Subsidiary power in the age of data driven decisions : an exploratory study

(1)

Subsidiary Power in the Age of Data Driven

Decisions: An exploratory study

By

Hristo Andreev

Student ID: 11439831

Date of submission: June 22 2018

MSc Business Administration: International Management

Master Thesis

University of Amsterdam

First supervisor: Dr. Markus P. Paukku

(2)

[- 2 -]

Statement of Originality

This document is written by Student Hristo Andreev who declares to take full

responsibility for the contents of this document.

I declare that the text and the work presented in this document are original and that

no sources other than those mentioned in the text and its references have been used

in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of

completion of the work, not for the contents.

(3)

[- 3 -]

In the last 50 years the field of international management, has experienced a gradual shift, from the home country and headquarters’ capabilities of controlling its subsidiaries, to an era where the subsidiary is the central piece in the MNC network (Kostova et al. 2016). Consequently, just as the subsidiary has become central for the MNE, the IB field in the last 10 years is also focused on the emerging concept of Big Data and Data Driven Decisions, which are seen as a way to optimize and make decision making efficient. This research paper aims to understand how exactly Data Driven Decisions affect the power of the subsidiary. To do exactly that, six semi-structured interviews were conducted with managers of two big multinational subsidiaries, to gain insight as to how exactly data driven managers are empowered by Big Data. The subsidiary perspective was chosen, because of the trend in IB literature in the last 20 years to focus on the subsidiary as a competence creating entity with its own strategic goals, differing from headquarters. The results reached in this thesis prove that data driven subsidiary units have higher subsidiary power, shorter budget allocation approval and experience more autonomy.

(6)

[- 6 -]

1. Introduction

In the last ten years, there has been a shift in management paradigm from intuition to Data Driven Decision making (Brynjolfsson & McCafee, 2012). But why is data so important is what Wamba et al. (2015) ask themselves. The answer to that that they give is that big data has the power to influence transformation processes within MNEs by generating business value (Wamba et al. 2015). Gobble (2013) calls Big Data the next big thing in innovation, to which Manyika et al. (2011) add that it is a frontier of competition and productivity. What strengthens these hyperboles defining data, is the capability of big data to change the competition environment, by transforming processes and altering the corporate ecosystem (Brown et al, 2012). Strengthening these notions, firms have started gathering extremely large quantities of data from their consumers in order to further develop their products and increase the viability of decision making (Brynjolfsson et al. 2011). Big Data has become a focus of modern business decisions and is even seen as one of the top tech trends in the 2010s. Different big companies are increasingly adopting the practice of Data Driven Decisions, as outside of operational data collection inside the company, everyday items like phones, vehicles and other electronic devices become a source of data influx (Brynjolfsson et al. 2011). Searches of the term “Big Data” have been ever increasing in the last few years from 252 000 in 2011 to an impressive 1.39 billion in 2012 (Flory, 2012). What has rapidly increased the adoption of Big Data is the ever-increasing adoption by consumers of mobile phones the rise of different platforms like Twitter and Facebook (Wamba et al. 2015). This technological revolution and adoption has transmitted itself into the business world causing a shift from intuition-based management to a

(7)

[- 7 -]

data driven approach to decision making (Brynjolfsson & McCafee, 2012). Consequently, Big Data has become more powerful than analytics used in the past as it brings more precision, accuracy and better predictive capabilities to companies implementing it (Brynjolfsson & McCafee, 2012). Increasingly Big Data is becoming the way to move forward in regard to company performance as those who have adopted Data Driven Decision have proven to perform 6% better than their competition (Brynjolfsson & McCafee, 2012). Elaborating on this Brynjolfsson & McCafee (2012) show that companies who regard themselves as data driven had this better performance after taking into account the contributions of labour, capital and services, which means organizational structure had no impact. Going further, companies implementing big data into analytics have proven to achieve a 15 to 20 percent higher Return of Investment (ROI) (Perrey et al. 2011 in Wamba et al. 2015). Seeing as Big Data has emerged as the next big thing in innovation (Gobble, 2013) it is interesting that the managerial implications of it are mostly unexplored with little research on how Big Data helps companies make better decisions. Moreover, even when looking at the concept of Data Driven Decisions, the implications have mostly been looked at from the point of view of an organizational management, without taking into account the shift in literature from headquarters to subsidiary (Kostova et. al. 2016) and the power struggle within an MNE (Mudambi & Navarra, 2004). Davenport et al. (2012) do note that Big Data unlocks organizational value, by putting new organizational capabilities and value to the table.

But in this notion, that data gives leverage to unleashing the full potential of a company, one underdeveloped theme is the power struggle between headquarters and subsidiary. Even though the importance of the subsidiary is clearly bigger than 50 years ago, when literature started viewing the MNE as a network, there has been no research on how big data and data driven decisions would affect the subsidiary and its autonomy and power.

(8)

[- 8 -]

Kostova et al. (2016) see the headquarter-subsidiary (HQS) relationship as one of the central points of the modern international management field. Birkinshaw (1996) strengthens this opinion by saying that an increasingly important issue for MNE headquarters is how to integrate their globally dispersed value-adding activities. The theory of HQS has gradually evolved in the last 50 years from the more formal and bureaucratic mechanisms to the more informal and networked models. While in the past, more emphasis had been put on the headquarters, and subsidiaries were looked at as their agents, since the rise of globalization, the role of the subsidiary has gradually been more emphasized in the HQS relationship literature (Kostova et al. 2016). A frequent notion argued in subsidiary research is that there must be tough control over subsidiary decision-making by headquarters, because of the invested capital by headquarters, but more importantly that subsidiaries left to their own design may hinder corporate strategy at a micro-level (Dorrenbacher & Gammelgaard, 2011). As Mudambi and Navarra (2004) note knowledge is power, and the large quantities of available data are knowledge that can be leveraged by subsidiaries in order to negotiate for rents and influence with headquarters.

Most of the big data management literature has emphasized on an increased profitability of the company as a whole and has not taken into consideration what effect would the implementation of data driven decisions by subsidiaries have on their relationship. With the velocity and accessibility big data has subsidiaries might no longer have the intangible resources needed to increase their power within the MNC, as well as this accessibility might level the playing field between the different subunits of the MNE. Continuing from data driven decisions might make it easier for subsidiaries to influence their parent company because of having an infinite resource of knowledge creation which Mudambi & Navarra (2004) have found to cause a shift in the level of subsidiary power. This leaves a clear gap in theory about the implementation of

(9)

[- 9 -]

big data on an intra-firm level and how, decisions based on data from the subsidiary affect the power it has when delegating with headquarters.

Based on the importance that data has played in the last few years as well as the shift in international management literature from headquarters to subsidiary, we will give an exploratory view on the effect Data Driven Decisions have on subsidiary power and autonomy. In the next paragraphs we will follow the introduction of literature on the parent-subsidiary relationship as it is a central topic in IM literature with the increasing role of subsidiaries as a competency creating unit (Mudambi & Navarra, 2004). Afterwards, the concept of subsidiary power and autonomy will be established. Lastly a literature overview will be given of Big Data and Data Driven Decisions, as the newest concepts in IM literature.

This thesis is qualitative and exploratory in nature, as the concepts of subsidiary power and data driven decision making have not been combined in a research up until this point.

1. Literature review

In the next paragraphs we will give a short introduction to the main topics this thesis is going to research. Starting with an overview of the parent subsidiary-relationship, which will show how the role of the subsidiary in the MNE has changed historically. Then we will review the literature on subsidiary power and autonomy which are the phenomena that are at the centre of this thesis. Finally, data driven decisions and big data will be defined.

2.1. Headquarter-subsidiary relationship

Headquarter-subsidiary relationships have long been the central topic of International Business theory (Kostova et al., 2016). Many definitions have been given about this field but for this thesis, the definition of Kostova et al. 2016 will be used. In their paper, the HQ-Subsidiary relationship is defined as the headquarters coordination and control of their geographically dispersed value-adding units.

(10)

[- 10 -]

The headquarter-subsidiary relationship literature has been changing for the last 60 years from focusing on the headquarters to focusing on subsidiaries and their knowledge creation mechanics, which we can see best in Kostova et al. (2016) paper, which takes a good look at it, analysing articles on the topic through the years. What we can take from the authors is that at the beginning of their research up until the 1990s the emphasis in the headquarters-subsidiary relationship was on the role of the headquarters (Kostova et al. 2016). Later on, that concentration of literature shifted to informal methods, which put the subsidiary in the spotlight (Kostova et al. 2016). This change in paradigm came to be because of two important changes- the change in the global environment the MNE operates in, and the change of focus of the International Management field (Kostova et al. 2016). The change of the global environment is a consequence of western MNEs globalization, the complexity of global operations, which changed the focus on subsidiary, and lastly the rise of the emerging markets and the importance of context in them (Kostova et al. 2016). The bigger shift came from the importance of subsidiaries to MNC performance overall. The perspective in the 2000 changed completely to the subsidiary as authors such as Lee & Williams (2007) put emphasis on the autonomous and entrepreneurial behaviour of subsidiaries as a driver for the MNE. Going even further into the 2000s the focus was on the knowledge assets that the subsidiary brings to the table, and how knowledge from multiple centres within the MNC can be used to create companywide innovation (Asakawa & Lehrer, 2003)

The change in the IM field is from the hierarchical model to a networked model. This shift is in line with what Birkinshaw et al. (1998) elaborate on- the contribution of subsidiaries to MNEs FSAs is growing. Subsidiaries have started playing an ever-increasing role in the development of firm-specific advantages (FSA) not only through their role as an agent of headquarters but also through their own subsidiary initiative (Birkinshaw et al. 1998). What they argue is that while research has concluded that FSAs originate from the parent rather than

(11)

[- 11 -]

the subsidiary, the subsidiary through its own initiative and specialized resources can contribute to the MNE developing FSAs (Birkinshaw et al. 1998). This is further strengthened by Andersson et al (2007) who take a good look at the influence of subsidiaries inside the federative MNC. Papers such as that in the mid-2000s are evidence of the evolution of the parent-subsidiary relationship theory moving from headquarters’ perspective to the subsidiary and its value creating activities.

Taking the role of the subsidiary a step further Mudambi et al. (2004) started talking about subsidiary bargaining power and the concept of knowledge transfer apart from the typical parent-subsidiary stream. The authors identify rent seeking from the subsidiary was the cause of opportunistic behaviour, which destroy value for the MNE. The paper identifies the role of knowledge as an intangible asset in the development of subsidiary specific advantages, which help when maximizing the subsidiary’s share of MNE rents (Mudambi et al. 2004). This paper contributes by going outside the box of the traditional knowledge flows- from parent to subsidiary and looks at three other streams of knowledge between headquarter and subsidiary. The first stream is from subsidiary to parent, which would increase the bargaining power of the subsidiary. This type of flow is called by Mudambi and Navarra (2004) knowledge transfer and are key for headquarters to leverage the local competencies and integrate knowledge into the MNE network. In the context of Big Data which has high velocity of movement and integration it can be seen how this knowledge flow can be affected as the rate at which the knowledge stream from subsidiary to parent will be affected.

The second and third flows are from subsidiary to location and vice versa. The flows from the location to the subsidiary are comprised of the way the subsidiary learns to leverage on local competence exploitation and utilization (Mudambi & Navarra, 2004).

(12)

[- 12 -]

The last flow is from other subsidiaries and the parent to subsidiary, which lowers the bargaining power of the subsidiary (Mudambi et al. 2004). The main point Mudambi et al. (2004) make is that control over knowledge within the MNE gives the subsidiary a competitive advantage when compared to other subsidiaries, which in return leads to a bigger rent share. Further examining the literature on subsidiary power Mudambi et al (2014) distinguish between two types of subsidiary power in the relationship with headquarters- functional and strategic. The first relates to the power, connected to a specific function in the company like R&D, while strategic power is in line with a wider span of influence and relates to firm wide decision-making (Mudambi et al. 2014). The authors also provide a look at headquarter-subsidiary relationship through the lenses of resource dependency theory. They look at the modern MNE as a set of differentiated networks, in which some subsidiaries have augmenting power while other control resources on which the MNE depends (Mudambi et al. 2014). The second type of subsidiaries are those on which the MNE is dependent, giving them power over the strategic decision-making in the company. After establishing the leading role of knowledge and knowledge transfer in the HQ-subsidiary relationship, Tavani et al. (2016) take a look at reverse knowledge transfer from subsidiary to HQ and its effect on subsidiary bargaining power and autonomy. The biggest contribution this paper makes is establishing the moderating role of internal and external embeddedness on subsidiary power.

The shifting view of international management scholars on the topic of parent subsidiary relationship, can be clearly seen in the review on IM literature above. Based on all these papers, it can be seen that the subsidiary is playing a bigger role than ever, as a part of the MNE. Consequently, the choice of using the subsidiary perspective when analysing the effects of Data Driven Decisions is justified. This notion is based on the fact that IM literature has steadily been moving in the direction of the subsidiary in the last 50 years (Kostova et al. 2016).

(13)

[- 13 -]

After we have taken a look at the parent-subsidiary relationship and how it has changed, it is time to introduce the concept of subsidiary power and autonomy and define the powers that might be most affected by data driven decisions.

2.2 Subsidiary power

With the focus in IB literature shifting from headquarters to subsidiary (Kostova et al., 2016) a growing sub-topic that is of interest to IB scholars is the concept of subsidiary power. The most commonly used definition of power is the one by Dahl (1957). Dahl defines power as the ability to force others to do what they otherwise would not. While this type of definition can usually be used, in the case of subsidiaries it is not the most appropriate to use as no matter how much power they have they are always vetoed by headquarters (Dorrenbacher & Gammelgaard, 2011). This is further strengthened by Baker, Gibbsons and Murphy (2002) who state that the subsidiary’s decision rights are “loaned not owned” which gives headquarters the right of veto and the power to overturn any decision made in the subsidiary (Mudambi & Pedersen, 2007). Based on this the most frequently used definition of Power is insufficient for the analysis of subsidiaries. For the purpose of the thesis, we will use the definition of Dorrenbacher & Gammelgaard (2006) and establish subsidiary power as the ability of the subsidiary to influence their parent in their strategic and organizational objectives (Dorrenbacher & Gammelgaard, 2011). Consequently, when talking about subsidiary power, we have to distinguish between formal and real authority (Aghion & Tirole, 1997). In the context of a subsidiary formal authority, the right to make decisions is impossible to have, which is why they focus on holding real authority (Dorrenbacher & Gammelgaard, 2011) and more specifically for the thesis intangible assets, knowledge.

In their paper, Dorrenbacher & Gammelgaard (2011) identify four types of subsidiary power- micro-political bargaining power, systemic power, resource-dependency power and institutional power. The concept of micro-political bargaining power relates to the way

(14)

[- 14 -]

subsidiaries influence headquarters by combining their initiatives and manipulative behaviour (Surlemont, 1998). Consequently, Mudambi and Navarra (2004) closely relate this type of power with the way subsidiaries use lobbying for rent appropriation. To further this, to be able to win allocation of resources from headquarters, the managers of the subsidiaries have to exercise their voice through lobbying (Cantwell & Mudambi, 2005). To further develop this notion, Mudambi and Navarra (2004) elaborate that the worse the performance of the subsidiary, the more likely it is to engage in lobbying as opposed to better performing ones. This position is supported also by Ling et al. (2005) who put the emphasis on “issue-selling” by the top management of the subsidiary, which they define as the behaviour to direct the attention of the top management to a specific strategic issue. The subsidiary’s managers can only turn this “issue selling” into power in the bargaining context between parent and subsidiary (Belanger & Edwards, 2006). Though, it is deemed important, Dorrenbacher & Gammelgaard (2011) give two reasons why micro-political bargaining power is not as strong as some of the other subsidiary power archetypes- firstly this bargaining power can easily be negated by the bargaining skills of headquarters and secondly to have access to this power the presence of valuable assets is necessary. The second aspect can easily be offset by the subsidiary by having intangible assets (knowledge) and developing many subsidiary competencies (Mudambi & Navarra, 2004). Lastly, the strength of micro-political bargaining power is seen as small as negotiating skills are not a unique asset and usually other subsidiaries might possess them (Dorrenbacher & Gammelgaard, 2011).

As headquarters give different functions within the network of subsidiaries and the subsidiaries specialize in different parts of the value chain activities, the second type of power arises- systemic (Dorrenbacher & Gammelgaard, 2011). This type of power is characterized as one that arises from the subsidiary holding power over a specific function that is of critical importance for the functionality of the value chain (Dorrenbacher & Gammelgaard, 2011). As

(15)

[- 15 -]

the case with micro-political bargaining power the effect of this type of subsidiary power is not very strong as it only functions within the intra-subsidiary system and has problems with implementation outside of it (Forsgren et al., 2005). On the other side of the spectrum though, this specialization over critical functions of the value-chain might help the subsidiary as it arises costs for headquarters to move this function to another subsidiary (Dorrenbacher & Gammelgaard, 2011). Based on this assumption systemic power held by the subsidiary is likely to be strong in subsidiaries that have control and knowledge of critical parts of the value chain. The third and most used by scholars type of power is the resource-dependency power (Dorrenbacher & Gammelgaard, 2011). This concept is connected to the rise of business network field in International Business. Yamin & Forsgren (2006) argue that by being embedded in different environments headquarters cannot control all the local subsidiaries that are critical for the performance of the MNE. Consequently, rather than being viewed as a hierarchy the MNE should be seen as a structure comprised of different sub-units which are competing for power and rent appropriation (Andersson et al., 2007). Viewing the MNE like that and emphasizing on the role of the subsidiary as a competence-creating unit, headquarters start facing resource-dependence predicaments. This type of resource-dependence of headquarters highlights the subsidiary’s ability to cope with critical problems that arise in its environment which HQ cannot affect (Pfeffer and Salancik, 1978). The second type of resource-dependency of HQ comes from the ability of the subsidiary to build economic opportunities from its local environment (Dorrenbacher & Gammelgaard, 2011). Finally, the final type of resource-dependency power the subsidiary develops is the fact it can specialize and create specialized knowledge as well as expertise and its own technologies (Andersson et al., 2007). This type of power is also being thought of as a strong and sustainable one as power over critical resources is essential for the performance of the MNE (Dorrenbacher & Gammelgaard, 2011). In the context of this thesis and Big Data this type of power plays an

(16)

[- 16 -]

essential role as Big Data is large volume of knowledge, which can be leveraged for developing intangible assets. At the same time though, the accessibility of Big Data might lower the resource-dependency headquarters has of the subsidiary’s resources as it can generate the knowledge it needs of a foreign market in-house from the home country.

Geppert & Williams (2006) define the last type of subsidiary power- institutional as the ability of local subsidiaries to leverage on the institutional framework of the host country. Dorrenbacher & Gammelgaard (2011) consequently add that what separates resource-dependency power from institutional is the fact that institutional power does not require a deep embeddedness in the economy of the host country. Institutional power can be harvested by the subsidiary in different ways and is overall thought of as a strong power (Dorrenbacher & Gammelgaard, 2011). This statement is strengthened by Zaheer & Kostova (1999) who argue that a critical issue for MNEs is the creation of legitimacy in the host country, as well as how the laws if the rules in the host country are not respected different problems arise for the MNE. Dorrenbacher & Gammelgaard (2011) argue that this type of power can be used by subsidiaries operating in an environment where institutional distance is big.

2.3 Subsidiary Autonomy

Subsidiary autonomy is one of the outcomes that comes from increased power within the subsidiary. There have not been a lot of studies on the topic but from what has been researched there are mixed results as to whether subsidiary autonomy is linked to better performance of the subsidiary (Gammelgaard et al. 2010). To counter that though Birkinshaw and Morrison (1995) find out that a certain degree of autonomy does indeed lead to better performance. More specifically they talk that high and low degree of autonomy lead to better performance. Even though subsidiary high subsidiary autonomy may lead to better results it also leads to more rent-seeking and appropriation behaviour that destroys value (Foss et al. 2003). The definition we are going to use for this thesis comes from Gammelgaard et al. (2010) and it defines

(17)

[- 17 -]

subsidiary autonomy as rights for decision making granted from the parent company. Mudambi & Navarra (2004) argue that autonomy is the result of increased bargaining power. In the context of this thesis data is a source of knowledge in large volume, which in return would have an effect on subsidiary autonomy, because when a manager negotiates for autonomy in decision-making, having big data analytics behind him would have a different effect than when using intuition methods

In the previous paragraphs the concept of subsidiary power and autonomy was established as well as the four different main types of power a subsidiary can gain. The strength of those powers was also established through the literature review to help identify which type of power would be the most affected by the implementation of Big Data and Data Driven Decisions.

2.4 Big Data and Data Driven Decisions

In the last ten years, Big Data has become a phenomenon that has drawn a lot of attention in the IT sector, while its implications for management have remained underdeveloped (Brynjolfsson & McCafee, 2012). To begin talking about Big Data though, we first have to define it and see how literature has looked at this phenomenon through the lenses of the MNE. Different scholars have given different definitions for Big Data ranging from defined as data that can be extracted from electronic devices, to being defined as datasets too big to be analysed by current analytical technology (Wamba et al. 2015). Overall most scholars agree that Big Data is connected to the so-called concept of “Vs” (Wamba et al. 2015). Russom (2011) first implemented the definition of Big Data through the three Vs as he differentiated Big data from what was usually called analytics (McCafee & Brynjolfsson, 2012). The 3Vs are the Volume, Variety and Velocity (Russom, 2011). Volume is the primary attribute of big data (Russom, 2011) being defined as “large volume of data that either consume big storage or consist of large number of records”. However, he agrees that the volume of the datasets is the primary attribute,

(18)

[- 18 -]

(Russom, 2011) and emphasizes on the importance of data variety and velocity. Variety, meaning generating data from different sources and formats (Russom, 2011) and Velocity defined as the frequency of data generation or delivery (Russom, 2011). In the context of the subsidiary power, which is the main research object of this thesis, the Velocity of Big Data will be one of the more important aspects that will affect it. This is based on the assumption that Mudambi and Navarra (2004) come to- that the higher the knowledge transfer from subsidiary to headquarters the higher the power of the subsidiary. Higher velocity of knowledge generation and delivery would mean that the subsidiary will have a faster way to assert its subsidiary power to headquarters through lower times of transfer as well as an infinite source of that power. When applying Volume to the relationship between headquarters and subsidiary the large amounts of data, can be viewed as a source of knowledge. To further elaborate on this Sagiroglu & Sinanc (2013) say that big data has the ability to transform every business by bringing knowledge and insights to companies’ arsenal. In the context of subsidiaries, the volume of knowledge that data brings to their decision-making, can be seen as what Salanick & Pfeffer (1977) view as critical resources. Control of these resources by the subsidiary, is said to give it bargaining power (Pederssen & Mudambi, 2014) so combining these notions it can be seen how two of the initial 3Vs affect the parent subsidiary relationship.

Wamba et al. (2015) later improved the concept of the 3Vs adding Veracity and Value into the mix of Big Data attributes. Veracity connects with the need to analyse big data to get a reliable prediction (Wamba et al. 2015) while Value defines the way in which the analysis of big data would lead to benefits for the company. Big Data itself has according to IDC (2013) three characteristics that define it- the data and its size, the analysis of the data and the way it presents a result, so the company can take action to achieve value creation. In this thesis the data itself will play a big role as it is a source for subsidiary power, being a huge knowledge resource, to subsidiary managers. This assumptions stem from a few definitions of Big Data given by

(19)

[- 19 -]

different scholars. Davenport et al. (2012) define Big Data as data that could be gained from absolutely everything varying from simple web clicks to medicine. Consequently Jacobs (2009) defines it as data that is too large to be placed in a database and has to be stored on hundreds of servers. All these definitions show how grand Big Data is in the context of the corporate world.

After establishing what makes Big Data “big” in the next paragraph, I will make a review of the managerial implications of Big Data and introduce the concept of Data Driven Decisions. In recent years the adoption of Data Driven Decisions (DDD) has more than tripled (Brynjolfsson & McElheran, 2016) which can be explained with the performance of firms that rely on Big Data which is higher than the performance of companies that still rely on intuition decision making or what Brynjolfsson & McCafee (2012) call the “highest-paid person opinion” (HiPPO). On average companies that implement big data into their decision-making are 6% more profitable (Brynjolfsson & McCafee, 2012) which can explain why the adoption of big data analytics has grown exponentially since 2010. Brynjolfsson, Hitt & Kim (2013) further elaborate that information used in making increases the ability of the decision-maker to identify the best possible outcomes by eliminating statistical noise.

However, even if the number of companies implementing big data into their management is rising, intuition is still the dominant way of making decisions in a company, which can be attributed to the cost of implementing big data and the time it takes to spread (Brynjolfsson & McElheran, 2016).

As a conclusion, big data and the data data driven decisions that come out of it, have proven to be the way of the future and scholars as well as companies have started to emphasize on their role in decision-making. Various authors have given their insights on why companies adopt data and how it affects managerial decisions and firm performance. Based on literature it can

(20)

[- 20 -]

be seen that, most scholar specialize in noting only the positive effects data has on a company- profitability, predicting and developing specialized products based on big data analysis and the stimulation big data has on innovation. But looking at it from the perspective of MNE overall profitability, the internal struggle between various subsidiaries for influence (Birkinshaw, 1996; Mudambi & Navarra, 2004) is left as a topic that is under developed. As

That leaves a clear gap as to how data driven decisions affect the power and autonomy of subsidiaries within MNEs.

After defining Big Data and Data Driven Decisions and the different context it has been used in, a clear gap in literature cane be seen. One of the fields international management has been moving to in the last 50 years is the increasing importance of the subsidiary in the parent-subsidiary relationship (Kostova et al., 2016). In the specific context of this increasing role of the subsidiary, scholars have yet to research how data driven decisions affect it and its power. Data has been. After conceptualizing all the variables in the next paragraphs a research question will be established and a theoretical framework will be built.

2.5 Research question

The purpose of this paper will be to examine the parent-subsidiary relationship, and how big data affects the subsidiary’s power. Brynjolfsson et al. (2011) note that data driven decisions give a performance to companies as they increase ROI and profitability on average. When applied to the level of subsidiaries cases in the past have proven that the better a subsidiary performs the more it can influence its headquarters. Kotter (1979) says that the higher an executive is regarded at an organization the more power they obtain at their disposal. To this

(21)

[- 21 -]

Boyatzis et al. (2006) add that this power can then be used to influence other people within the corporation. Based on this and the review of literature review done in the previous paragraphs a gap in literature presents itself on what are the implications of data driven decisions on the subsidiaries of MNEs. A subtopic of this is how the phenomena of data driven decisions can be used in the context of subsidiary power in negotiating with headquarters, and the ability of the subsidiary to be autonomous in their own decision making. From this we have come up with the research question of this thesis:

What is the effect of Data Driven Decisions made by subsidiary managers, have on the subsidiary’s power?

2.6

Working

propositions

Based on literature, Big Data has become a driving force for decision-making and a source of power for companies, resulting in better performance on average (Brynjolfsson & McCafee, 2012). In the context of subsidiary power, decisions based on Data can be seen as a way to accelerate the speed at which subsidiaries gain power. As the velocity of generating and integrating knowledge is faster than before, the subsidiary will have the backing of data when trying to leverage their intangible knowledge. Forsgren & Pedersen (2000) argue that the greater the competences of the subsidiary the greater its strengthened position within the MNC. Mudambi & Navarra (2004) add to this the notion that the more control over knowledge assets of the MNC a subsidiary has, the higher the bargaining power exercised by this subsidiary. Jaffe et al. (1993) argue that knowledge is highly localized, but in the context of data driven decisions and big data, this does not apply. This notion is developed by going through two of the 5Vs of Big Data-Volume and Velocity, which means it is highly accessible and comes from everything, it is generated from varying sources (Russom, 2011). At the same time even though

(22)

[- 22 -]

it is accessible, human capabilities to analyse it are argued to be needed to generate stronger analysis and knowledge (Beulke, 2011). When all this is done, this knowledge can be used in decision-making to make data driven decisions which will affect the bargaining power of subsidiaries. More knowledge, more power (Mudambi & Navarra, 2004), and decisions based on Big Data use a large volume of knowledge (Wamba et al. 2015). Based on this, the first working proposition of this thesis is:

WP1: Data Driven Decisions are expected to give subsidiary managers higher bargaining power

Brynjolfsson & McCafee (2012) bring forward the importance of data driven decisions for company performance when compared to intuition-based decisions. Big Data has Volume, Velocity and Veracity (White, 2012), which gives managers, more knowledge to work with, makes them get it faster, and gives them more means of analysing it to reach insights. Since the MNC is a field of competing interests between headquarters and subsidiary (Mudambi & Navarra, 2004) one of the most important things that subsidiary managers bargain for using their resources is rent appropriation. Blyler & Coff (2003) support this thought by saying that rent appropriation is determined by bargaining power. Data is evidence, and data driven decisions are on the basis of evidence rather than intuition or experience (Brynjolfsson & McCafee, 2012). When we apply all this to the subsidiary level, it can be hypothesized that the continuous integration of data in decision making and when bargaining with headquarters for budget allocation will give managers more credibility as they will have evidence-based insights to support why they would want a certain sum for their subsidiary or unit. This in turn would make the process of budget allocation approval shorter. Based on all those things the third working proposition of this thesis is:

(23)

[- 23 -]

W2: Subsidiary units that use Data Driven Decisions go through a shorter approval process for budget allocation.

Gammelgaard et al. (2010) define subsidiary autonomy as the rights of decision-making given from the parent to the subsidiary. Consequently, Mudambi & Navarra (2004) say that autonomy is the result of two different approaches. Firstly, of high bargaining power. This subsidiary autonomy comes from the fact that the subsidiary has bargained itself a certain degree of ownership of its own decision-making, rather than having everything given by headquarters. The second approach is the one Foss & Foss (2002) talk about which is delegated decision making ownership. This type of autonomous decision-making rights mean that headquarters has the power of veto and the power to overturn autonomous decisions made by the subsidiary- decision rights are “loaned not owned” (Mudambi & Navarra, 2004). Since data driven decisions and big data have an impact on subsidiary performance (Brynjolfsson & McCafee, 2012), and give managers large quantities of knowledge that are generated fast, it can be argued that data driven decisions will impact the subsidiary bargaining power. Not only that, but since data is evidence it can be used to persuade headquarters that a certain decision is optimal. Based on that we argue that, data driven decisions will affect the delegated rights of decision making given by headquarters, as subsidiaries that are data driven would perform better. Going further, these delegated rights will become higher as the subsidiary would prove itself to be able to perform on its own based not on intuition, but on evidence-based insights. Coming from this last point the third working proposition is:

W3: Data Driven Decisions give subsidiary units a high degree of subsidiary autonomy.

Lastly, Dorrenbacher & Gammelgaard (2006) state that subsidiary power arises from headquarters being resource-dependent on the subsidiary. This is also strengthened by

(24)

[- 24 -]

Pederssen & Mudambi (2014) who elaborate that resource dependency is traced back to the control of resources, that lead to hieararchical power of the subsidiary. We take a look at the resource dependency power which has three variations by being able to tackle critical issues in its environment, leveraging the creation of own capabilities and expertise, and lastly create economic opportunities in its environment. The second type, creation of own capabilities and expertise, arises from the development of specialized knowledge and expertise that are later used throughout the whole MNE (Andersson et al. 2007 in Dorrenbacher & Gammelgaard, 2006). In the context of the research on data driven decisions and big data, the volume of data that flows within the MNE and its implementation, we argue, would allow the subsidiary to have higher resource-dependency power, than if intuition-based decisions are used. Going further, the velocity of data generation, that is then analysed and used, would help reach this stage of resource dependency by headquarters faster. Not only that but since data is available to everyone, the type of resource-dependency power will be different as it will not be specific to the location knowledge but rather something universally applicable. Rather than the business context, the Veracity dimension (Beulke, 2011) that the subsidiary has control over, would be the driving factor of resource-dependency power. Based on this the last working proposition is created:

W4: Data Driven Subsidiary Units have high resource-dependency power.

Based on the research question that was established as well as the different working proposition elaborated on in the last few paragraphs, a theoretical framework has been established:

Data Driven Decisions Subsidiary power Bargaining Power

Subsidiary Autonomy Resource Dependency Power

(25)

[- 25 -]

2.7 Expected contribution

The expected contribution this paper has is that it will bridge the gap between two highly important for modern international business theory fields- Big Data and parent-subsidiary relationship, as well finding a connection between Data Driven Decisions and the subsidiary power concept.

3. Methodology

In this segment of the thesis the methodology will be discussed. More specifically an overview on the strategy, the research design, the data collection and analysis will be discussed in detail.

3.1 Research strategy

In order to answer the question posed inside this thesis a research strategy has to be established. The main purpose of this thesis is to bring forward research on an interesting topic, that is relevant in the 21st Century which is- “How Data Driven Decisions affect subsidiary power and autonomy”. What this research questions want to bring light on is how the implementation of data driven decision making by subsidiary managers when negotiating with headquarters has affected their power- whether the gives them more power to negotiate, or maybe it gives them resource dependency based on the knowledge data brings forward. Having data behind a potential decision, might make it easier for headquarters to allow a subsidiary project to go into development or give them more influence over headquarters themselves, as managers inside

(26)

[- 26 -]

the subsidiary will be deemed competent to analyse data and just give the insights to the managers from corporate HQ. Another more in-depth proposition is that by proving themselves to be data driven, subsidiaries will get more autonomy over their own decision-making, as companies that implement data driven decisions are said to perform better (Brynjolfsson & McElheran, 2016), which in turn might lead to autonomy. The overall purpose is to explore whether that is indeed the case, so a qualitative research is the best way to do that. Miles, Huberman & Saldana (2003) state that in order to tackle complex themes a qualitative research is the best way to go. On top of that the theme of data driven decisions in the context of the subsidiary has never been researched before so qualitative research is applicable as a means of answering the research questions. Lastly, Pratt (2009) articulates that when trying we are trying to answer a “how” or “what” question rather than a “how many” one qualitative research is a very good way of addressing the issue. Based on everything written above a qualitative research method proves to be the best way to tackle the research question.

3.2 Research Design

In creating the research design, a critical factor for that is to align the research question of the research with the research strategy and collection methods (Saunders, Lewis and Thornhill, 2009). While there are a lot of ways in which we can perform a research, the type of study we are performing is the primary distinction that should be made in the first stages of research design. There are three types of purposes for research according to Yin (2013)- exploratory, explanatory and descriptive. Yin (2003) argues that a good approach to conducting exploratory studies is by using a case study. Consequently, he states that when trying to answer a “what” question which tries to develop propositions for further research, an exploratory case study is a justifiable rationale (Yin, 2003). To further elaborate on the choice of a case study for this research, one of the applications of it is in a case where we are exploring situations in which the intervention being evaluated has no clear outcomes (Yin, 2003). Based on the fact that the

(27)

[- 27 -]

Parent-Subsidiary relationship has not been looked at through the concept of DDD, a qualitative approach to this paper is expected to yield the best results (Basal & Corley, 2012).

This is further strengthened by Robson (2002, p.178) who defines a case study as “the empirical investigation of a particular phenomenon within its real-life context”. Lastly, since the purpose of this paper is to uncover more rich insights from the data collection a qualitative research is perfect to get that much more detailed information (Gephart, 2004).

On top of using case study as a research strategy, multiple cases will be analyzed in this research. Eisenhardt (1989) notes that each case is synonymous to one experiment, and that in replication, cases that confirm the relationships bring forward more confidence in the validity of those said relationship. This is further validated by Scotland (2012) who states that the multiple case study approach gives the researcher the power to see differences and similarities between and within cases. This study explores how data driven decisions from subsidiary managers affect their power and autonomy. Since this is an exploratory study and companies have a different degree of data driven decisions implementation, work in different sectors and have different business contexts a multiple case study can take a look at that and find similarities and differences in their approach to DDD. Using a multiple case study also makes a theory easier to be generalized, also noted by Yin (2003) to be external validity. Soy (2015) brings further evidence for this notion, as he elaborates that the bigger variety of places and people a case study can handle and has the same results the more external validity there is. Semi-structured interviews will be used in order to get the necessary information from. The cases I selected are from two different subsidiaries of big MNEs based in two different countries- Bulgaria & The Netherlands.

In the next paragraph the quality criteria that is meant to ensure the credibility of the findings of this thesis will be given an overview.

(28)

[- 28 -]

3.3 Quality Criteria

For a study to be credible, it is needed to have certain quality criteria. It is a researcher’s duty in such a case to ensure that his decisions while writing the research paper are justified and credible. There are four types of quality tests to justify the choices a researcher makes- external and internal validity, reliability and construct validity. In this part of the thesis an overview of them will be given.

The first test Yin (2003) as well as Saunders & Lewis (2012) outline is construct validity, which they define as accuracy of measuring for the intended purpose. To stand up to this criterion, a research paper needs data triangulation as it helps reach the intended results (Gibbert and Ruigrok, 2010). In order to do that in the context of this thesis, semi-structured interviews with managers from subsidiaries located in different countries were conducted. This gives the data different perspectives of subsidiary managers operating in different business contexts. In addition, since the main purpose of this thesis is to understand if there is a relationship between data driven decisions and subsidiary power, the subsidiary managers picked were ones that had a record of being data driven and from a well performing subsidiary.

The second test is internal validity which tests the power of the relationship between results and variables (Yin, 2003). This part of the test outlined by Yin (2003) helps the researcher take a look at the explanations of relations in the study. The way to have internal validity is to compare the findings with previous literature. In this study this is done by grounding the working propositions in previously developed theories and afterwards testing and comparing them with the results of the data analyzed.

Third comes external validity which Calder et al. (1982) define as “the process of identifying whether an observed causal relationship should be generalized across different measures, persons, settings and times”. Gibbert et al. (2008) further this thinking by saying that in order to assess the external validity of a research, the case selection has to be explained as thoroughly

(29)

[- 29 -]

as possible. In this thesis, the case selection will be given in the next theme thoroughly. One of the weaknesses that might come up using interviews is whether 6 interviews from only two companies would be enough for a theory to be generalizable. In this case the subsidiaries operate in different environments but use data in decision making in the same way. This would prove to bring external validity. Additionally, the interviews and the off the record conversations I had with the subsidiary managers, brought similar results across the six interviews, and no new knowledge was gained after the 5th one.

Lastly, the fourth test defined by Yin (2003) is reliability, which is defined as the repetition of the same data collection methods that would yield the same results if done by other researchers. Gibbert et al. (2008) go more in-depth by elaborating that the most important aspects of reliability are replication and transparency. In the case of this thesis, all interview protocols as well as the transcribed interviews are given as appendixes, which means there is full transparency on the type of questions asked as well as the answers of the interviewees. One problem that might occur when trying to replicate the results of this study would be the use of semi-structured interviews to gain the data necessary, as different researchers might ask different supporting questions in order to get more insights on the topic.

After establishing the quality criteria in the next part an overview on the subsidiaries used in this research will be given as well as justification on why they were chosen for this research.

4. Case Selection: IBM Bulgaria & Elsevier Amsterdam

As noted in the research question part, the main objective of this research is the effect data driven decisions of subsidiary managers have on their subsidiary power and autonomy. In order for this research to be realized, narrowing down the list of potential sources of information was needed. When looking at subsidiary power, two points of view can be seen- the headquarters

(30)

[- 30 -]

one, and the subsidiary one. The literature review established that the role of the subsidiary as an extension of headquarters inside foreign markets has become increasingly more popular in IM literature (Kostova et al. 2016). This is why I take the subsidiary approach in this research. In order to analyze the state of subsidiary power in the age of big data and data driven decisions two main points were the criteria for choosing samples:

Firstly, it had to be a subsidiary that has access to a lot of data, both analytics and big data, and has a record of using it in decision-making. For this purpose, I chose two companies that have been transforming adequately into the digital age - IBM Bulgaria, a subsidiary of IBM, and Elsevier Amsterdam, a subsidiary of RELX Group LTD.

Secondly, the managers that were going to be interviewed had to be data driven in their decision-making for which I used the so-called snowball sampling, which is used to get to hard to reach groups (Browne, 2005). Existing contacts from my network were asked to nominate and help me reach managers with data driven reputation across the company. As of the moment of the interviews in Elsevier Amsterdam, I was doing an internship there, which helped me build the network necessary to reach high standing managers that are known to be data driven decision-makers.

In the case of IBM Bulgaria EOOD, I used again the same approach with building a network of people who would nominate managers that use data driven decision-making and would be willing to share sensible information about the empowerment effect it might have.

IBM was chosen because has a long history of being a pioneer in innovation and the implementation of new approaches to business and technology. As of 2017, it was considered the most innovative company in the USA by patents granted at 9043 (Statista, 2018). The sectors the chosen managers worked in were financial product development and financial analysis. The two sectors were chosen as finance is a field where data towards managers is

(31)

[- 31 -]

flowing without a break and it is a field that has additional layers of depth that benefit from specialization.

Lastly, when choosing the cases for the research the most important criterion was that the manager being interviewed would be under the supervision of another manager from headquarters with whom they communicate frequently, and exchange insights into decision-making. This proves to be critical about the success of this research as a manager who does not negotiate with headquarters would not bring any knowledge of value in the context of this thesis about subsidiary power.

To summarize, in order to bring high quality data on the topic of subsidiary power and autonomy, three criteria were used to choose interviewees. First, the managers had to be ones from subsidiaries, second, they had to be data driven in their decision making and third, managers have to work under the direct supervision of a higher manager from headquarters in order to have a subsidiary perspective on the influence data driven decisions might have on them. With all that established, the data collection approach will be explained in the next part.

4.1 Data collection

4.1.1 Semi structured interviews

According to Harrell & Bradley (2009), when a researcher is trying to understand a topic deeply and to understand fully the answers that are provided, a semi-structured interview is the best way to do that. The questions for a semi-structured interview are standardized and the researcher chooses the order, but rather than collecting knowledge based on preset answers,

(32)

[- 32 -]

this type of interview gives the freedom of having more of a conversation with the interviewee (Harrell & Bradley, 2009).

In total, six interviews were conducted with six members of senior management. Four of the interviewees were from Elsevier Amsterdam, a subsidiary of RELX group, while the remaining two were done with managers from IBM Bulgaria EOOD, a subsidiary of IBM. The interviews were conducted between March 2018 and June 2018, ranging in length from 31 minutes and 43 seconds to 43 minutes and 13 seconds, depending on the amount of information the manager was willing to share about the research question. Four interviews were conducted in Amsterdam at Elsevier’s office there, while the other two interviews were conducted in Sofia, Bulgaria with IBM Financial Center managers.

All interviews were conducted face to face, in English with non-native speakers, as it was preferable for them to use English. The questions asked were by templates, with additional questions asked by the interviewer where deemed necessary, based on previous answers to the scripted questions. This is in line with the flexibility semi-structured interviews bring, as scripted questions help steer the discussion on the topic, without going out of theme.

All of the interviews were done with managers of subsidiaries that communicate on a regular basis and negotiate with managers from headquarters about subsidiary decisions, budget allocation, exchanging knowledge about customers. On top of that, all the managers had described themselves and had been described by other employees as data driven, and using data driven decisions and analytics in their decision-making process. The interviewees asked that their names would not be disclosed so only their position inside the company as well as their experience will be provided in table 1. For the sake of transparency, it should also be noted that the researcher for this thesis had worked in Elsevier for eight months as a Business analyst intern and has worked for the last two months at IBM Bulgaria as a financial analyst.

(33)

[- 33 -]

For this reason, where projects have been discussed during the interviews, in the transcriptions they have been blurred out and have been changed with codes.

In the next table a brief description of the interviewed managers will be given:

Table 1: Participant Affiliation Overview

Participants Company Position Experience

P1 Elsevier Software Engineering Manager 7 years Manager P2 Elsevier Head of Data Science & Director of

Content

9 years Manager P3 Elsevier Product Manager Funding Solutions &

Team Lead

5 Years Manager P4 Elsevier VP Product Management 20 years

Manager P5 IBM Manager Global Technologies Services

MEA

7 Years Manager P6 IBM Manager Global Technologies Services

SPGI

13 Years Manager

4.2 Data Analysis

In order to answer the research question of this topic, it is critical to use a proper data analysis strategy, that would ease the filtering of relevant information that was gathered in the process

(34)

[- 34 -]

of data collection. For the research at hand, six semi-structured interviews were conducted among senior to top management of subsidiaries in different countries, working in different sectors as well in order to have triangulation. The positions the managers interviewed were at range from- Market Financial Managers at IBM to Vice president and Directors at Elsevier. The criteria for the sample was that the managers have to be in constant contact with headquarters’ managers and negotiate with them. By picking that criteria, it was assured that the managers took part in the process of subsidiary bargaining power and had first row seats in the process of negotiating.

The approach to data analysis used in this thesis is the one outlined by Ritchie & Spencer as the “Framework Analysis”. The process is divided into five stages- familiarization, identifying a thematic framework, indexing, charting, mapping and interpretation (Rabiee, 2004). The first stage encompasses the data collection and generating of rich data from interviews, as well as familiarizing yourself with the data by listening to the recordings and reading the transcripts. Secondly, comes the identifying of thematic framework. This is done by writing the concepts that arise from the statements in the interview, outlining the general topics (Rabiee, 2004) The third stage in the framework analysis is indexing, which means sifting through the data and highlighting quotes and looking for relationships within and between cases (Rabiee, 2004). After doing that, we compiled the relevant quotes for each working proposition within the case and between cases. Based on that the results will be given in the next chapter of the thesis.

5. Results

This chapter of the research will present the results of the studies that were conducted with the six managers from IBM Bulgaria & Elsevier that were introduced in the previous paragraphs. The results will be structured by working proposition to show a clear picture of the outcome of

(35)

[- 35 -]

the data analysis. Firstly, the results from the interviews with Elsevier will be given and afterwards with IBM, consequently a cross case study analysis will be done to find similarities between answers from managers at different subsidiaries.

5.1 IBM Case findings

IBM Bulgaria EOOD is a subsidiary of IBM, that was reopened in 1994. The financial center in Sofia for IBM is one of the two biggest finance hubs in Europe, growing from 75 people in 2013 when it opened to above 300 in 2018. The managers interviewed for this thesis are market managers within the subsidiary for Middle East and Africa (Participant 5), and SPGI (Spain, Portugal, Greece and Israel); (Participant 6). These managers were chosen as they heavily implement data in their decision making, have a proven track record within the company of using analytics, big data and being agile. Going further, they are a direct line between the subsidiary and the CFOs responsible for those markets from headquarters, which as they elaborated means they negotiate with the CFOs about different projects and decision making. This makes them perfect candidates as they have a real-life picture of how data in their decision making and negotiation with headquarters affects their respective teams within the Bulgarian subsidiary.

5.1.1. Bargaining Power

On the topic of subsidiary bargaining power when the market managers were asked how big data and data driven decisions affected the subsidiaries bargaining power, both P5 and P6 confirmed that Data Driven Decisions and Big Data analytics improve the negotiation process with headquarters by bringing proof of the potential decision for the subsidiary or the result of getting a certain budget will have on overall performance. P5 elaborated on the role of the Sofia Subsidiary and how subsidiaries at IBM function overall and where data fits into the negotiation process overall. Later on, when asked about bargaining power and how DDD and

(36)

[- 36 -]

Big Data affect it he stated: “When it comes to the Sofia subsidiary and its management we are

trying to use data in order to influence headquarters, which means if we want to increase our capacity here we need to show factual evidence and data, why they should allow us to do so”.

Consequently, when asked on what type of data they use when negotiating he provides the type of analytics that are used when negotiating with headquarters for more responsibilities or rents. Continuing from that he states as well that: “What I am saying is that proof is really important

when negotiating, and if you bring these statements to headquarters, along with the analysis and insights it becomes a lot easier to negotiate”. These statements are further proved more

than a singular case by P6 who stated that Data and analytics help simplify the process of negotiating and again bringing more proof to the negotiation table: You have the data, make

the analysis, present it and when you do that in an appealing manner, it gets much more easily accepted from the CFO, and the credibility and trust goes to a higher level.” These statements are all in line with Surlemont (1998) and Cantwell & Mudambi (2005) who say that bargaining power is the exercising of voice by subsidiaries through lobbying. From the statements given by the managers, data in potential decisions affects the acceptability of their proposals, making it higher. This is further strengthened by statements by P6 about using anything other than data: “If I would only go and say this is my intuition and I think that things will simply work like this

because I have a very good feeling I would not get anything out of this meeting”. The statement

by P6 is also supported by the data gathered from P5 who elaborated that: “But when it comes

to decisions that we try to push to headquarters, intuition is not applicable, experience has little moderation, but data gives us the easiest way to make our point.” Both P5 and P6

provided information that supports the proposition that data driven decisions give subsidiary units more bargaining power. The reasons they provided for this is that it increases trust by providing factual evidence as well as that it shows you are not basing decisions on your gut. It

(37)

[- 37 -]

should be noted that both managers state that you would not even get a reaction from HQ managers if you are not backed by data.

5.1.2 Budget Allocation

When asked about budget allocation from headquarters both managers noted that the processes in IBM about budget allocation are not really flexible, as there is a framework for budget and financial resources allocation, which operates in a certain window. But they also noted in critical situations where allocation is critically needed, having data and analytics to back up what they ask for is essential. P6 when asked about budget allocation elaborated that: “This is

a quite restricted process in IBM right now, to have any kind of control on spending related to projects and hiring. So, unless there is a very good reason backed by evidence, which data is and a good plan based on metrics, that shows how this decision to give you budget will turn out to be long-term productive you will not get any additional budget allocation.” On top of

saying that even though the process is restricted, having data and analytics to back you up can help you get more budget allocation, she also elaborated that the process itself is easier to approve when data and analytics are involved. She stated that when going to headquarters to negotiate for budget allocation if metrics are shown on how this allocation would improve performance of the MNE and revenue the chances of getting it approved are high. Also, she continues to state that if you try to negotiate using anything but data there will be no approval at all, since you have no forecast or prediction on the value this allocation will bring. Statements by P5 support this stance as well, but he notes having data insights does not by default give you more budget allocation but: “you only having data does not make you get it, but data rather

gives you the weapon to argue why this budget allocation is necessary.” He also elaborates on

the fact that if you try to negotiate using anything other than data the result will not be positive. P5 states the following about using other methods to negotiate with headquarters for budget

Subsidiary power in the age of data driven decisions : an exploratory study