Digital Leadership and Firm Performance: a Meta-Analysis

Digital Leadership and Firm Performance: a

Meta-Analysis

Master’s Thesis

MSc BA - Strategic Innovation Management/SIM

Economics and Business, University of Groningen

Supervisor: dr. Q. (John) Dong

Co-Assessor:

P. (Pere) Arque-Castells, PhD

1. Abstract

Digitalization is one of the most important innovations for firms in recent years. Managing the digital transformation has therefore become a key strategic issue. More and more firms try to manage this digital transformation by appointing digital leaders at multiple levels within their organization. Such digital leaders are managers in a digital setting who can be present within organizations at different levels: at top levels as a CEO, at functional levels as the CIO, or at project levels as team or project managers. The idea behind appointing digital leaders is that leaders with a digital background will focus on digital initiatives and increase the speed of the digital transformation. By enhancing the level of digital

2. Introduction

Digitalization is one of the ongoing megatrends that widely affect organizations in the last decades. As it is expected that the impact of digitalization on organizations will continue to do so at an even greater pace, the prediction is that how organizations deal with such digital changes in their operations determine its future competitiveness. Within firms, digital technologies are disrupting value chains, organizational structures, processes of operations, and revenue models. Basically, businesses as a whole are affected in every industry and in every region. Managing this digital transformation will decide which company will survive in the future and which will not (Boyd et al., 2015). One way for

organizations to be able to adapt to such a changing environment is through having managers that are aware of the latest business developments (Bankewitz et al., 2016). Many organizations already prepare themselves for a move towards a more digital future by focusing on appointing digital leaders. In 2015, only 6 percent of 2,500 of the world’s largest public companies appointed an executive to lead their digital processes. One year later this same number grew to 19 percent already. Moreover, sixty percent of the digital leaders that were identified in studies by Strategy& were appointed since 2015 (Acker et al., 2017). Such numbers indicate the level of priority for firms to appoint digital leaders within their organization in recent years. Digital transformation processes are influenced by managers, and

organizations have started to appoint digital leaders as they expect that digital leadership will positively impact the digital transformation of organizations.

But what does digital leadership exactly mean? In their paper, Ding et al. (2014) provide four conceptualizations of IS strategic leadership: one that stated it as leadership that focuses on CIO job responsibilities, one that discussed it as leadership focusing on technology and business, one that focused on leadership through effectiveness, and the final one that explained it as leadership through influence on firm performance at the level of top executives (Ding et al., 2014). Moreover, these four

meant, spread out on different levels: at top levels as a CEO, at functional levels as the CIO, or at project levels as team or project managers.

The reasoning behind this mechanism of managers and their influence on organizational outcomes can be found in the upper echelons theory, which explains that organizational outcomes, like strategic choices and performance levels, can be predicted by managers’ characteristics (Hambrick & Mason, 1984). Recent studies have found support for the importance of leadership in other domains like the effect of sustainability related expertise on environmental performance (HomRoy & Slechten, 2016), merger and acquisition expertise on improved decision making (Field & Mkrtchyan, 2017), and industry expertise on successful internationalization strategies and an enhanced strategic change (Oehmichen et al., 2017), amongst others. However, the question whether digital leadership within a firm is actually creating value remains unanswered. As digitalization becomes more and more important for organizations and an increasing number of organizations is reacting by appointing digital leaders, it is important to understand whether digital leadership leads to improved organizational performance. At this point in time,

digitalization is considered as one of the most important trends for organizations with a huge impact on business. Combined with the observation that a large number of organizations are responding by appointing digital leaders and since they are expecting to benefit from having such leaders on board, it becomes interesting to understand the value creating mechanism behind this idea. However, there remains a gap in research concerning insights and empirical support for this mechanism of appointing digital leaders and its effect on improved firm performance. Digital leadership is a topic that is related to multiple levels of leadership within a firm. As these different levels occur, a diverse range of literature exists related to the topic. This study is important to guide further research and to check what has already been found related to this topic in current literature by synthesizing these diverse findings.

As a response to this lack of research related to leadership in the digital domain and because it is unknown how digital leaders affect business in practice, this study conducts a meta-analysis of previous studies that include a variable related to digital leadership, while also reporting a variable related to organizational performance. By doing so, previous literature is catalogued, variables that affect the findings are identified, and implications for managers as well as suggestions for future research directions are discussed. The objective is to find a pattern for strong leadership resulting in an improved

theory holds in a digital setting as well. This topic can be linked to my studies in Strategic Innovation Management for the following reason. Digitalization is a key strategic innovation in recent years and going through the digital transformation is one of the most important strategic issues for firms. Managing such a digital transformation by appointing digital leaders is a strategic and innovation related topic, and thus related to the field of Strategic Innovation Management. Therefore, this topic fits well with my studies.

3. Literature Background

3.1 Digitalization

The impact of digitalization on the business world can be widely observed. Within organizations, several operational activities including firms’ business models have been influenced by the increased level of digital activities. Different types of cooperation, relationships, and collaboration between

companies, between companies and customers, and between customers and employees have been enabled by the digitalization process resulting in new products and services (Rachinger et al., 2019). A survey by the European Central Bank shows how important the adoption of digitalization already has become for large organizations. Companies across all sectors are already making use of digital activities like big data, e-commerce and cloud computing, while most sectors also have adopted artificial intelligence, the

internet of things, and robotics. Most important, organizations believe to be able to benefit financially from the digitalization of business operations. Firms point out that the better access to consumers it creates is regarded as the most important channel of digitalization affecting their sales (ECB, 2018). Such digitalization practices and digital transformations impact businesses in all sectors and amongst all firms, from shipping (Lambrou et al., 2019), chemical engineering (Isaksson et al., 2018), to manufacturing (Zhou, 2013), and from start ups (Isaksson et al., 2018; Neubert, 2018), small- to medium sized

enterprises (Bouwman et al., 2018), to the larger businesses in this world (Das et al., 2018). Most of these companies expect to benefit from the impact of digital technologies as digitalization is considered as being a creator of value for companies. More than half of the companies expect it to give rise to a slight increase in sales in the next three years while one third expect a significant increase (ECB, 2018). Such insights show how important digitalization already has become for organizations, how organizations believe to benefit from it, and how this is expected to become even more important in the future.

Therefore, a need exists for organizations to organize themselves differently to be capable of transforming to their changing environment in terms of digitalization.

3.2 Leadership

Leadership is in many theories discussed as an important factor within a firm. The first theory focusing on leaders as a crucial mechanism in creating firm outcomes is the upper echelons theory. Within this theory, the firm is considered as a reflection of its leaders: the organizational outcomes, the choices made strategically, and organizational performance levels can be predicted by the backgrounds of its managers (Hambrick & Mason, 1984). The main mechanism of this theory is that the leaders’

(Hambrick, 2007). As aforementioned, leaders with a background in sustainability create a better environmental performance within their organizations (HomRoy & Slechten, 2016), with expertise in merger and acquisitions enhance the decision making performance (Field & Mkrtchyan, 2017), and high levels of industry experience leads to improved internationalization strategies and enhanced strategic change (Oehmichen et al., 2017). Combining the general mechanism of the upper echelons theory with such support in different industries and amongst different concepts, it becomes clear that this mechanism might hold in the case of digital leadership as well. Within the domain of digital leadership, the upper echelons theory would expect that appointing leaders with a digital background in the firm will lead to a higher level of strategic choices that can be linked to digitalization. Thus, the expectation will be that a firm with digital leaders will start more digital initiatives within the firm compared to organizations with a lower number of digital leaders. Such digital leaders will be more likely to see the opportunity of digital initiatives than leaders without a digital background and will therefore initiate more digital initiatives. This idea can be linked to the finding that managers are influencing the implementation of new technology, like digital initiatives within the firm, and that digital leaders are more likely to encourage subordinates to adopt digital technologies (Leonard-Barton & Deschamps, 1988). Therefore, this study expects that digital leaders will focus more on digital problem solving by initiating digital activities within the firm compared to leaders that have a lower level of digital background.

3.3 Digital Leadership and Firm Performance

As aforementioned, digital leadership can be explained as leadership in a digital context. This includes managing firms in a digital setting or by leaders with a personal digital background. Such digital leadership exists at multiple levels, at CEO, CIO and project or team level. But how does an increase in digital leaders, who are more likely to initiate digital activities, lead to improved organizational

performance? The main idea behind this mechanism is related to the upper echelons theory, which implies that such digital leaders are more likely to make decisions based on their digital expertise (Hambrick & Mason, 1984). This theory would imply that when the leaders’ characteristics are digital or related to IS, these leaders will have a so-called IS vision or digital vision, as this is related to their own perceptions, values, and experiences. Thus, the upper echelon theory suggests that a digital leader will make choices based on his/her digital vision and will theoretically improve IS quality (Ding et al., 2014; Hambrick & Mason, 1984). As a result, digital leadership would lead to an increase of digital initiatives, and an increase in quality of such initiatives within a firm.

To make a link from here to benefits for firm performance, it is important to understand the different phases of digital transformation. In business perspectives, two crucial phases of digital

explained as how organizations use IT or digital technologies to change business processes while digital transformation goes a step further and can be described as a companywide change triggered by IT or digital technologies that leads to new business models for the organization. In digitalization, IT is seen as an enabler for seizing opportunities by altering existing business processes, and thus leading to value for the organization. In the concept of digital transformation, it is explained that digital technologies help to create a competitive advantage for the organization by transforming the organization and enabling the leverage of core competences or developing new core competences (Verhoef et al., 2019). While understanding these phases of digital transformation, it becomes clear that leaders can influence this transformation. Leaders that understand digital processes better will be more likely to make use of digital technologies within the firm. A theory explaining how digital technologies like IT can lead to improved organizational performance is the information-processing perspective. This perspective views

organizations as information processing systems facing uncertainty (Nonaka et al., 1996). In order to operate effectively as an organization, there must be a fit between the information processing

requirements facing the organization, and the information processing capacities of organizational design (Egelhoff, 1991). Thus, digital initiatives like the use of big data for business purposes and other digital technologies that enhance communication and collaboration increase the level of information processing capacity of the firm. On the other hand, environmental uncertainty will increase the information

processing requirements of the organization. Therefore, the information processing capacity can be seen as how well an organization is able to handle the uncertainty created by its environment (Egelhoff, 1991).

In this perspective, IT and other digital technologies facilitate a firm to become a ‘knowledge-creating company’ (Nonaka et al., 1996), leading to an increased ability to handle uncertainty and improved decision making which in turn will lead to a more effective and efficient organization (Egelhoff, 1991). Additionally, digital initiatives also stimulate globalization and simplify business development practices, which enhance the commercial activities of organizations. As the upper echelon theory explains that leaders with digital expertise predict to stimulate IT initiatives (Hambrick & Mason, 1984), it can be argued that a leader with a digital background leads to improved information processing capacity which, in turn, leads to improved organizational performance. To summarize, digital leadership is expected to increase the number and quality of digital initiatives within a firm, which leads to

4. Methodology

4.1 Meta-Analysis

A analysis is used as the statistical technique in this research. By conducting a meta-analysis, it is possible to identify what currently available research says about a certain mechanism (Bangert-Drowns, 1992). Meta-analyses use statistical methods that pool and integrate results of primary empirically conducted studies (Stewart & Roth, 2001). A meta-analysis is chosen as the research method for this study to identify whether leadership in the digital domain and its effect on organizational

performance has been studied before and to find out what current literature says about this proposed mechanism. While previous research has not focused on the specific research question whether digital leadership leads to improved firm performance, many studies do include measurements for leadership and firm performance, while also focusing on a digital setting. As the meta-analysis technique provides the opportunity to pool results from existing research in a certain research domain, sample sizes can be increased without the need to collect primary data (Bangert-Drowns, 1992). By making use of a meta-analysis, it is possible to include the correlations of relationships that are found within such studies that did not focus on the proposed research question but that do include both variables. Therefore,

measurements from these previously conducted studies can be bundled and used in a new statistical model in order to answer the question that is not posed by the individual studies. By doing so, the pattern of the answer to this question by all individual studies is examined by synthesizing the data across the included studies. Choosing for a meta-analysis in this case of finding a pattern between digital leadership and firm performance makes sense as many individual studies include usable measurements for both variables while none of them are answering the question specifically. Therefore, making use of a meta-analysis will help to combine these measurements into one model. This will be statistically stronger and will make use of a larger population than choosing for an analysis in which data for both variables have to be identified primarily. Also, precision is improved as the effect is based on more information than by making use of a single study.

4.2 Sample

Systems) basket of eight that includes the journals European Journal of Information Systems (EJIS), Information Systems Journal (ISJ), Information Systems Research (ISR), Journal of the Association of Information Systems (JAIS), Journal of Information Technology (JIT), Journal of Management Information Systems (JMIS), Journal of Strategic Information Systems (JSIS), and Management Information Systems Quarterly (MISQ) and both the journals Decision Support Systems (DSS) and Information & Management (I&M) are regarded as leading in the field of information systems research. This basket of eight journals is introduced by senior information systems academics to indicate the highest quality journals within the field of information systems (Amiri & Moqri, 2018). Studies published in these journals are considered as the most important articles in the field of information systems.

Therefore, the first sample selection criteria relates to including studies from these ten journals only to make sure that all studies selected for the sample are leading articles within the digital domain.

The second sampling criteria refer to the time span regarding the date of publication of the studies. As this study focuses on the digital domain, and this domain is generally regarded as an emerging field, the time span for the date of publication is chosen to be within the last 20 years; from 2000 - 2019, respectively. This time frame is sufficient as the world has changed radically since the emergence of commercial internet a quarter century ago. Since then, the digital journey has played a major role in every company. As digitalization within businesses took a leap forward in the last decade, choosing for this twenty year time frame will include all relevant studies on this topic (McKinsey, 2019).

Third, relevant keywords have to be selected in order to identify related papers within the selected journals and time frame. As the proposed mechanism to study concerns the effect of leadership in the digital domain on organizational performance, all articles that include the keywords leader or leadership are selected. By using these two keywords, all articles in the digital domain (within Information Systems research) that hold information about leaders or leadership will be included in the sample. By doing so, it is expected that all relevant articles that include a leader or leadership related variable are included in the sample.

After finishing the first three sampling steps, a total of 501 articles across all ten journals were identified. All these articles are within the digital domain, as they are from the ten most important journals in information systems research. Also, all articles are leader or leadership related since they contain the keyword leader or leadership. However, this first sample includes articles that are not

empirical, which are not on firm level, do not hold a variable that measures leadership or do not include a variable that measures organizational performance as well. Therefore, all 501 articles have to be checked for the aforementioned criteria.

performance were deleted from the sample. After removing these articles, a final sample of 26 useful articles remained spread over six different journals.

Finally, I randomly select one correlation between digital leadership and firm performance from 26 articles, as some articles may contain multiple leadership and/or performance variables. The analysis with 26 measurements includes only one measurement per article in the sample and this observation is chosen randomly. By doing so, it is possible to compare this with the results from the sample with all observations, including multiple measurements per article, and check whether they are significantly different. If so, it is from a statistical point of view better to only use the reduced sample of 26

observations for the WLSRs. Therefore, checking and possibly correcting for the repeating measures from the same article is performed as a last step in the sampling procedure.

See Figure 1 for the distribution of the first and the final sample across all journals, including information about the total number of articles that were identified in the first sample and the total number of articles that were left in the final sample after checking all articles for the related criteria.

Figure 1: Distribution of articles per journal used in sample

See Figure 2 for the distribution of the articles in the final sample in terms of the year of publication. As you can see, the publication of useful articles for this meta-analysis took off after 2006 and kept going until recent years. As can be observed in the figure, a temporal trend seems to exist which indicates that the topic was populair from 2006-2012 and after 2014. This implies that the choice to limit the time span of publication of the articles from 2000 to 2019 is a justified choice.

Figure 2: Distribution of articles per year of publication

See table A1 in the appendix for the complete details of the final sample, including the number that was appointed to the article in the analysis and the complete reference of the related article.

4.3 Coding

Per observation retrieved from the articles in the final sample, several pieces of information were reported. First, information for the level of leadership at either top level as the CEO, functional level as the CIO, or project level at project and team level is coded. All variables that measure a leadership aspect on CEO level are coded as CEO. This includes all variables that explain a digital leadership aspect from a CEO, like the CEO’s knowledge of IT. Also, all variables that measure a digital leadership aspect on CIO level are coded as CIO. These variables include leadership information from the CIO perspective, like CIO leadership or CIO human capital. Finally, all variables that measure digital leadership aspects at team or project level are coded as project/team leaders. These variables include leadership aspects at team or project level, like project managers’ practical intelligence. By doing so, leadership aspects at the three different organizational levels are distinguished. See table A2 in the appendix for the details of each measurement regarding the leadership variables.

Second, information related to the type of firm performance is checked and divided between financial performance, project performance, efficiency, and effectiveness. Variables coded as financial performance include all measurements that deal with financial organizational outcomes, like ROA, net benefits, or firm performance for example. Variables coded as project performance include all

measuring project outcomes or project performance. Third, variables coded as efficiency consider all observations that deal with a more efficient organization. Examples of such observations are decision making efficiency, cost performance, or IT contribution to firm efficiency. Finally, variables coded as effectiveness are measurements dealing with a more effective firm, like organizational effectiveness, task outcomes, or decision making effectiveness.

As the mechanism researched in this study focuses on the influence of digital leadership on organizational performance, correlation coefficients of the leadership variable with the organizational performance variable were reported. The choice to make use of Pearson correlations as input for the model is based on the ideas by Hunter and Schmidt (1990), who state that correlations should be encouraged as inputs in meta-analyses as such correlations are not influenced by other variables that are included in the model as well.

Fourth, the industry sectors of the sample in the original study were checked. Five different industry sectors were identified, divided between manufacturing, services, information technology (IT), healthcare, and non profit organizations (NPOs). These industry sectors were based on how the original study coded the industry characteristics of the sample. For example, industries such as the software industry were coded as belonging to the IT sector, industries like shipbuilding were coded as belonging to the manufacturing sector, and industries like governments were coded as NPOs.

Fifth, the countries in the sample of the original studies were coded in three different categories: the United States, emerging countries, and other developed countries. Most countries used in the samples from the original studies were China and the United States, as both countries together were 25 times present in the samples out of the 26 studies. In this study, countries like China, India and Brazil were coded as emerging countries and countries like Canada and France were coded as other developed countries.

Sixth, a differentiation was made between the data collection methods in either the group of survey based data collection or in the group of archival data. In order to code this variable, data collection information of the original studies was checked and was coded based on whether the study made use of survey data or archival data like data from databases.

The next coded variable explains if the observation in the sample consists of either longitudinal or cross sectional data. For this variable, the original sampling methods were checked for whether the data was collected in a single point of time or over a longer time span, and were coded as longitudinal or cross sectional respectively.

modeling (CB SEM). For this variable, the analysis part of the original research was checked in order to find out what type of analytical technique was used.

4.4 Analysis Strategy

Descriptive Statistics

Before testing the proposed mechanism, this paper summarizes the correlation effect of each variable on organizational performance by making use of descriptive statistics. Information like the mean, number of observations (N), standard deviation, minimum- and maximum value of the correlations are all reported. By studying the mean of the descriptive statistics, a first indication of the effect of the related variable on firm performance can be observed. The higher the mean, the higher the general correlation effect of the variable on firm performance is. By reporting the number of observations (N), it can be checked how many times a correlation of the variable was found in relation to firm performance. The higher the number of correlations with firm performance, the more times the variable was correlated with firm performance in the original studies. The standard deviation shows how the measurements of the grouped variable were spread out from the mean. A low number for standard deviation means that most measurements were close to the mean while a high number means that the measurements were more spread out from the mean. The minimum value shows the lowest value for the correlation of the grouped variable with firm performance while the maximum value shows the highest value for the correlation of the grouped variable with firm performance.

Pearson Correlation Matrix (PCM)

Additionally, this study reports the Pearson correlations of all variables to show relationships amongst them. By doing so, all correlations amongst the individual variables used in this study are presented and can be pairwise compared amongst each other. Positive correlations in the PCM indicate that when one of the related variables increase in value, the other increases as well while negative correlations indicate that when one of the related variables increases in value the other will decrease. The closer the value is to one, the more perfect the correlation of the paired variables is. The asterisks in the table indicate whether the correlation is significant: one asterisk shows correlations significant at the 0.01 level while two asterisks show correlations significant at the 0.05 level.

Weighted Least Squares Regression (WLSR)

5. Results

5.1 Correcting for repeating measurements

Before presenting the final results of the analysis, a check for the influence of repeating

measurements from the same article was carried out. This was done by first conducting the WLSRs with the full sample of 56 observations (including repeating measurements from the same article) and

afterwards conducting the same WLSRs with the reduced sample that includes no repeating

measurements form the same article, leading to 26 observations in total. In order to create this reduced sample, one measurement from each article was randomly chosen as input. Afterwards, the WLSR results of both samples were compared against each other. As the results of the full sample showed some

differences compared to results of the reduced sample, it was decided that this research was continued with the reduced sample. By doing so, the regression is more robust and the results are more reliable, leading to an increased validity.

5.2 Descriptive Statistics

Table 1: Descriptive Statistics of all variables

Group Variable Mean N Std. Deviation Minimum Maximum

Leadership Level CIO Leaders 0,160 9 0,121 0,014 0,390 Leadership Level CEO Leaders 0,369 12 0,220 0,065 0,800 Leadership Level Project/Team

Leaders 0,389 7 0,176 0,170 0,610 Performance Type Financial Performance 0,285 13 0,237 0,014 0,800 Performance Type Project Performance 0,380 7 0,177 0,170 0,610 Performance Type Efficiency 0,320 8 0,183 0,029 0,610 Performance Type Effectiveness 0,419 4 0,212 0,195 0,610 Sector Manufacturing 0,298 15 0,243 0,014 0,800 Sector Services 0,290 10 0,271 0,014 0,800 Sector IT 0,390 8 0,198 0,117 0,590 Sector Healthcare 0,331 3 0,209 0,181 0,570 Sector NPO 0,266 4 0,152 0,117 0,470

Country United States 0,328 19 0,227 0,014 0,800 Country Emerging

Countries

0,304 9 0,180 0,117 0,590

Country Other Developed Countries 0,390 1 - 0,390 0,390 Data Collection Method Survey Data 0,311 22 0,187 0,029 0,610 Data Collection Method Archival Data 0,301 7 0,259 0,014 0,800

Research Setting Longitudinal 0,255 6 0,278 0,014 0,800 Research Setting Cross-sectional 0,331 20 0,191 0,029 0,610

Some first impressions can be observed from table 1. When looking at the group of leadership variables, the means show that CIO leadership related variables tend to have a lower correlation with firm performance compared to the CEO and project or team related leadership variables. Also, correlations of the CIO leadership variable are more centered around the mean compared to the other two leadership variables. This would indicate that CIO digital leadership has a smaller effect on firm performance compared to the other two levels of digital leadership.

Second, taking a look at the performance type variables provide some first impressions as well. The mean of correlation of the highest level of performance type, financial performance, shows the lowest mean of correlation with firm performance of all four levels of performance. Effectiveness shows the highest mean of correlation, although this variable is only included in four observations. Project or team performance shows a higher mean of correlation with firm performance as well, compared to financial performance. This gives the impression that variables for financial performance, the highest level firm outcome, are less affected compared to lower level variables like project or team performance.

Third, the descriptive statistics of the group of variables concerning the different industry sectors do not provide clear first impressions. The means of the different groups are close to each other which indicates that not much difference in correlations between the different sectors and firm performance occurs. One interesting aspect might be that the mean of the IT sector is slightly higher compared to the other groups, which gives a first impression that digital leadership might be more important in the IT sector compared to the other sectors. Also, the number of observations per grouped variable show that the sectors NPO and healthcare have only four and three numbers of observations respectively.

Fourth, the grouped variables concerning the countries in the samples show no interesting results. All means of the correlations with firm performance are close to each other while the group of other developed countries consists of only one observation.

Fifth, no first impressions can be observed from the groups of variables concerning the data collection methods. The means are all close to each other, and most included studies rely on survey data rather than archival data. Also, the minimum and maximum values do not provide interesting first impressions.

Sixth, some minor interesting first impressions from the descriptive statistics related to the research setting variable groups can be observed. The mean of the cross sectional research setting correlations looks like to lead to a stronger effect on firm performance compared to the longitudinal research setting correlations. This would indicate that by making use of cross sectional data stronger results are produced. However, the longitudinal research setting variable consists of only six

Finally, some interesting information from the descriptive statistics of the groups for analytical statistics can be derived. The mean of the correlations look to be somewhat lower for the regression group compared to the CB SEM group and the PLS SEM group, which are both SEM based analytical

techniques. This indicates that measurements gained by using both of the SEM based analytical techniques show stronger results for firm performance compared to the regression based analytical technique. However, the CB SEM group consists of only five measurements.

5.3 Pearson Correlation Matrix

Some first impressions can be derived as outcomes from observing this PCM. The first insight that can be observed is that the leadership variable at project or team level is significantly negatively correlated with financial performance and significantly positively with project or team performance. This would imply that the leadership variable at project or team level is correlated with project or team performance on many occasions while not so many times with financial performance.

Second, the leadership variable at the CIO level is significantly negatively correlated with project or team performance which implies that CIO leadership is not many times correlated with project

performance.

Third, the variable of financial performance is significantly positively correlated with the country variables for both the US and emerging countries. This would imply the US and emerging countries are many times correlated with financial performance.

Fourth, effectiveness is significantly positively correlated with NPOs and negatively significantly correlated with the US. This would imply that effectiveness is many times correlated with NPOs and only a few times with the US.

Fifth, the NPO variable is significantly negatively correlated with the US. This would mean that NPO variables are not so many times correlated with the US variables.

Finally, the survey variable is significantly positively correlated with the cross sectional data variable. This suggests that survey collected data is many times correlated with cross sectional data.

5.4 Weighted Least Squares Regression

Table 3: WLSR Analysis Results WLSR Model Sample Size Adjusted R Square Variable Unstandardized Coefficients

Beta t-value p-value Result

B Std. Error

As can be observed in the table, some interesting results become clear. In the first WLSR model that was run, concerning the different levels of digital leadership, the digital leadership at CEO level variable was used as the baseline. The results of the WLSR indicate that digital leadership at CIO level has a smaller effect on firm performance compared to digital leadership at the CEO level, while this finding is significant at below the 1% level. Additionally, the effect of digital leadership at project or team level on firm performance is not significantly different compared to the same effect from digital

leadership at the CEO level. This is in line with the first impressions derived from the descriptive statistics.

The second model, concerning the different performance types, was run with the project/team performance variable as the baseline. Looking at the results, the first impression is that the variables of financial firm performance and efficiency show a slightly weaker effect and that effectiveness shows a slightly stronger effect, however, none of these findings are significant. Therefore, it can be concluded that none of the grouped variables show significant differences with the project/team performance variable. The first impression of a weaker effect of financial performance is in line with the first impressions from the descriptive statistics, although this finding does not hold statistically.

The third model, concerning the different industry sectors, made use of the manufacturing sector as the baseline. The first impression observed in this model is that almost all variables are close to the baseline variable. One exception might be the IT sector variable which shows a somewhat stronger effect on firm performance, which is in line with the first impression of the descriptive statistics. However, this finding is not statistically supported as none of the variables show significant results. Therefore, none of the industry sectors have a statistically significant different effect on firm performance compared to the manufacturing sector.

In the fourth model, regarding the countries in the sample, the US was chosen as the baseline variable. Within this model no significant results were observed, which is in line with the first

impressions from the descriptive statistics. Therefore, it can be concluded that the differences between countries in the sample do not lead to significantly different effects on firm performance.

The fifth model concerns the differences amongst data collection methods. Within this model, survey data was chosen as the baseline variable. As the other variable of archival data is very close to the baseline variable, the first impression is that the data collection method has no influence on the results. This is supported by the findings of the WLSR as no significantly different effects are observed in the model.

sectional data leads to stronger results, which is in line with the first impression of the descriptive

statistics and the PCM. However, this first impression was not supported by the results as the finding was not significant, stating that the differences in research settings do not lead to significant differences effects on firm performance.

The final model concerns the analytical techniques used in the analysis of the original studies. Here, the regression analytical technique was chosen as the baseline variable. The first impression of this model is that both the variables regarding PLS and CB SEM show stronger effects on firm performance compared to the baseline variable of regression. The significance levels of both variables compared with the baseline show support for this first impression. Variables related to a PLS analytical technique have a significantly stronger effect on firm performance compared to regression at a significance level of below 5%. Variables related to a CB SEM analytical technique have a significantly stronger effect on firm performance compared to regression at a significance level of below 10%.

Table 4: WLSR Full Model

Sample Size

Adjusted R Square

Variable Unstandardized Coefficients Beta t-value p-value Result

6. Discussion

6.1 Main Findings

Levels of Leadership

The results indicate that stronger digital leadership in general leads to stronger firm performance. The first finding of this study concerns differences between the levels of digital leadership within this mechanism. Three levels of digital leadership were distinguished and compared: CEO leadership, as the top level of leadership in this study, CIO leadership, as the leadership variable at functional levels in this study, and project or team leadership, as the leadership variable at project levels in this study. By

distinguishing these three levels of digital leadership and comparing its effect on organizational performance, it can be demonstrated at which level stronger digital leadership results in the most beneficial effect on firm performance. The results of the WLSR indicate that digital leadership on the CEO level is more beneficial for organizational performance compared to leadership at the CIO level. Also, digital leadership at the project or team level show no significant differences related to its effect on organizational performance compared to leadership at the CEO level. Thus, it can be said that in the digital context, leadership of a CIO has a smaller influence on organizational performance compared to leadership of the CEO or leadership of the project or team managers.

The first finding, concerning a stronger effect of leadership at the CEO level on organizational performance compared to leadership at the CIO level is in line with existing theories regarding

managerial influence on firms. It is no surprise that digital leadership of CEOs will lead to stronger firm performance as it is already proven that managers influence the implementation of new technology and usage by subordinates (Leonard-Barton & Deschamps, 1988), strong TMTs can be associated with firm performance (Smith et al., 2006), and that power concentration strongly affects decision making (Greve & Mitsuhashi, 2007). Combining such insights with the upper echelons theory, it can be said that digital leadership leads to more digital initiatives within the firm which, in turn, will lead to stronger firm performance. However, these insights do not clearly explain why digital leadership of the CEO leads to a higher level of firm performance improvements compared to digital leadership of the CIO. An

Hu, 2014). Therefore, a reason behind the stronger effect of digital leadership at the CEO level can last in the power distribution within TMTs; CEOs and CIOs need to agree on decisions, which results in a smaller direct power of a CIO. In the end, the CIO needs permission from the CEO for starting top level digital initiatives as a final call in the decision making process.

However, the digital leadership effect at the team or project level was not significantly different compared to the CEO level. This implies that the digital leadership of team or project managers is at a comparable level associated with firm performance as digital leadership at the CEO level. Multiple explanations for this observation can be addressed. First, project and team managers are directly influencing subordinates which are supervised at the operational level of the organization. By doing so, they can influence and direct decisions made at this level, as the project and team members are likely to follow their direct leaders (González-Cruz et al., 2019). Also, strategic influence, both down- and upwards, relies within middle managers and especially the downward strategic influence of middle managers can lead to improved firm performance (Floyd & Wooldridge, 1997). Such effects discussed by multiple studies of both direct supervisors and middle managers on firm performance (Yang et al., 2010), indicate the main reasoning behind a relatively strong effect of digital leadership at the project or team level on firm performance. Moreover, when launching digital initiatives within the organization, lower level managers are of utmost importance. When adopting new technologies, lower level managers are influencing the extent to which such an innovation is adopted and ultimately used by their subordinates (Leonard-Barton & Deschamps, 1988). Focusing on this concept may provide additional arguments for the similar effect of digital leadership at the CEO and project or team level. The digital leadership of the CEO may lead to the launch of digital initiatives within the organization. However, in order to have a beneficial effect on firm performance, such initiatives need to be adopted at operational levels as well. In order to implement and adopt such initiatives at operational levels, project and team managers are necessary as they are able to influence their subordinates. The stronger the digital leadership of such a project or team manager, the more likely he or she will see the importance of the digital initiative and the more likely it will be that he or she puts effort in implementing the initiative at the operational level. Moreover, both levels of digital leadership might work alongside eachother: top level leadership launches digital initiatives and project and team level leadership make sure the initiatives are put to practice. This can ultimately lead to a higher level of firm performance. Therefore, not only digital leadership at the CEO level matters but at project and team level as well.

with project and team performance on many occasions while not with financial performance. Therefore, this lower level leadership of project or team managers is proven to be beneficial for lower level firm performance like project and team performance while not on overall financial performance. This is in line with aforementioned insights that lower level managers influence their subordinates and have therefore influence on lower level operations only. It could be that this digital leadership of team or project managers is also affecting higher level financial performance as well, but to find support for this assumption more research is needed.

Types of Performance

The second finding of this study concerns the different types of firm performance. Four types of firm performance were distinguished and compared: financial performance, project and team

performance, efficiency, and effectiveness. By distinguishing these four levels of performance and comparing the level of its effect, it can be shown at which type of organizational performance the performance gains become most apparent. The results of the WLSR show that there are no significantly different effects related to the different types of performance and their gains on overall firm performance. Therefore, it cannot be determined from this study which type of firm performance is related to the highest level of performance gains. More research is needed to find support for the first impression that financial performance showed weaker effects compared to the other, lower level, performance types.

Research Context

The third finding of this study concerns the research context. By research context, aspects like the industry sector and country characteristics of the sample are meant. By distinguishing such research contexts and by comparing its effect on organizational performance, it can be shown whether differences in country or industry sector contexts matter. If differences occur, it becomes clear in which countries or industry sectors the digital leadership aspects result in the strongest effect on organizational performance. The results of the WLSR show that neither differences in the country context, nor the industry sector contexts make a difference regarding their effect on organizational performance. Therefore, this study indicates that the effect of digital leadership on firm performance does not depend on a certain country or industry sector context. More research is needed to find support for the first impression that the IT sector showed stronger results.

Research Methods

by either survey or archival data, the research setting, grouped by either a longitudinal or cross sectional setting, and the analytical technique, grouped by either PLS SEM, CB SEM, or regression based

techniques. Results of the WLSR regarding the data collection methods indicate that no significant differences between the usage of survey and archival data exists related to their effect on firm

performance. Results of the WLSR concerning the research settings result in no significant findings as well, indicating that the usage of either longitudinal or cross sectional obtained data in the sample has no effect on the analysis. To find support for the first impression that cross sectional data lead to stronger results more research is needed. However, the results of the WLSR related to the analytical technique do show significant differences. As the results indicate, PLS SEM and CB SEM lead to stronger results compared to observations that have used a regression analysis as analytical technique. This demonstrates that SEM based analytical techniques help researchers to find stronger results compared to when they make use of regression based analytical techniques. Possible explanations for this difference between SEM based and regression based analysis become apparent when taking a look at the advantages of SEM over regression. The main advantage of using SEM over regression methods is that SEM integrates the measurement model and the structural model into a unified assessment. SEM is able to include the error terms into a simultaneous model. Subsequently, this model is estimated, either in one single model like CB SEM does or iteratively like in PLS. These results are displayed in one model including path

estimates of measurements and structural models. By doing so, measurement and structural relationships are measured better in CB SEM and PLS. Thus, SEM is able to create better estimations compared to regression (Gefen et al., 2011). Such advantages of using SEM based analytical techniques compared to regression based analysis are in line with the stronger results of the observations that made use of SEM based analytical techniques. Therefore, the stronger results of the SEM based analytical techniques are possibly the result of the advantages that SEM based analytical techniques hold over regression based analytical techniques.

6.2 Theoretical Implications

Digital Leadership and Firm Performance

digital leadership seems to be beneficial for firm performance. Finding support for this mechanism can be explained as an expansion of the upper echelons theory in the field of information systems research. In general, the main implication of this study is that the trend of digital leadership and its association with stronger firm performance is identified. This finding is in line with findings from the past like IT

awareness within the firm and its effect on stronger firm performance (Yayla & Hu, 2014) but shapes its focus more to the leadership aspect and is including lower level management as an important trigger for digital leadership as well. Further, the observation of this trend leads to more questions regarding the upper echelons theory in a digital setting. For example, it can be studied whether an increase in digital leaders result in more digital initiatives within the firm. When also including a firm performance variable, it can be assessed whether digital leadership leads to more digital initiatives within the firm, and whether these initiatives in turn leads to an improved firm performance. Thus, it can be interesting to study

whether the number of digital initiatives have a mediating role within the relationship of digital leadership and firm performance.

Currently, not many studies focus on the phenomena of digital leadership in the sense of having a leader with a digital background in place within firms. Having observed this trend of digital leadership leading to firm performance is an interesting case for scholars in the information systems research field as leadership is a domain that can be explored further. However, more research concerning this observation is necessary. Future researchers could study this trend more in depth, for example by using a large sample of firms and including its leaders' backgrounds by labeling it on the level of digital backgrounds (either through digital professional experience, digital educational experience, or both) and including measures for firm performance.

Levels of Digital Leadership

Second, not only digital leadership at the highest level matters, like the upper echelon theory would explain (Hambrick & Mason, 1984), but at project and team level as well. This indicates that backgrounds of leaders are not only influencing decisions made at top and functional levels, but at project levels as well. When relating this mechanism to digital leadership the strength of this project level effect is comparable to the CEO effect. Thus, not only TMT executives but lower level digital managers also significantly affect firm performance.

CIOs and the digital leadership context, as CIOs are usually the more ‘digital’ managers present within a firm.

Generalizable Findings

The non-significant findings for the variable groups of performance type, industry sector, country, data collection method, and research setting indicate that the findings of this study are generalizable among these aspects. As the differences of these groups in the samples did not lead to significant differences in the results, it can be expected that differences in such groups are not important and that the results of this study is generalizable to other contexts amongst. Thus, in the case of the group of performance type, it can be expected that digital leadership will positively affect other performance measures that are not included in this study as well. For industry sectors, it can be expected that digital leadership will also have a positive effect on firm performance in industry sectors that are not included in this study. Related to countries, the expectation will be that in other countries the same results will hold. Also, it is expected that differences in data collection and the research setting will not influence results. This would imply that it can be expected that when using other variables within these groups the same results will hold.

However, as the first impression of the descriptive statistics and the PCM indicated some

differences between groups, it could be the case that when making use of larger samples some differences between these groups do become apparent. Thus, more research is advised to find potential differences amongst the type of firm performance, research setting, and industry sectors and their effect on firm performance in this digital leadership context. Performing a study in the future which is able to capture a larger sample size might find significant differences amongst these variables and is therefore necessary.

interesting as well. It might be the case that after appointing digital leaders in a firm, it takes time until their influence affects firm outcomes. Building on this idea, it can be logical that cross sectional data finds stronger effects on firm performance. Thus, it will be interesting to find support for this assumption in the future whether it might be the case that choosing for cross sectional data leads to stronger results.

6.3 Practical Implications

The practical implications of this study impact both researchers and managers. For managers, the implications of this study limit to focusing on the digital backgrounds of CEOs, CIOs, and project and team managers. As the results show, a higher level of digital leadership of the CEO or the project and team managers lead to stronger benefits in firm performance compared to a higher level of digital

leadership of the CIO. This indicates that appointing a CEO or project and team members with a stronger digital background has a strong positive effect on the performance of the firm. Appointing a CIO with a stronger digital background has a smaller positive effect on the performance of the firm compared to the CEO or project and team managers. Therefore, the results of this study show that it is wise to appoint either or both CEOs or project and team managers with a strong digital background before appointing a CIO with a stronger digital background in order to boost firm performance. Moreover, enhancing digital leadership within a firm will lead to improved firm performance in all studied levels of performance: financial performance, efficiency, effectiveness, and project and team performance. As no differences between these four groups were identified, managers can choose to increase the digital leadership within their firm and can, as a result, expect to positively affect all of these four performance dimensions.

Another important insight related to digital leadership is that not only leadership at top management team levels lead to increased firm performance but at project or team level as well at a comparable level. Therefore, lower level managers might have a larger influence on firm performance than initially thought. Future research could focus on studying this phenomenon and gaining more insights in the effect of project and team leadership on firm performance compared to top level leadership.

Additionally, it is important to know that most other grouped variables showed non-significant findings in the WLSR. This would mean that for the related groups of variables, the findings are

generalizable in different contexts. Thus, for the grouped variables of types of firm performance, industry sectors, countries, data collection methods, and research settings the findings of this study are

generalizable in different contexts.

6.4 Limitations

This study holds some limitations as well. First, the number of observations used in the sample can be seen as a limitation. The final sample consists of 26 unique observations, which is not a very large sample size. Usually, such a sample size is sufficient for a meta-analysis however a larger sample might lead to more robust results.

Second, the strong presence of China and the US as countries in the sample might be seen as a limitation. Almost all original studies, nineteen times the US and six times China, make use of a sample with data from one of these countries. As a result, it might be questioned whether the results of this study are generalizable to firms that originate from other countries. However, this seems to be the case as the results showed no differences amongst country contexts.

Third, not all observations are perfectly measuring digital leadership as the leadership variable. Most of the variables used in this study are measuring the digital background of the leader while all are measuring a leadership aspect in a digital setting. Therefore, for future research a perfect measure for all leadership variables is needed and could include either or both professional- or/and educational digital backgrounds.

7. Bibliography

Acker, O., Herzog, M., & Péladeau, P. (2017). The new class of digital leaders. In Tech & Innovation (Issue 88). https://www.strategy-business.com/article/The-New-Class-of-Digital-Leaders?gko=820f1 Aladwani, A. (2002). An Integrated Performance Model Information Systems Projects. Journal of

Management Information Systems (Vol. 19, Issue 1, pp. 185–210). https://doi.org/10.1080/07421222.2002.11045709

Amiri, P., & Moqri, M. (2018). A Citation Analysis of the Basket of Eight IS Journals.

Argo, J. J., & Main, K. J. (2004). Meta-Analyses of the Effectiveness of Warning Labels. In Journal of Public Policy & Marketing (Vol. 23, Issue 2, pp. 193–208).

https://doi.org/10.1509/jppm.23.2.193.51400

Bangert-Drowns, R. L. (1992). Review of developments in meta-analytic method. In Methodological issues & strategies in clinical research. (pp. 439–467). https://doi.org/10.1037/10109-034 Banker, Banker, Hu, Pavlou, & Luftman. (2011). CIO Reporting Structure, Strategic Positioning, and

Firm Performance. In MIS Quarterly (Vol. 35, Issue 2, p. 487). https://doi.org/10.2307/23044053 Bankewitz, M., Aberg, C., & Teuchert, C. (2016). Digitalization and Boards of Directors: A New Era of

Corporate Governance? In Business and Management Research (Vol. 5, Issue 2). https://doi.org/10.5430/bmr.v5n2p58

Bouwman, H., Nikou, S., Molina-Castillo, F. J., & de Reuver, M. (2018). The impact of digitalization on business models. In Digital Policy, Regulation and Governance (Vol. 20, Issue 2, pp. 105–124). https://doi.org/10.1108/dprg-07-2017-0039

Boyd, P., Harting, A., Hiebeler, M., Kolev, N., & Redweik, J. (2015) Survival Through Digital Leadership. Deloitte Digital. Accessed at 3/6/2020. Retrieved from:

https://www2.deloitte.com/content/dam/Deloitte/de/Documents/technology/Survival%20through%2 0Digital%20Leadership_safe.pdf

Chan, R. Y. K., & Ma, K. H. Y. (2017). Impact of executive compensation on the execution of IT-based environmental strategies under competition. In European Journal of Information Systems (Vol. 26, Issue 5, pp. 489–508). https://doi.org/10.1057/s41303-017-0052-3

Chen, D. Q., Preston, D. S., & Xia, W. (2010). Antecedents and Effects of CIO Supply-Side and Demand-Side Leadership: A Staged Maturity Model. In Journal of Management Information Systems (Vol. 27, Issue 1, pp. 231–272). https://doi.org/10.2753/mis0742-1222270110 Das, P., Verburg, R., Verbraeck, A., & Bonebakker, L. (2018). Barriers to innovation within large

financial services firms. In European Journal of Innovation Management (Vol. 21, Issue 1, pp. 96– 112). https://doi.org/10.1108/ejim-03-2017-0028

Ding, F., Li, D., & George, J. F. (2014). Investigating the effects of IS strategic leadership on

organizational benefits from the perspective of CIO strategic roles. In Information & Management (Vol. 51, Issue 7, pp. 865–879). https://doi.org/10.1016/j.im.2014.08.004

Dong, S., Xu, S. X., & Zhu, K. X. (2009). Research Note—Information Technology in Supply Chains: The Value of IT-Enabled Resources Under Competition. In Information Systems Research (Vol. 20, Issue 1, pp. 18–32). https://doi.org/10.1287/isre.1080.0195

Egelhoff, W. G. (1991). Information-Processing Theory and the Multinational Enterprise. In Journal of International Business Studies (Vol. 22, Issue 3, pp. 341–368).

https://doi.org/10.1057/palgrave.jibs.8490306

from: https://www.ecb.europa.eu/pub/economic-bulletin/focus/2018/html/ecb.ebbox201807_04.en.htm

Fan, J., Zhang, P., & Yen, D. C. (2014). G2G information sharing among government agencies. In Information & Management (Vol. 51, Issue 1, pp. 120–128).

https://doi.org/10.1016/j.im.2013.11.001

Field, L. C., & Mkrtchyan, A. (2017). The effect of director experience on acquisition performance. In Journal of Financial Economics (Vol. 123, Issue 3, pp. 488–511).

https://doi.org/10.1016/j.jfineco.2016.12.001

Floyd, S. W., & Wooldridge, B. (1997). Middle Management’s Strategic Influence and Organizational Performance. In Journal of Management Studies (Vol. 34, Issue 3, pp. 465–485).

https://doi.org/10.1111/1467-6486.00059

Gefen, Gefen, Rigdon, & Straub. (2011). Editor’s Comments: An Update and Extension to SEM Guidelines for Administrative and Social Science Research. In MIS Quarterly (Vol. 35, Issue 2, p. iii). https://doi.org/10.2307/23044042

Gemino, A., Reich, B. H., & Sauer, C. (2007). A Temporal Model of Information Technology Project Performance. In Journal of Management Information Systems (Vol. 24, Issue 3, pp. 9–44). https://doi.org/10.2753/mis0742-1222240301

González-Cruz, T. F., Botella-Carrubi, D., & Martínez-Fuentes, C. M. (2019). Supervisor leadership style, employee regulatory focus, and leadership performance: A perspectivism approach. In Journal of Business Research (Vol. 101, pp. 660–667). https://doi.org/10.1016/j.jbusres.2019.01.065

Greve, H. R., & Mitsuhashi, H. (2007). Power and Glory: Concentrated Power in Top Management Teams. In Organization Studies (Vol. 28, Issue 8, pp. 1197–1221).

https://doi.org/10.1177/0170840607075674

Hambrick, D. C. (2007). Upper Echelons Theory: An Update. In Academy of Management Review (Vol. 32, Issue 2, pp. 334–343). https://doi.org/10.5465/amr.2007.24345254

Hambrick, D. C., & Mason, P. A. (1984). Upper Echelons: The Organization as a Reflection of Its Top Managers. In The Academy of Management Review (Vol. 9, Issue 2, p. 193).

https://doi.org/10.2307/258434

Hedges, L. V., and Olkin, I. (1985). Statistical Methods for Meta Analysis, San Diego, CA: Academic Press.

HomRoy, S., & Slechten, A. (2016). Board Expertise, Networked Boards and Environmental Performance. SSRN Electronic Journal. https://doi.org/10.2139/ssrn.2873760

Hunter, J.E. and Schmidt, F.L. (1990). Methods of Meta-Analysis: Correcting Error and Bias in Research Findings. Newbury Park, CA: Sage.

Isaksson, A. J., Harjunkoski, I., & Sand, G. (2018). The impact of digitalization on the future of control and operations. In Computers & Chemical Engineering (Vol. 114, pp. 122–129).

https://doi.org/10.1016/j.compchemeng.2017.10.037

Jiang, J. J., Klein, G., & Discenza, R. (2002). Pre-project partnering impact on an information system project, project team and project manager. In European Journal of Information Systems (Vol. 11, Issue 2, pp. 86–97). https://doi.org/10.1057/palgrave/ejis/3000420

Karahanna, E., & Preston, D. S. (2013). The Effect of Social Capital of the Relationship Between the CIO and Top Management Team on Firm Performance. In Journal of Management Information Systems (Vol. 30, Issue 1, pp. 15–56). https://doi.org/10.2753/mis0742-1222300101

Technology: A Knowledge-Based View of Behaviors, Outcome, and Consequences. In Journal of Management Information Systems (Vol. 23, Issue 3, pp. 129–162). https://doi.org/10.2753/mis0742-1222230306

Lambrou, M., Watanabe, D., & Iida, J. (2019). Shipping digitalization management: conceptualization, typology and antecedents. In Journal of Shipping and Trade (Vol. 4, Issue 1).

https://doi.org/10.1186/s41072-019-0052-7

Langer, N., Slaughter, S. A., & Mukhopadhyay, T. (2014). Project Managers’ Practical Intelligence and Project Performance in Software Offshore Outsourcing: A Field Study. In Information Systems Research (Vol. 25, Issue 2, pp. 364–384). https://doi.org/10.1287/isre.2014.0523

Leidner, D. E., Preston, D., & Chen, D. (2010). An examination of the antecedents and consequences of organizational IT innovation in hospitals. In The Journal of Strategic Information Systems (Vol. 19, Issue 3, pp. 154–170). https://doi.org/10.1016/j.jsis.2010.07.002

Leonard-Barton, D., & Deschamps, I. (1988). Managerial Influence in the Implementation of New Technology. In Management Science (Vol. 34, Issue 10, pp. 1252–1265).

https://doi.org/10.1287/mnsc.34.10.1252

Liang, H., Wang, N., Xue, Y., & Ge, S. (2017). Unraveling the Alignment Paradox: How Does

Business—IT Alignment Shape Organizational Agility? In Information Systems Research (Vol. 28, Issue 4, pp. 863–879). https://doi.org/10.1287/isre.2017.0711

Lu, Y., & Ramamurthy, K. (2010). Proactive or reactive IT leaders? A test of two competing hypotheses of IT innovation and environment alignment. In European Journal of Information Systems (Vol. 19, Issue 5, pp. 601–618). https://doi.org/10.1057/ejis.2010.36

McKinsey Global Institute. (2019). Twenty-five years of digitization: Ten insights into how to play it right. Accessed at 10-06-2020. Retrieved from:

https://www.mckinsey.com/~/media/mckinsey/business%20functions/mckinsey%20digital/our%20i

nsights/twenty-five%20years%20of%20digitization%20ten%20insights%20into%20how%20to%20play%20it%20ri ght/mgi-briefing-note-twenty-five-years-of-digitization-may-2019.ashx

Mithas, Mithas, Ramasubbu, & Sambamurthy. (2011). How Information Management Capability Influences Firm Performance. In MIS Quarterly (Vol. 35, Issue 1, p. 237).

https://doi.org/10.2307/23043496

Neubert, M. (2018). The Impact of Digitalization on the Speed of Internationalization of Lean Global Startups. In Technology Innovation Management Review (Vol. 8, Issue 5, pp. 44–54).

https://doi.org/10.22215/timreview/1158

Nonaka, I., Umemoto, K., & Senoo, D. (1996). From information processing to knowledge creation: A Paradigm shift in business management. In Technology in Society (Vol. 18, Issue 2, pp. 203–218). https://doi.org/10.1016/0160-791x(96)00001-2

Nwankpa, J. K., & Datta, P. (2017). Balancing exploration and exploitation of IT resources: the influence of Digital Business Intensity on perceived organizational performance. In European Journal of Information Systems (Vol. 26, Issue 5, pp. 469–488). https://doi.org/10.1057/s41303-017-0049-y Oehmichen, J., Schrapp, S., & Wolff, M. (2017). Who needs experts most? B oard industry expertise and

strategic change—a contingency perspective. In Strategic Management Journal (Vol. 38, Issue 3, pp. 645–656). https://doi.org/10.1002/smj.2513