1
The moderating effect of big data analytics in
mitigating R&D diminishing returns to the
innovation performance
Master Thesis MSc Business Administration – Change
Management
Dionisius Denizar S3347877 Antaresstraat 21-01
9742LA, Groningen, The Netherlands d.denizar@student.rug.nl
University of Groningen Faculty of Economics and Business
MSc Business Administration – Change Management
June 2018
Supervisor: dr. John Q. Dong Co-assessor: prof. dr. Jordi Surroca
2 ABSTRACT
The emergence of big data analytics in recent days as an information processing technology has become one of the interesting topic to be explored in business research. Especially the empirical research on how big data analytics could create business value in firms. The business value of big data can be investigated from the perspective of R&D management and innovation process. The study is based on the diminishing return characteristic of firm’s R&D investment, and this study theorize how the use of big data analytics, and its interaction with firm’s existing R&D resources, would decreasing the speed of R&D investment diminishing return, by referencing to the system theory. Using firm level dataset from 2016 Mannheim Innovation Panel database, this study shows that the diminishing return of R&D investment also happens in innovation process. This study also supports that big data analytics usage in firms would mitigate the diminishing return of R&D investment, and firms that use higher level of big data analytics decrease the rate of R&D investment diminishing return. Interpretation and discussion of the findings along with the practical and theoretical implication also will be the part of this study.
3 TABLE OF CONTENTS ABSTRACT ... 2 1. INTRODUCTION ... 4 2. THEORETICAL BACKGROUND ... 7 2.1. R&D investment ... 7
2.2. Big data analytics utilization ... 8
2.3. System theory and complementarity perspective ... 9
3. HYPOTHESES ... 10
3.1. The Relationship between R&D Intensity and Innovation Performance ... 10
3.2. Role of Big Data Analytics in Innovation Process by Reducing Diminishing Returns to R&D Intensity ... 11 4. METHODS ... 14 4.1. Data ... 14 4.2. Measures ... 14 5. RESULTS ... 17 5.1. Hypotheses Testing ... 17 5.2. Robustness Check ... 20 6. DISCUSSION ... 22 6.1. Main Findings ... 22
6.2. Contributions and Implications ... 22
6.3. Limitations and Future Studies ... 24
7. CONCLUSION ... 26
REFERENCES ... 27
4
1. INTRODUCTION
Nowadays, the incorporation and adoption of external knowledge has become vital for firm’s product innovation (Chao-Ton et al., 2006; Robert & Candi, 2014; Lichtenthaler, 2016; Zhan et al., 2016). In addition, the rapid advancement of technology recent days in all over the world has generated remarkable growth of data volume, variety, and velocity which can be called as “big data” that offer new possibilities for innovation process (Brown et al., 2011). Such data can be obtained in many ways in this modern era of Internet-of-Things (IOT), for example, the everyday technological products that consumer use have its own usage record and usage pattern that can be retrieved by firms that manufactured such products; another example of the vast amount of data can be taken from the sensors or networks of machines in firm’s manufacturing plant to analyse the performance from manufacturing machines. Due to the rapid development surrounding big data, firms should adapt themselves to accommodate the big data importance into organizational aspects, especially the R&D functions of the firm. Hence, this study will be important in adding new perspective on how big data analytics usage in a firm, which is a novel way to capture and process knowledge, would change the ordinary R&D process.
The emergence of big data has heightened organizations’ demand for business analytics to efficiently manage and use big data for business purpose. Business analytics interpreted as the “extensive use of data, statistical and quantitative analysis, explanatory and predictive models, and fact-based management to drive decisions and actions” (Davenport & Harris, 2007). The adoption of big data analytics is associated with signficant financial investments for firms (Müller et al., 2018). For example, the three-year total of ownership cost for an IBM PureData System for Analytics (applicance for big data processing with eight server racks and 1,500 terabytes of storage capacity) is approximately to US$ 39 million, and the total overall costs for a comparable Cloudera Hadoop cluster (with 750 nodes), for the same period with IBM PureData System sum up to more than US$ 50 million (Asthana & Chari, 2015). Hence, there is a main question whether the usage of big data analytics would bring more financial gain or would be a sunk cost to the firm.
5
vast amount of data, management of product portfolio, and R&D strategy alignment with business strategy (El Sawy et al., 2016; Leiponen & Helfat, 2011). Most studies regarding R&D investment have only focused on the relationship between R&D investment and the firm performance in general (e.g. Lin et al., 2006; Kostopoulos et al., 2011; Tsai, 2001; Belderbos et al., 2004). These studies also only looked into the linear perspective of the R&D process, assumed that R&D investment could be done continuously and neglecting the limitation of surrounding areas of R&D especially the management capability to maintain the focus on R&D as the driver on the innovation process in the firm. The study of curvilinear relationship of R&D investment with innovation performance was done previously by Ravichandran et al. (2017) to observe the possibility of diminishing return in R&D investment, and this study will provide another insight of such relationship, which will be described in the theoretical section of this study.
Past studies examined the effects of IT on many core organizational mechanisms and organized connection with the improvements in productivity, profits, customer satisfaction, and Tobin’s q (Dedrick et al., 2003; Mithas et al, 2012, 2016; Ravichandran & Lertwongsatien, 2005; Tambe & Hitt, 2012; Ho et al., 2017), while study aimed at assessing the effect of IT on knowledge intensive processes such as R&D and product development is till in a relatively beginning stage (Nambisan et al., 2017). Study regarding the effect of IT on innovation performance is limited to the general terms of IT, rather than specific technology (see Ravichandran et al., 2017). The big data analytics, which will be a specific technology investigated in this research, have been one of the major technology that enables innovation, competition, and productivity (Brown et al., 2011). Recent studies regarding big data analytics is more dominated by technical aspects and solutions to analyze big data (see Chen et al., 2012; Baesens et al., 2014; Lau et al., 2016; Günther et al., 2017), and according to Abbasi et al. (2016), there is lack of empirical studies assessing the business value of big data analytics.
6
firm-level? (2) How does the various levels of big data analytics usage in a firm would moderate the relationship between R&D investment and innovation performance? Because of the importance of innovation in firm growth, it is crucial to orderly theorize and empirically investigate the role of big data analytics utilization in influencing innovation performance and bring added value to the firm, which is the aim of this study.
This paper will contribute to provide more insights to the R&D management field, especially in giving another perspective of R&D diminishing return from innovation performance side. Another contribution of this study is to fill the study gap about specific technology that is possible to mitigate the diminishing return rate of R&D investment. For the practical contribution, this study provides new insights regarding how firms can optimize the amount R&D investment that they can disburse. In addition, the study would give managerial perspective on the benefits of using big data analytics as R&D resources to increase innovation performance of firms, and in what level the firms should use big data analytics.
7
2. THEORETICAL BACKGROUND 2.1. R&D investment
R&D investment is one of the two major business decisions alongside marketing investment that intensively influences a firm’s performance, although the performance indications of those strategic decisions are still obscured and vary according to performance measures and possibly diverse between industries. Firms may utilize a strategic decision (e.g. R&D investment) to gain competitive advantage. The only way for them to do so is to continuously create innovations (Lin, Lee, and Hung, 2006).
In addition to business decision function, R&D investment mainly functions as a necessary input of absorptive capacity in the innovation process in previous researches (Cohen & Levinthal, 1990; Lin et al., 2012; Kostopoulos et al., 2011). Absorptive capacity, according to Cohen & Levinthal (1990), is the “ability to recognize the value of new information, assimilate it, and apply it to commercial ends.” From the definition coined by Cohen & Levinthal (1990), the absorptive capacity is a byproduct of R&D process, which could increase a firm’s ability to innovate by the process of effective exchange from the firm’s existing knowledge to the improvement of firm’s current products and technologies (Ritala & Hurmelinna-Laukkanen, 2013). Following the approach of previous research (Cohen & Levinthal, 1990; Tsai, 2001; Fosfuri & Tribo, 2008; Youndt et al., 2004), this research uses R&D intensity (R&D investment as a proportion of sales) as the operationalization of absorptive capacity, because the more a firm invests in R&D, the more it builds human capital that enables innovation process. It is implied that a firm’s high R&D intensity level leads to higher internal technological capacity and more linkage exploitation with other firms and (consequently,) knowledge creation (Lin et al., 2012). Absorptive capacity should not be the sole goal of a firm. It can generate important organizational outcomes, such as innovation performance (Fosfuri & Tribo, 2008; Tsai, 2001) This will be the focus of our/this research.
8 2.2. Big data analytics utilization
Big data analytics is emerging as vast academical research attention has increased in recent years. (McAfee & Brynjolfsson, 2012; Chen, Chiang, & Storey, 2012; Wamba et al., 2017). According to Bumblauskas, et al. (2017), the term “big data” is usually associated with the collection and analysis of large datasets. The exact definition varies between studies/authors due to the ambiguity and flexibility of the term. Maynika et al. (2011) defined big data as “datasets whose size is beyond the ability of typical database software tools to capture, store, manage, and analyze.” On the other hand, Mauro et al. (2016) proposed the definition of big data as “the information asset characterized by such a high volume, velocity, and variety to require specific technology and analytical methods for its transformation into value.” From the two written definition in this research and several studies (Chern et al., 2005; McAfee & Brynjolfsson, 2012; Russom, 2011; Bumblaukas et al., 2017), we derive the following characterizations which help distinguish “big” datasets:
1) Volume of the datasets
2) Velocity, which the speed of data flows from sources such as business processes, machines, networks, and human interaction with technological artifact (social media, cloud computing);
3) Variety of data sources and types, which can be structured or unstructured (e.g., free-form text, sensor data, graphics, media files);
4) Veracity as the uncertainty of data;
5) Value, which the datasets can generate economically worthy insights and or benefits through extraction and transformation
9
2.3. System theory and complementarity perspective
An approach that will be used in this research is system theory which was utilized by Dong and Yang (2017) in a previous study. System theory implies that all systems are composite things, which have interacting components within the systems themselves (Ackoff, 1971; Checkland, 1999). It follows that a system should own properties that are derived from interactions among its components (Churchman, 1971; Gharajedaghi, 2006). The main point of system theory is centered on the notion of synergies, which describes the phenomenon of super-additive value or sub-additive cost generated by interactions among components of a system (Tanriverdi & Venkatraman, 2005). A system is said to have super-additive value if the combined value of two components is greater than the sum of each component’s individual value. In contrast, a system is said to have sub-additive cost if its cumulative cost is less than the sum of its individual costs (Farjoun, 1998; Robins & Wiersema, 1995).
Beside the system theory approach in observing the relation between big data analytics and existing R&D resources, this study also consider a recent research by Ravichandran et al. (2017) stated that IT is complementary to the R&D processs. This perspective posits that two types of resources complement each other when the marginal return of one resource increases along with the level of the other. Hence, the value of having both resources at the same time exceeds the sum of the value of having each resource individually. In the case of value creation in business, IT usage is considered complementary to the organizational practices, for instance, incentive pay and human resources analytics (Aral et al., 2012). Ravichandran et al. (2017) also argued that IT complements R&D resources by overcoming the challenges in scaling R&D efforts with the enablement of better information processing capacity, better communication between organizational members and units, and better integration of knowledge. They also suggested that firms should have its own information processing capacity for innovation activities, which are more data-driven and data-intensive. Hence, the role of big data analytics in increasing a firm’s information processing capacity is essential to innovation activities management in the firm and later the increase of R&D productivity, particularly at higher levels of R&D.
10
3. HYPOTHESES
3.1. The Relationship between R&D Intensity and Innovation Performance
R&D activities done by a firm can contribute to the innovation process due to the inventive nature over which technological knowledge is built and scientific discoveries are made. Eventually, the new product/service that comes from the innovation service would affect the firm growth. Thus, investment in R&D is actually an investment in intangible assets that support the long-term growth of the firm (Lin et al., 2011).
There are previous studies which show positive correlation between R&D investment with firm performance (Branch, 1974; Erickson & Jacobson, 1992; Long & Ravenscraft, 1993). On the other hand, there are also some research which stated that R&D investment has a significant negative effect on firm performance (Ayadi, Dufrene, & Obi, 1996; Gou et al., 2004; Lin & Chen, 2005). Most of the studies regarding the relationship between R&D investment and firm performance considered linear relationship between them. However, empirical evidence showed that this relationship could not last perpetually (Huang & Liu, 2005). In other words, the relationship will either lead to an endless increase in R&D investment or no R&D investment at all, both of which are irrational (Lin et al., 2011).
According to Lin et al. (2011), there are three important factors which might be neglected in the past studies dealing with R&D-firm performance linear relationship. First, R&D investment will draw significant expenses to a firm. Increase in R&D investment may bring profit, but it will also increase the firm’s aggregate R&D cost (Shy, 1995). Firm performance cannot be simply considered as linearly dependent on R&D investment per se. The second factor is based on the theory of the growth of a firm, which stated that it is impossible for a firm to extend itself endlessly due to the limitation of management capabilities (Penrose & Pitelis, 2009). Hence, although R&D investment generates positive value, it is unreasonable to go beyond the limit(s) of management capabilities and persist endlessly. Finally, the third factor is based on Foster’s (1986) S-curve theory, which high R&D investment cannot definitely achieve its optimal performance, because when R&D investment reaches a certain critical point, productivity will plunge and cause “decreasing returns to R&D” (Becker & Speltz, 1983)
11
According to the United States Census Bureau (2014), more than 50% of R&D budget is allocated to employee salaries and other employee benefits1. This phenomenon also
accompanies an increase in the number of knowledge workers involved in R&D process. Hence, additional R&D investments tend to be ineffective if not coupled with appropriate levels of matching capabilities in other areas (Henderson & Cockburn, 1994; Argyres & Silverman, 2004). According to Ravichandran et al. (2017), when the R&D functions in a firm grow, it would experience functional differentiation that leads to decreasing level of creativity, collaboration, and knowledge sharing within R&D functions. Another explanation for diminishing return on R&D investment is explained by Faff et al. (2012). There are three specific factors justifying the diminishing return of R&D investment from the financial perspective. First, there is evidence that R&D-intensive firms have a higher systematic risk compared to less R&D-intensive firms, mainly due to greater intrinsic business and operating risk. Second, overinvestment in R&D can raise the cost of borrowing in the financial sector, since the lenders demand a guarantee for the lack of marketable collateral and this can lead to a non-optimal capital structure for an R&D-intensive firm. Finally, overinvestment in R&D would lead to underinvestment in complementary assets if the firms deal with resource constraints for strategic investment. The underinvestment of complementary assets might reduce the value obtainable from strategic assets created during the R&D process.
Considering the arguments from stated relevant studies, this research seeks to test the relationship between R&D intensity and innovation performance is and observe whether the diminishing return of R&D intensity also occurs in the product/service innovation process in firm-level. It is highly likely that this relationship follows an inverted Ushape. This leads to the following hypothesis:
H1: R&D Intensity has an inverted U-shaped relationship with innovation performance.
3.2. Role of Big Data Analytics in Innovation Process by Reducing Diminishing Returns to R&D Intensity
Knowledge and communication exchange in an R&D process are intensive in nature. Hence, the use of big data analytics as one of IT tools that aggregates data from the internal and external side of the firm can be a significant enabler in managing R&D. In a previous
12
research, Ravichandran et al., (2017) stated that IT contributes to R&D resources by tackling the challenges in scaling R&D efforts through the enablement of better information processing capacity, better communication between organizational members, and better knowledge integration in a firm. Thus, from the perspective of system theory, the combination between high-level R&D intensity and the big data analytics usage would boost the amount of innovation performance. The usage of big data analytics in the innovation process itself may increase information processing capacity within the firms, and it would help existing R&D resources such as R&D human resources, which make up the majority of R&D expenditure. Big data analytics helps ease data and information exchange within R&D functions in a firm. For example, in new product development big data analytics helps process the vast amount of customer insights from the firms’ existing products. It reduces the effort of gathering and curating customer insights introduced by surveying and other slower and more error-prone methods requiring manual data processing. Another example which shows that R&D functions can benefit from big data analytics is when the firm’s researchers want to improve existing products. They can track the consumer behavior data from a real-time dashboard available in the firm and identify the improvements that can be done for a particular product. Big data may also enable R&D to respond to unexpected events before they become crises by providing early warning . For instance, a firm can utilize big data analytics for social media monitoring in order to gauge customer reaction of newly launched products and keep an eye on potential issues regarding the new products (Blackburn et al., 2017).
The innovation performance of a firm can be boosted using big data analytics as a complement to the R&D process for a diverse type of products. This is especially important for firms with export orientation, for instance, to formulate products that are tailored for each target market. Blackburn et al. (2017) took an example from a consulting company that maintains vast databases of worldwide regulations related to food, pharmaceuticals, medical devices, and packaging along with raw materials used in these areas. The data are translated into 40 languages and they permit R&D organizations which use their services to develop products with greater assurance that the product will be approved by regulatory agencies in the target market.
13
input of innovation process, thus more R&D investment can drive more innovation. On the other side, the complexities and rigidity rise rapidly as R&D investment scales up, resulting in the diminishing returns to R&D investment. The role of big data analytics usage would increase higher information processing capacity and knowledge acquisition in R&D process, which is complex and rigid. The role of big data analytics also adding the superadditive value to the relationship between R&D investemnt and innovation performance by complementing existing R&D resources. Hence, it would flatten the inverted-U shaped relationship depicting returns to R&D investments. The second hypothesis is therefore:
14 4. METHODS 4.1. Data
The data from German Innovation Survey 2016 is used to test our hypotheses. The Centre for European Economic Research (ZEW) has conducted this survey on an annual basis since 1993. This survey is also commissioned by the German Federal Ministry for Education and Research (BMBF) in cooperation with Infas (Institute for Applied Social Science) and the Fraunhofer Institute for Systems and Innovation Research. In line with the Oslo Manual (OECD and Eurostat, 2005), the survey was based on a stratified random sample of firms with five or more employees from a broad area of economic activities. The data of this survey is sourced from the same set of firms, which is sampled every year to allow a panel structure. The 2016 dataset had questions on the use of digitalization and obstacles that hinder the implementation of new digital technologies, as the public attention towards digitalization and its likely impacts on innovation has been growing again (Behrens et al., 2017), which become a focus of this research. Hence, this survey data set is also called Mannheim Innovation Panel (MIP) database. The sample size of German Innovation Survey 2016 is 25,392 firms with 5,294 firms completed the survey questionnaire, leading to 23.6 percent response rate. The sample was drawn from MIP database with the criteria that the data for R&D investment and innovation performance are available. Finally, it results a sample of 4,685 firms with some variables containing missing values.
MIP database is the German part of the survey that is incorporated into Community Innovation Survey (CIS), which has been conducted in several member states of the European Union (EU) since 1993 and using the Oslo Manual as the guideline to conduct the surveys. The Oslo Manual guides the similar instrument usage on each survey incorporated in CIS; the understandability, the reliability, and the validity of the survey questionnaires have been extensively tested using the pilot test in each country. (Ziegler & Nogareda, 2009). MIP database has been used in past research (e.g., Bohringer et al., 2012; Dong & Netten, 2017) Horbach, 2008; Kaiser, 2002; Ziegler & Nogareda, 2009; Rennings & Rammer, 2011).
4.2. Measures
4.2.1. R&D intensity. R&D intensity a firm/a company’s total R&D investment as a share of total sales. This calculation of R&D intensity is in line with the definition in Eurostat (2016)2 and Hughes (1988), which is also the primary driver of innovation and also one of key indicator
15
used to monitor resources devoted to science and technology. This variable is the independent variable in this study
4.2.2. Innovation performance. Considering the complexity in measuring innovation performance, which is still not wholly comprehensive (Sabidussi, et al., 2014; Coombs et al., 1996) and challenging (Hauser, 1998; Hagedoorn & Cloodt, 2003), the MIP database would provide the comprehensive picture in measuring it. The MIP database that is based on the CIS measurement refined and expanded the innovation performance measurement by using the proportion of total turnover from new or clearly improved products to differentiate specifically the result of the innovation process regarding the financial measure; this is consistent with prior research that used CIS (Laursen & Salter, 2006; Leiponen & Helfat, 2010; Krzeminska & Eckert, 2016) A 9-point numerical scale represents the data in this variable from 0 to 8, in which each scale has its percentage range ([0] 0%, [1] 0 to < 5%, [2] 5 to < 10%, [3] 10 to < 15%, [4] 15 to < 20%, [5] 20 to < 30%, [6] 30 to <50%, [7] 50 to <75%, and [8] 75 to 100%). However, in calculating this dependent variable, there was an appropriation conducted with adjustment of each range to its midpoint to increase the accuracy of the testing process. This variable is the dependent variable in this study.
4.2.3. Big data analytics usage. The role of big data analytics in innovation is the main subject of this research, and it plays a significant role in R&D process, by linking data from internal and external sides of the firm to observe the information processing capacity of firms. The big data analytics usage measure is also provided in MIP database. A 4-point Likert scale ranging from 0 to 3 is assigned to each firm, where 0 represented no big data analytics use in the firm, 1 represented low usage, 2 represented medium usage, and 3 represented high usage of big data analytics. This variable is the moderating variable in this study.
16
intensity, which is calculated as turnover from abroad divided by total turnover, to see the impact of international linkage to global market to innovation performance (Fosfuri & Tribo, 2008). This measure ranged from 0 to 0.85, which 0.85 is a truncated number for the same reason of anonymization with the previous control variable. The next control variable is firm size, which is the natural logarithm of a firm’s number of employees. This may influence a firm’s human resource availability for innovation process (Laursen & Salter, 2006). Finally, the proportion of employee who has a higher educational degree (education level) which is represented in interval scale from 0 to 8 in percentage range ([0] 0%, [1] 0 to < 5%, [2] 5 to < 10%, [3] 10 to < 15%, [4] 15 to < 20%, [5] 20 to < 30%, [6] 30 to <50%, [7] 50 to <75%, and [8] 75 to 100%). Table 1 presents the descriptive statistics and correlations for the variables.
Table 1: Descriptive Statistics and Correlations
Mean SD (1) (2) (3) (4) (5) (6) (7) (1) Innovation performance 0.085 0.181 1 (2) R&D intensity 0.009 0.029 0.480*** 1 (3) Big data analytics Usage 0.486 0.775 0.234*** 0.139*** 1 (4) East/West 0.333 0.471 0.014 0.042** -0.032* 1 (5) Labor productivity 0.266 0.169 0.046** -0.038* 0.109*** -0.176*** 1 (6) Export intensity 0.142 0.245 0.252*** 0.258*** 0.101*** -0.125*** 0.289*** 1 (7) Size 3.345 1.478 0.153*** 0.057*** 0.234*** -0.099*** 0.254*** 0.287*** 1 (8) Education level 0.227 0.260 0.194*** 0.286*** 0.159*** 0.102*** -0.007 0.053* -0.121***
Note: *p < 0.05; **p < 0.01; ***p < 0.001. The correlations in bold are used as the proxy of
17 5. RESULTS 5.1. Hypotheses Testing
All measures in this research have a single measurement item. Hence it is not necessary to estimate a separate measurement model in structural equation modeling (SEM). This research uses ordinary least square (OLS) regression to test the hypotheses, which is substantially equivalent to estimating a structural model with a single measurement in SEM. The hypothesis testing of the inverted U-shaped relationship between R&D intensity and innovation performance was done following a quadratic regression model described in Aiken and West (1991:63) and Haans et al. (2016:1181). Table 2 details the regression results. The first step in the regression process is the estimation of a control model which showed that all variable had a statistically significant effect on product/service innovation performance. The firm size variable that is represented by employee number has a positive relationship with innovation performance because large firms usually have more resources put into innovation process (Chiang & Hung, 2010). However, the labor productivity as a control variable in the first regression is negatively associated with innovation performance.
Table 2: OLS Regression Results
Control Variable (1) Product/ Service
Innovation Performance (2a) Product/ Service Innovation Performance (2b) Product/ Service Innovation Performance Moderated by Big Data (3) R&D intensity 2.266*** (0.103) 5.403*** (0.400) 5.806*** (0.538) R&D intensity2 -22.590*** (2.782) -26.375*** (3.780)
Big data usage 0.026***
(0.004) Big data usage X
R&D intensity
-0.950* (0.444) Big data usage X
18 (0.002) (0.002) (0.002) (0.002) Education level 0.107* (0.014) 0.040** (0.013) 0.036** (0.013) 0.028* (0.014) R2 0.1757 0.3080 0.3237 0.3406 Adj. R2 0.1687 0.3017 0.3173 0.3331 F 25.200 48.720 50.430 45.540 n 2,983 2,873 2,873 2,676
Note. *p < 0.05; **p < 0.01; ***p < 0.001. Standard errors are in parentheses. Dependent variable is percentage of improved products/services in total turnover
Following the test of control variables, R&D intensity and squared R&D intensity variables were consecutively added to the model. The regression result also strongly supports the hypothesis 1, which asserts that R&D intensity is curvilinear – taking an inverted U shape – relation to innovation performance. First, the R&D intensity variable has a significant and positive relationship with innovation performance, and this is consistent with prior research by Laursen & Salter (2006) and Ravichadran et al. (2017) which tested the effect of R&D to innovation performance. Second, the squared R&D intensity is also significant, showing that when firms spend too much on R&D investment regarding their turnover, there are diminishing returns. Figure 1 shows, in the case of innovation performance, the point where the level of R&D intensity appears to have negative impact on innovation performance. The parabola reaches its maximum when the R&D intensity is approximately at 0.12. Therefore, if the firm’s R&D intensity value is more than 0.12, the negative returns set in. Nevertheless, although the model predicts negative return, the conclusion on this part of the result is there is diminishing return from a negative and significantly squared term. In this hypothesis 1 testing, the control variables are still statistically significant.
Figure 1. Predicted relationship between R&D intensity and innovation performance
19
20
Figure 2. Predicted relationship between R&D intensity and innovation performance with various levels of big data analytics as the moderator
5.2. Robustness Check
The key concern regarding cross-sectional survey data is common method bias, because the same survey respondent answered all questions. The method that is used to test common method bias is marker variable approach, which has been recommended as one of the most effective methods to assess common method bias in IS research (Malhotra et al., 2006). The step to conduct the marker variable approach according to Lindell and Whitney (2001) is by using the smallest correlation (i.e., 0.014) and, more conservatively, the second smallest correlation (i.e., 0.042) as the proxies for common method variance (CMV). It is found that after partialling out CMV from the zero-order correlations between innovation performance and several variables, partial correlations remain statistically significant after an increase of up to 30% and 89% (see Table 5 on the next page)
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 0.11 0.12 0.13 0.14 0.15 0.16 0.17 0.18 0.19 0.2 0.21 0.22 In n o vat ion P erf o rm an ce R&D Intensity
21
Table 5: Assessment of Common Method Bias
Antecedents of innovation performance
Zero-order correlation
First smallest correlation as the proxy of CMV (0.014)
22
6. DISCUSSION
6.1. Main Findings
The emergence of big data analytics and its utilization in the business environment has attracted the managers on how they can take benefits of big data analytics for firms’ business process, particularly R&D process. R&D is the main process in a firm to generate new ideas, technologies, and even products that would contribute to the firm financial performance. Because of that reason, most firms will invest more in R&D resources, however, the decline in R&D productivity from the overinvestment in R&D has been an obstacle in improving innovation output for firms. The aim in this study is to investigate whether big data analytics usage can mitigate the diminishing returns to R&D investments, and how the moderating effect intervene in the R&D investment relationship with innovation performance.
Utilizing MIP database from 2016, with responses from 5,294 firms from various industry sectors in Germany, this study provides several new insights. First, there is a curvilinear relationship between R&D investment and innovation performance, which there is a turning point where can be considered as an optimal point to invest in R&D resources. When a firm invest in R&D beyond that turning point, empirically and theoretically that firm conducts an overinvesting in R&D and bring down the innovation level lower in each additional R&D investment. This finding is consistent with prior studies that showed the diminishing return of R&D (Ravichandran et al., 2017; Lin & Chen, 2015; Graves & Langowitz, 1993; Faff et al., 2013). Second, the analysis as expected in hypothesis 2 shows that big data analytics usage in a firm moderates the inverted-U shaped relationship between R&D investment and innovation performance. The usage of big data analytics can mitigate the diminishing return to R&D, in other word that the marginal returns to R&D decrease at a slower rate for firms using higher level of big data analytics. This finding emphasize that big data analytics that will boost the combined value of existing R&D investment and the big data analytics usage, bring super additive value, and subsequently also boost the firm’s innovation performance.
6.2. Contributions and Implications
23
from the interaction between those resources in innovation process. The super-additive value would enhance the the firm’s innovation performance and also flattening the relationship curve between R&D investment and innovation performance. This can be explained in the This theoretical description contributes to the understanding of how big data analytics and existing R&D resources creates value, which is accompanied with other resources.
The findings from this study also contribute to the research on R&D management, specifically on the relationship of R&D investment and the innovation performance. While previous studies stated that R&D investment has its diminishing returns characteristics to various dependent variable (Graves & Langowitz, 1993; Huang & Liu, 2005; Faff et al., 2013), these studies more emphasizing on operational aspects of the firm. This research gives new insight on how firm investment in R&D can be more optimized to prevent the overinvestment due to the declining level of productivity. On the other hand, there have been limited attention paid to the mechanism that can intensify R&D productivity. Insufficient capability and the presence of rigidity have been the main resaons of diminishing return of R&D investment (Ravichandran et al., 2017). The previous research of Ravichandran et al. (2017) brought about the benefit of IT in managing R&D, however it did not address specific technological artefact or systems that are possible to mitigate the diminishing return of R&D investment. This study complements the previous research and shows that big data analytics is one of the specific technology that mitigates the diminishing return of R&D, and the higher level of big data analytics usage by a firm, the flatter the curve of R&D investment in regard to innovation performance.
24
product satisfcation would reducing the time and cost taken for doing customer research for the innovation purpose. The fast-changing business environment triggers the business itself to change and adapting to the technology that is more appropriate to the business environment. This study would suggest managers that the large investment in big data analytics, will boost the innovation performance and adapting the change of how R&D process works while reducing the risk of overinvestment in R&D at the same time
6.3. Limitations and Future Studies
This study has some limitations. First, the innovation performance is measured only by firm’s proportion of total turnover from new or clearly improved products. The proportion of total turnover from new or clearly improved products is appropriate for this study because managers’ need to examine whether the effect of investment expenditures that contribute to the new or clearly improved products/services brings financial profitability to firms. However, the proportion does not fully capture the knowledge and human capital factor to the innovation process. There is abundant room for further progress to include the alternative measures for innovation performance, such as patent count (e.g. Ravichandran et al., 2017) to test this study’s findings.
Second, the MIP dataset, which was collected from a large-scale sample firms from various industries in a recent year, has a cross-sectional design. This cross-sectional dataset only assumed firm behavior innovation process at one point of time, neglecting the possibility of change of innovation strategy and inclusion of new technology. The anonymization of the MIP datasets is also make the continuous study regarding this topic less possible, due to the unavailability of firm identifier. Hence, the panel dataset construction is not possible from MIP dataset. The causality that underlies the hypothesized relationships cannot be examined comprehensively because of the generalization of how big data analytics would moderate the relationship between R&D investment and innovation performance. The MIP dataset also did not define the exact measure of the usage level of digital technology (i.e. big data analytics) by the firm. There are possibilities of the improvement of the dataset to include specific measure of digital technology level usage, such as the investment value of that each of the specific technology. There are more possibilities on how big data analytics usage technically benefits the innovation process that could be investigated on further studies.
25
26
7. CONCLUSION
27
REFERENCES
Abbasi, A., Sarker, S., & Chiang, R. H. (2016). Big data research in information systems: Toward an inclusive research agenda. Journal of the Association for Information Systems, 17(2), i-xxxii.
Ackoff, R. L. (1971). Toward a system of systems concepts. Management Science, 17(11), 661-671.
Aiken, L., & West, S. (1991). Multiple Regression: Testing and Interpreting Interactions. London: Sage.
Aral, S., Brynjolfsson, E., & Wu, L. (2012). Three-way complementarities: Performance pay, human resource analytics, and information technology. Management Science, 58(5), 913-931.
Asthana, A., & Chari, S. (2015, April). Cost-Benefit Analysis: Comparing the IBM PureData System with Hadoop Implementations for Structured Analytics. Retrieved from Cabot Partners Group, Inc. : http://www.cabotpartners.com/Downloads/TCO-Study-Pure-Data-versus-Hadoop-May-2015.pdf
Ayadi, O., Dufrene, U., & Obi, C. (1996). Firm Performance Measures: Temporal Roadblocks to Innovation? Managerial Finance, 22(8), 18-32.
Baesens, B., Bapna, R., Marsden, J. R., Vanthienen, J., & Zhao, J. L. (2014).
Transformational issues of big data and analytics in networked busines. MIS Quarterly, 38(2), 629-631.
Becker, H., & Speltz, L. (1983). Putting the S-curve concept to work. Research Management, 26(5), 31-33.
Behrens, V., Berger, M., Hud, M., Hunermund, P., Iferd, Y., Peters, B., . . . Schubert, T. (2017). Innovation Activities of Firms in Germany – Results of the German CIS 2012 and 2014: Background Report on the Surveys of the Mannheim Innovation Panel Conducted in the Years 2013 to 2016. Mannheim: Zentrum für Europäische Wirtschaftsforschung.
Belderbos, R., Carree, M., & Lokshin, B. (2004). Cooperative R&D and firm performance. Research Policy, 33(10), 1477-1492.
Bharadwaj, A. (2000). A resource-based perspective on information capability and firm performance: An empirical investigation. MIS Quarterly, 24(1), 169-196.
Blackburn, M., Alexander, J., Legan, J. D., & Klabjan, D. (2017). Big Data and the Future of R&D Management. Research-Technology Management, 60(5), 43-51.
Branch, B. (1974). Research and Development Activity and Profitability: A Distributed Lag Analysis. Journal of Political Economy, 82(5), 999-1011.
28
Bumblauskas, D., Nold, H., Bumblauskas, P., & Igou, A. (2017). Big data analytics:
transforming data to action. Business Process Management Journal, 23(3), 703-720. Chao-Ton, S., Yung-Hsin, C., & Sha, D. Y. (2006). Linking innovative product development
with customer knowledge: A data-mining approach. Technovation, 26(7), 784-795. Checkland, P. (1999). Systems Thinking, Systems Practice. Chichester: John Wiley & Sons. Chen, H., Chiang, R. H., & Storey, V. C. (2012). Business intelligence and analytics: from
big data to big impact. MIS Quarterly, 36(4), 1165-1188.
Chiang, Y.-H., & Hung, K.-P. (2010). Exploring open search strategies and perceived innovation performance from the perspective of inter‐organizational knowledge flows. R&D Management, 40(3), 292-299.
Churchman, C. W. (1971). The Design of Inquiring Systems: Basic Concept of Systems and Organizations. New York, NY: Basic Books.
Cohen, W. M., & Levinthal, D. A. (1990). Absorptive Capacity: A New Perspective on Learning and Innovation. Administrative Science Quarterly, 35(1), 128-152. Coombs, R., Narandren, P., & Richards, A. (1996). A literature-based innovation output
indicator. Research Policy(25), 403-413.
Davenport, T. H., & Harris, J. G. (2007). Competing on Analytics: The New Science of Winning. Boston: Harvard Business Press.
Davenport, T. H., Barth, P., & Bean, R. (2012). How 'Big Data' is Different. MIT Sloan Management Review, 54(1), 22-24.
Dedrick, J., Gurbaxani, V., & Kraemer, K. L. (2003). Information technology and economic performance: A critical review of empirical evidence. ACM Computing Surveys, 35(1), 1-28.
Dong, J. Q., & Netten, J. (2017). Information technology and external search in the open innovation age: New findings from Germany. Technological Forecasting and Social Change, 120, 223-231.
Dremel, C., Wulf, J., Herterich, M. M., Waizmann, J., & Brenner, W. (2017). How AUDI AG Established Big Data Analytics in Its Digital Transformation. MIS Quarterly Executive, 16(2), 81-100.
El Sawy, O., Kraemmergaard, P., Amsinck, H., & Vinther, A. L. (2016). How Lego built the foundations and enterprise capabilities for digital leadership. MIS Quarterly
Executive, 15(2), 141-166.
Erickson, G., & Jacobson, R. (1992). Gaining Comparative Advantage through Discretionary Expenditures: The Returns to R&D and Advertising. Management Science, 38(9), 1264-1279.
29
Farjoun, M. (1998). The independent and joint effect of the skill and physical bases of relatedness in diversification. Strategic Management Journal, 19(7), 611-630.
Fosfuri, A., & Tribo, J. A. (2008). Exploring the antecedents of potential absorptive capacity and its impact on innovation performance. Omega, 36(2), 173-187.
Fosso Wamba, S., Gunasekaran, A., Akter, S., Ren, S. J.-f., Dubey, R., & Childe, S. J. (2017). Big data analytics and firm performance: Effects of dynamic capabilities. Journal of Business Research, 70, 356-365.
Foster, R. (1986). Innovation: The Attacker's Advantage. New York, New York: Summit Books.
Gharajedaghi, J. (2006). Systems Thinking: Managing Chaos and Complexity: A Platform for Designing Business Architecture. Burlington, MA: Butterworth-Heinemann.
Gou, B., Wang, Q., & Shou, Y. (2004). Firm Size, R&D, and Performance: An Empirical Analysis on Software Industry in China. International Engineering Management Conference. Singapore: IEEE.
Guest, D. E., Michie, J., Conway, N., & Sheehan, M. (2003). Human Resource Management and Corporate Performance in the UK. British Journal of Industrial Relations, 41(2), 291-314.
Günther, W. A., Mehrizi, M. H., Huysman, M., & Feldberg, F. (2017). Debating big data: A literature review on realizing value from big data. The Journal of Strategic
Information Systems, 26(3), 191-209.
Haans, R. F., Pieters, C., & He, Z.-L. (2016). Thinking about U: theorizing and testing U- and inverted U-shaped relationships in strategy research. Strategic Management Journal, 37(7), 1177-1195.
Hagedoorn, J., & Cloodt, M. (2003). Measuring innovative performance: is there an advantage in using multiple indicators? Research Policy, 32(8), 1365-1379. Hauser, J. (1998). Research, development and engineering metrics. Management Science,
44(12), 1670-1689.
Ho, J., Tian, F., Wu, A., & Xu, S. (2017). Seeking value through deviation? Economic impacts of IT overinvestment and underinvestment. Information Systems Research, 28(4), 850-862.
Huang, C. J., & Liu, C. J. (2005). Exploration for the relationship between innovation, IT and performance. Journal of Intellectual Capital, 6(2), 237-252.
Huang, C. J., & Liu, C. J. (2005). Exploration for the relationship between innovation, IT, and performance. Journal of Intellectual Capital, 6(2), 237-252.
30
Kafouros, M. I., Buckley, P. J., Sharp, J. A., & Wang, C. (2008). The role of
internationalization in explaining innovation performance. Technovation, 28(1-2), 63-74.
Klingebiel, R., & Adner, R. (2015). Real options logic revisited: The performance effects of alternative resource allocation regimes. Academy of Management Journal, 58(1), 221-241.
Knott, A. M. (2012). The Trillion-Dollar R&D Fix. Harvard Business Review, 90(5), 77-82. Kostopoulos, K., Papalexandris, A., Papachroni, M., & Ioannou, G. (2011). Absorptive
capacity, innovation, and financial performance. Journal of Business Research, 64(12), 1335-1343.
Krzeminska, A., & Eckert, C. (2016). Complementarity of internal and external R&D: is there a difference between product versus process innovations? R&D Management, 46(S3), 931-944.
Lau, R. Y., Zhao, J. L., Chen, G., & Guo, X. (2016). Big data commerce. Information and Management, 53(8), 929-933.
Laursen, K., & Salter, A. (2006). Open for innovation: the role of openness in explaining innovation performance among U.K. manufacturing firms. Strategic Management Journal, 43(5), 867-878.
Laursen, K., & Salter, A. (2006). Open for innovation: The role of openness in explaining innovation performance among U.K. manufacturing firms. Strategic Management Journal, 27, 131-150.
Leiponen, A., & Helfat, C. E. (2010). Innovation objectives, knowledge sources, and the benefits of breadth. Strategic Management Journal, 31, 224-236.
Leiponen, A., & Helfat, C. E. (2011). Location, decentralization, and knowledge sources for innovation. Organization Science, 22(3), 641-658.
Lichtenthaler, U. (2016). Absorptive capacity and firm performance: an integrative
framework of benefits and downsides. Technology Analysis & Strategic Management, 28(6), 664-676.
Lin, B., & Chen, J. (2005). Corporate Technology Portfolios and R&D Performance Measures: A Study of Technology Intensive Firms. R&D Management, 35(2), 157-170.
Lin, B.-W., Lee, Y., & Hung, S.-C. (2006). R&D intensity and commercialization orientation effects on financial performance. Journal of Business Research, 59, 679-685.
Lin, C., Wu, Y.-J., Chang, C., Wang, W., & Lee, C.-Y. (2012). The alliance innovation performance of R&D alliances—the absorptive capacity perspective. Technovation, 32, 282-292.
31
Conference on Information Systems 2011 (pp. 1-15). Brisbane: Queensland University of Technology.
Long, W., & Ravenscraft, D. (1993). Lbos, Debt and R&D Intensity. Strategic Management Journal, 14, 119-135.
Malhotra, N. K., Kim, S. S., & Patil, A. (2006). Common Method Variance in IS Research: A Comparison of Alternative Approaches and a Reanalysis of Past Research.
Management Science, 52(12), 1865-1883.
Mauro, A. D., Greco, M., & Grimaldi, M. (2016). A formal definition of Big Data based on its essential features. Library Review, 65(3), 122-135.
McAfee, A., & Brynjolfsson, E. (2012). Big data: the management revolution. Harvard Business Review, 90(10), 60-68.
Mithas, S., Krishnan, M., & Fornell, C. (2016). Information technology, customer
satisfaction, and profit: Theory and evidence. Information Systems Research, 27(1), 166-181.
Mithas, S., Tafti, A., Bardhan, I., & Goh, J. (2012). Information technology and firm profitability: Mechanisms and empirical evidence. MIS Quarterly, 36(1), 205-224. Müller, O., Fay, M., & Brocke, J. v. (2018). The Effect of Big Data and Analytics on Firm
Performance: An Econometric Analysis Considering Industry Characteristics. Journal of Management Information Systems, 35(2), 488-509.
Nambisan, S., Lyytinen, K., Majchrzak, A., & Song, M. (2017). Digital innovation
management: Reinventing innovation management research in a digital world. MIS Quarterly, 41(1), 223-238.
Penrose, E., & Pitelis, C. (2009). The Theory of the Growth of the Firm. Oxford: Oxford University Press.
Ravichandran, T., & Lertwongsatien, C. (2005). Effect of information systems resources and capabilities on firm performance: A resource-based perspective. Journal of
Management Information Systems, 21(4), 237-276.
Ravichandran, T., Han, S., & Mithas, S. (2017). Mitigating Diminishing Returns to R&D: The Role of Information Technology in Innovation. Information Systems Research, 28(4), 812-827.
Ritala, P., & Hurmelinna-Laukkanen, P. (2013). Incremental and Radical Innovation in Coopetition—The Role of Absorptive Capacity and Appropriability. Journal of Product Innovation Management, 30(1), 154-169.
Robert, D. L., & Candi, M. (2014). Leveraging social network sites in new product development: Opportunity or hype? Journal of Product Innovation Management, 31(S1), 105-117.
32
Russom, P. (2011). TDWI Best Practices Reports: Big Data Analytics. Renton: The Data Warehousing Institute.
Sabidussi, A., Lokshin, B., Leeuw, T. d., Duysters, G., Bremmers, H., & Omta, O. (2014). A comparative perspective on external technology sourcing modalities: The role of synergies. Journal of Engineering and Technology Management(33), 18-31. Shy, O. (1995). Industrial Organization: Theory and Applications. Cambridge,
Massachusetts: The MIT Press.
Stock, G. N., Greis, N. P., & Fischer, W. A. (2001). Absorptive capacity and new product development. Journal of High Technology Management Research, 12, 77-91.
Tambe, P., & Hitt, L. (2012). The productivity of information technology investments: New evidence from IT labor data. Information Systems Research, 23(3), 599-617.
Tanriverdi, H. (2005). Information technology relatedness, knowledge management capability, and performance of multibusiness firms. MIS Quaterly, 29(2), 311-334. Tanriverdi, H. (2006). Performance effects of information technology synergies in
multibusiness firms. MIS Quarterly, 30(1), 57-77.
Tanriverdi, H., & Venkatraman, N. (2005). Knowledge relatedness and performance of multibusiness firms. Strategic Management Journal, 26(2), 97-119.
Tsai, W. (2001). Knowledge transfer in intraorganizational networks: Effects of network position and absorptive capacity on business unit innovation and performance. Academy of Management Journal, 44(5), 996-1004.
Van de Vrande, V., Vanhaverbeke, W., & Duysters, G. (2011). Technology in-sourcing and the creation of pioneering technologies. Journal of Product Innovation Management, 28(6), 974-987.
Wagner, H. (2004). Internationalization speed and cost efficiency: evidence from Germany. International Business Review, 13(4), 447-463.
Youndt, M. A., Subramaniam, M., & Snell, S. A. (2004). Intellectual Capital Profiles: An Examination of Investments and Returns. Journal of Management Studies, 41(2), 335-361.
33 APPENDICES
Industry Classification
Industry Classification Frequency Percentage
1) Mining 197 4.20 2) Food/Tobacco 223 4.76 3) Textiles 157 3.35 4) Wood/Paper 151 3.22 5) Chemical 126 2.69 6) Plastics 154 3.29 7) Glass/Ceramics 101 2.16 8) Metals 332 7.09 9) Electrical equipment 292 6.23 10) Machinery 194 4.14 11) Retail/Automobile 102 2.18
12) Furniture/Toys/Medical Technology and Maintenance 273 5.83
13) Energy/Water 311 6.64
14) Wholesale 208 4.44
15) Transport equipment/Postal Service 409 8.73
16) Media services 246 5.25
17) IT/Telecommunications 187 3.99
18) Banking/Insurance 186 3.97
19) Technical service/R&D Services 310 6.62
20) Consulting/Advertisement 283 6.04
21) Firm-related services 243 5.19
