SMEs: The impact on innovation performance. Journal of Small Business Management 50: 283–309. [CrossRef] Poot, Tom, Dries Faems, and Wim Vanhaverbeke. 2009. Toward a dynamic perspective on open innovation: A longitudinal assessment of the adoption of internal and external innovation strategies in the Netherlands. International Journal of Innovation Management 13: 177–200. [CrossRef] Rangus, Kaja, Mateja Drnovsek, and Alberto Di Minin. 2016. Proclivity for open innovation: Construct development and empirical validation. Innovation 18: 191–211. [CrossRef] Salge, Torsten Oliver, Tomas Farchi, Michael Ian Barrett, and Sue Dopson. 2013. When does search openness really matter? A contingency study of health-care innovation projects. Journal of Product Innovation Management 30: 659–76. [CrossRef] Salmi, Pekka A. S., Carl J. Kock, and Marko T. Torkkeli. 2008. Open innovation practices in Finnish firms: A survey. Paper Presented at the 1st ISPIM Innovation Symposium, Singapore, December 14–17. Santoro, Michael D. 2000. Success breeds success: The linkage between relationship intensity and tangible outcomes in industry-university collaborative ventures. Journal of High Technology Management Research 11: 255–73. [CrossRef] Santoro, Gabriele, Alberto Ferraris, and Daniel John Winteler. 2019. Open innovation practices and related internal dynamics: Case studies of Italian ICT SMEs. Euromed Journal of Business 14: 47–61. [CrossRef] Tether, Bruce S. 2002. Who co-operates for innovation, and why: An empirical analysis. Research Policy 31: 947–67. [CrossRef] Van de Vrande, Vareska, Jeroen P.J. de Jong, Wim Vanhaverbeke, and Maurice de Rochemont. 2009. Open innovation in SMEs: Trends, motives and management challenges. Technovation 29: 423–37. [CrossRef] Vanhaverbeke, Wim. 2006. The interorganizational context of open innovation. In Open Innovation: Researching A New Paradigm. Oxford: Oxford University Press, pp. 205–19. West, Joel, and Marcel Bogers. 2017. Open innovation: Current status and research opportunities. Innovation 19: 43–50. [CrossRef] © 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).