Marketing Automation: Considerations and Choices

1.0 Executive Summary

This paper is a study in how large enterprises can ease the implementation and utilize the untapped potential of marketing automation. The purpose of this is to give management groups from other enterprises a tool box they can use to streamline the complex, lengthy and costly process of implementing marketing automation for successful adoption in their context. By conducting an exploratory single case study in Microsoft, we want to look at which strategic considerations and choices were prevalent for their successful implementation of marketing automation in Denmark, and to what extent these can be applied to other large corporations wanting to implement marketing automation.

 

After conducting a literature review we identified a gap in the literature, which worked as an offset for our problem formulation and research. In line with interpretivism as our paradigmatic base assumption, we began our research in Microsoft. Firstly, we conducted observations and informant interviews from which we gained an initial understanding of the context and the phenomenon. This did not only prepare us for writing our interview guide, it also clearly strengthened the quality of semi-structured interviews, which we did with both employees at management level as well as employees on a more executional level.

 

After revising and summarising the interviews, we had our empirical data ready for processing. From the summaries we extracted the most important strategic considerations and choices were prevalent for Microsoft’s successful implementation of marketing automation. This resulted in five very concrete considerations and choices identified which worked as our recommendations for other enterprises. Those recommendations were:

 

1. Prioritization
2. Knowledge sharing
3. Content creation
4. Change Management
5. Organizational structure

 

In line with an exploratory and inductive approach, this study have created new theory which now can be used for future research. With the foundation in our research, it is encouraged to conduct a more deductive and confirmatory study in other enterprises, which would solidifying the reliability of the results.

 

2.0 Introduction

 

Today a lot of companies are trying to keep up with the latest IT trends in an attempt to stay innovative and grow their business with the aid of technology. As new technologies constantly evolve, and firms have limited resources, it is important to aid implementation and adoption of those technologies.

Companies perpetually look to upgrade their marketing efforts with the help of IT, aiming to increase effectiveness while decreasing costs. Marketing automation is frequently being mentioned in this context, as automation of any process can help facilitating increased effectiveness and/or decreased costs. For some companies, marketing automation is a fully integrated part of their marketing strategy, while others still lack behind as marketing automation might just have caught their attention. It is evident that marketing automation, for those who have already started using it, has becoming an increasingly lucrative way for many firms to generate solid leads (Watson, 2015). Especially large enterprises are benefiting from automating their marketing efforts, as they can substitute much of their sales force and thereby bring down labour cost.

In line with this technological advancement, we see a lot of potential in using marketing automation as a way to automate manual operations. Furthermore, with the concept of artificial intelligence looming in the near future, we see a lot of impending possibilities of combining those  with marketing efforts and building upon an already established platform for automation of marketing efforts. Though, it all sounds very promising, implementation of marketing automation can be challenging since it requires restructuring of numerous processes, organizational changes, and integration within the existing marketing setup (Chae et al., 2014). We are puzzled by how the large corporations that successfully utilize the potential of marketing automation, have implemented this innovative technology, and which considerations and prioritizations they have made. With this paper we want to give management groups from other large firms a tool box to streamline the complex, lengthy and costly process of implementing marketing automation for successful adoption in large enterprises. This includes some very tangible guidelines as well as recommendations on what to be aware of before, during and after the implementation process.

By conducting a qualitative single case study in Microsoft Denmark, we want to find an answer to the following research question:

 

Which strategic considerations and choices were prevalent for Microsoft’s successful implementation of marketing automation in Denmark, and to what extent can these be applied to other large corporations wanting to implement marketing automation?

 

As mentioned above, artificial intelligence offers many interesting opportunities for increasing the level of intelligence in marketing technologies, but to limit our scope, we have chosen to exclude perspectives on artificial intelligence from our research. With this paper we solely want to determine the factors influencing successful first time implementation of a marketing automation solution. The reasons why we have chosen to conduct our research in a large enterprise, compared to an SME, would be that they, first of all, have a larger existing setup to integrate with, which increases the complexity of the implementation and, second of all, normally have a much larger amount of data they can leverage from and thereby fully derive advantages from marketing automation (Wilson, 2015).

2.1 Case Description

Microsoft Denmark is in the process of implementing a marketing automation system, using the platform called Marketo, and rolling it out as a pilot of their Azure product line.

Azure is a product focusing on the developer ecosystem and secure cloud server hosting and is highly business to business oriented (“What is Azure”, 2016).

The overarching strategy and decision is made by the parent organization called Microsoft Corp, who define the strategic outset for each year on a global scale. It is then up the the local subsidiaries, hereunder Microsoft Denmark, to adopt and implement in the most beneficial way, where they have autonomy is within the tactical sphere. Thus, Microsoft Corp sets the KPIs (Key Performance Indicators) and determines some factors of the overall marketing campaigns and communications, but largely – the tactical aspects are executed on the subsidiary level. Because the subsidiaries are much more attuned to their local markets as relationship management with partners and clients is of paramount importance.

Microsoft Denmark, being a small market within the EMEA (Europe, Middle-East & Africa) area of the Microsoft Corporation, is also managed by a sub-division of Microsoft Corp called the Western Europe department, which oversees adoption of Microsoft Corp’s overall strategies for a larger group of smaller European subsidiaries. However, this research will focus on the Danish subsidiary and their efforts to maximize the potential of marketing automation. The following research will seek to map best practises of implementing marketing automation in a large corporation, through qualitative interviews with the people working directly with this task in Microsoft Denmark.

3.0 Theory

3.1 Interpretivism

The purpose of this research is to determine the different strategic factors that influence successful implementation of marketing automation. As user adoption very often comes down to opinions and actions of different social actors, we believe that understanding this phenomenon “…is far too complex to lend itself to theorising by definite ‘laws’” (Saunders et al, 2009, p. 115). We also believe that business situations like this are very complex and so unique that they have to be studied by entering ”…the social world of our research subjects and understand their world from their point of view” (Saunders et al., 2009, p. 116).

In order to gain an understanding of the socially constructed reality which lies within Microsoft, we have chosen interpretivism as our paradigmatic base assumption. We believe that the answer to our research question can be found by asking the social actors within Microsoft’s own context (Saunders et al., 2009). Figure 1. “The Research Onion” shows how research philosophies are the outermost layer, and therefore underpin the rest of the research design (Saunders et al., 2009). The alignment between philosophy and research design will be further described and elaborated in section 4.1 Research Design. However, before strategic considerations and choices can be identified, it is necessary to first understand the employees’ views on this specific phenomenon (Saunders et al., 2009). We believe that conducting research in line with the interpretivist paradigm, will give us the necessary depth to map the strategic considerations and choices prior to the successful implementation of marketing automation. As interpretivism acknowledges that there are multiple realities and that these may change over time, we will try to provide a snapshot of the current state of Microsoft Denmark (Saunders et al., 2009). This snapshot will describe Microsoft Denmark as a social context where reality is determined by the intersubjective consensus between the social actors.

 

From an axiological point-of-view our research is value-bound. We are operating in a sphere based in the subjective/regulatory realm, thus, our main goal is to extract knowledge to strengthen our understanding of the implementation process. It could be argued that our value-bound approach will put us, as researchers, too close to the research subjects to attain credible data, however, we view this as one of the advantages of our research philosophy. Our proximity will put us in a position where we are more flexible and able to adapt to surprising information, opening up for alternative explanations for the phenomenon (Saunders et al, 2009).  

Figure 1. “The research Onion”

In short, we believe, that by studying the social actors in their context, our understanding of the strategic considerations and choices that have influenced the successful implementation of marketing automation in Microsoft will deepen.

 

3.2 Literature Review

For this research project we utilized a review of previously written literature on the topic of choices and considerations behind adoption of marketing automation, with two main goals in mind. Firstly, we wanted to make certain that the topic had not been covered extensively already, thus, making our research redundant. Secondly, to sharpen the scope of our own research by drawing upon studies already conducted in the field and utilizing them as a theoretical framework.

 

We started by looking at the major contributions from recognized IS journals and using Google Scholar to see if other relevant articles might have been overlooked. We quickly realized that making a complete review of all the previously written literature was almost impossible, and there are numerous other challenges associated with constructing a literature review that we would have to accept. Firstly, our search criteria themselves were tainting objectivity from the outset, as we would never encounter articles that could be relevant to our research topic if they did not align with our search criteria. Secondly, our research is based on theory development within a very narrow segment of the intersection between, business/IT alignment, IT portfolio management, automation of marketing processes and business strategy, thus, our review would have to encompass literature in IS, business strategy, psychology, management and more, which added to the difficulty of maintaining the integrity and thoroughness of the literature review.

 

To help future researchers understand our thought process and facilitate constructive criticism of our review, we have decided to include most of our conducted searches to increase the transparency of the research. Our search started with a basic “implementation of marketing automation”, where result were very heavy on the technical aspects of marketing automation and the implementation into an enterprise architecture, the difference between other providers of marketing automation and to certain extend also user-adoption of marketing automation, most of which was beyond the scope of our research. As our searches widened and incorporated search words such as, “IT strategy”, “adoption” and “performance”, we began to notice several articles relevant to our research topic, however, none that had conducted an inductive, qualitative study of the decision-making process behind successful implementation of marketing automation in a large company. The most relevant articles are reviewed below, sorted based on the article’s main themes and concepts, with references to the articles at the end of the paper.      

 

3.2.1 Concept Matrix

Authors	Concepts
	IT Business Value	Knowledge Creation	Organizational Performance	Marketing Execution	IT in Marketing
Biegel				X	X
Hair Jr.		X			X
Chae et al.	X		X
Melville et al.	X		X
Wierenga & Oude Ophuis	X	X			X
Xu et al.	X			X	X
Chang et al.			X		X

We have identified five key concepts in the previous literature that relates to our research topic, stated in the concept matrix above and further elaborated on the below.

3.2.1.1 IT Business Value

IT business value revolves around the perceived business value gained from investments in IT. The value is notoriously difficult to measure using classic business parameters such as ROI, as investments in information systems can optimize processes without the effects showing up in the next quarterly. Investments in IT infrastructure are often long term investments where the direct business value may be obscured in the near future after implementation. They usually require extensive training of employees in order to utilize the investments to the fullest and will always carry the risk of cultural opposition within the sphere they are meant to optimize. An example of this would be process-optimizing IT investments, where the end goal is to cut costs by decreasing the labor force. The dynamic size of IT business value is one of the main sub-topics we want to examine in our research. In relation to our research question the notion of IT business value is central, as we are examining the choices and considerations made by the management prior to implementing a vast IT investment and delving into the prioritizations, perceived risks and barriers, success factors and evaluations surrounding the process.

The previously written literature that we have reviewed has supported the above. According to Chae et. al (2014), the link between IT capabilities and firm performance are difficult to link, and an investment in IT will not always manifest itself in a firm’s financial performance.  Melville et. al (2004) argues that IT is valuable, but is dependent upon “internal and external factors”, especially organizational resources and competitive environment, as well as the resources of its business partners. Wierenga & Oude Ophuis (1997) talk briefly about IT business value, and have conducted a study similar to ours, however, their study is based around rigorous testing of a hypothesis, whereas our research is based around an analysis of a company that has already implemented a marketing support system in an effort to build theory around the positives that can be taken from their approach. Xu et al. (2002) brushes upon IT business value from the aspect of why a firm should implement a customer relationship management system. This aspect is very relevant for our research question, however, our research should be seen as a building block upon the why. When a firm knows why they should implement a specific IT system our research can give them insights into how to prioritize, and which crucial considerations to make.    

3.2.1.2 Knowledge Creation

According to Hair Jr. (2007) data mining and predictive analytics are the keys to survival in a knowledge based economy, more specifically the ability to convert information to knowledge. The difference between information and knowledge lies in is the analysis of the information at hand, which makes the information applicable to problem anticipation, problem solving, and optimization.

The concept of knowledge creation is also relevant to our research topic, and in more than one way. Firstly, a significant part of the competitive advantage gained from implementing a process-optimizing IT system is dependent on the creation of knowledge around the new system. Secondly, to conclude that the implementation has been successful, knowledge has to be created through the new system. We are aiming to study two different aspects of knowledge creation in this paper. Namely, which knowledge is needed to implement marketing automation successfully, as well as the knowledge sharing culture and its impact on the issues with adoption and implementation of a new marketing automation system. Are successful firms using experts and go-to-guys to aid in issues with learning to use the new IT systems, or are they more likely to encourage knowledge sharing to strengthen the overall competencies of the employees as the system rolls into increasing effect and usage? Wierenga & Oude Ophuis (1997) also allude to the goal of obtaining new information through the use of marketing decision support systems, however, the differentiation between the information obtained and the knowledge created therefrom is lacking.  

3.2.1.3 Organizational Performance

The concept of organizational performance relates closely to the concept of IT business value, as financial performance remains a big part of the quantifiable results a firm produces regularly. However, organizational performance can also be measured in customer service, social responsibility and other non-financial performance improvements. Chae et. al (2014), found, as mentioned above, that there is not always a direct correlation between IT capabilities and organizational performance. This view can be contradicted by the fact that IT leaders still remain the leaders in spite of not showing significantly better financial performance improvement than the control group in their study, thus, the position of the IT leaders at the forefront of the market could be due to better performance in non-financial aspects, thereby, offering a better value proposition and differentiating themselves from the competition.    

Melville et. al (2004), say that most contemporary research on the contribution to organizational performance stemming from IT, conceptualize key constructs and relationships divergently. This view is very relevant to our research as it shows the multidimensional nature of measuring the impact IT investments have on organizational performance. Our research is based on the pretense that the implementation in question has been successful, thus, it is implied that organizational performance has been improved in either a financial or non-financial aspect. Therefore, the concept of organizational performance is not directly considered in our research, yet it is a prerequisite to understanding the considerations and prioritizations this research revolves around.  

Chang et al. (2008), focus on CRM technology, which is similar to a marketing automation technology, but not identical. The article is still very relevant, in relation to our research, as is they operate within the sphere of IT enhanced sales and marketing. Both technologies also suffer from the lack of empirical research showing the effects these technologies have on operational performance. According to Chang et al. (2008), CRM technology has the potential to improve operational performance when certain prerequisites are in place, namely, a customer-centric organizational culture and management system. These previously written articles underline the fact that merely investing in the newest IT system and expecting an increased organizational performance is hazardous, thus, investments in IT systems, especially within the sphere of marketing and sales, have to be based in strategy and built upon a foundation where these systems will upgrade processes instead of hindering core operations and increasing complexity.

3.2.1.4 Marketing Execution

Marketing execution is another key concept that we deducted from the previously written literature. As marketing is subject to the same principles as other operational segments of a firm, cutting costs, optimization, getting return of investments, etc., the direct impact on the marketing execution is central when implementing marketing automation. According to Biegel (2009) the marketing environment is gradually becoming more complex and in order to cater to a more complex and individualized consumer, thus, marketers will need to adopt more sophisticated marketing tools and technologies. This idea lies as the foundation for the marketing execution aspect of our research topic, more specifically, when a firm needs to implement more sophisticated marketing technologies, which is the best approach and what can be learned from the market leaders?. Xu et al. (2002) mainly talk about the organizational performance, as noted above, however the article contains substantial insights on marketing execution. They conclude that organizational performance hinge more on the actual usage of the marketing IT, more precisely the organizational culture and management system underlying the use of technology to optimize marketing execution in businesses. The study is quite narrow, as it is made only on Korean firms, thus, organizational culture is likely to be different when compared to western firms, however, the emphasis on an organizational culture which supports the marketing execution should be applicable regardless of a specific culture in a specific firm.

3.2.1.5 IT in Marketing

IT in marketing is one of the more vague concepts in this literature review, albeit, it is a concept which is pervasive throughout the relevant marketing literature. Biegel (2009), Hair Jr. (2007), Xu et al. (2002), Chang et al. (2008), and Wierenga & Oude Ophuis (1997) all talk about the concept of IT in marketing, however, on many different levels, in different settings, and with different purposes. The main finding in the previously written literature, on the concept of IT in marketing, is that the most important part of implementing IT is to base it in strategy, thus, being clear on why a particular system will boost marketing efforts, which system or service provider to hire, how to put it to use, when to upgrade, etc., will facilitate the implementation of a new process-optimizing IT system, and make sure the investment is solid. Strategy behind the implementation is the core of our research, therefore, we feel we can build upon what is at the core of the previously written literature, even if our actual contribution is only moderate.

4.0 Methodology

In this section we will describe our research design as well as demonstrating how it is linked to our research objective and research question. As illustrated in figure 2 “The Research Onion Outlined”, this section will focus on the two outer layers of the research onion. Firstly, we will define our overall research design and how it aligns with our paradigmatic assumptions mentioned in section 3.1 Interpretivism, secondly, we will explain the data collection methods used to collect empirical data. Overall, this section will describe how we have chosen to conduct our research while justifying our methodological choices.

 

Figure 2. “The Research Onion Outlined”

4.1 Research Design

4.1.1 Research Strategy: Single Case Study

With interpretivism as our paradigmatic base assumption we have chosen to conduct an exploratory single case study to identify the strategic considerations and choices that were prevalent to the successful implementation of marketing automation.

We want to explore which strategic considerations and choices that were prevalent to the implementation to increase our understanding, rather than describing or explaining the causal relationship between the choices and the success factors. In order to understand that particular phenomenon we believe that we will have to conduct an empirical investigation in the context of Microsoft (Saunders et al., 2009). We believe the only way to understand the phenomenon is by studying its social actors within their own given context.

 

The justification for choosing to conduct a single case study, compared to a multiple case study, is mainly due to the time constraints of this project. A multiple case study would enable us to determine whether our findings from Microsoft is an isolated incident or they actually are transferable to other large corporations in the industry. However, we strongly believe that Microsoft have a very strong and unique position when it comes to being on the forefront of new marketing technologies. Not only do they have the resources to work with these kinds of tools, but they also have a volume that allows it to manifest.

 

The purpose of this research is to determine the different strategic factors that influence successful implementation of marketing automation. As user-adoption very often comes down to opinions and actions of different social actors, we believe that understanding this phenomenon “…is far too complex to lend itself to theorising by definite ‘laws’” (Saunders et al, 2009, p. 115). We also believe that a business situation like this is very complex and so unique that it has to be studied by entering ”…the social world of our research subjects and understand their world from their point of view” (Saunders et al., 2009, p. 116).

 

Due to the time constraints of our research project, as well as our chosen research strategy, we found it evident to make a cross-sectional study where we examined our particular phenomenon at a particular point in time. This should of course also be seen in the light of the achieved benefits we could have gained from a longitudinal study, where we would have followed the change and development of Microsoft over a specific course of time. This might have given us more accurate insights on the different underlying considerations prevalent to the implementation of marketing automation. But due to our limitations and constraints in time we have chosen to look at Microsoft and take a snapshot at that particular point in time, in order to understand which considerations and choices were seen as pivotal for the implementation of marketing automation.

4.1.2 Research Methods: Multi Methods

By applying different methods for data collection, we aim get more accurate and valid results. To some extent we have been doing a methodological triangulation as we have used observations, informant interviews, and semi-structured one-on-one interviews. This triangulation enables us to validate our data through cross-verification from each of our data collection methods and thereby gaining more accurate insights of Microsoft’s context (Patzer 1996). If we had chosen a more pragmatic research design, we could also have combined qualitative and quantitative data collection methods to narrow down and clarify our results even further (Saunders et al., 2009). Of course this would be more time-consuming, however, we also believe that we have gained stronger understanding of the phenomenon.

 

According to Guba & Lincoln (1994), both qualitative and quantitative methods may be used regardless of research paradigm, however, as the two methods are very different in their approach, the advantages of using either one will be based on the application. The qualitative method is superior when the scope of the research is an exploratory in-depth investigation with a small sample size of data, mainly focused on understanding a phenomenon. The quantitative method is superior when the scope is focused around highly structured measurements concerning large sample sizes, with an emphasis on maintaining objectivity. Thus, we chose the qualitative method for this research in order to utilize the strengths of the subjective interpretivist approach, as our research seeks deeper understanding of a phenomenon.  

 

By first conducting observations and informant interviews we gained an initial understanding of the context which clearly strengthened the quality of our interview guide for the semi-structured interviews. As we have chosen to make an exploratory study, the obvious way to collect data would either be to conduct unstructured or semi-structured interviews as those will help “find out what is happening [and] to seek new insights” (Robson, 2002). Furthermore, the reasoning behind the choice of semi-structured interviews is that we wanted to be able to gain rich and in-depth knowledge of the context. As we inexperienced in the field of marketing automation from the beginning of this project, we did not want to limit ourselves to our interview guide, but instead wanted the freedom and flexibility to deviate from our list of planned questions, in case the respondent brought other more meaningful topics into the conversation. For the same reason, and in order to gain a nuanced view on the reality, we also asked a lot of open questions, in order for the respondents to bring up other essential topics or themes.

 

4.2 Data Collection Methods

4.2.1 Observations

Before conducting any interviews, we had one group member sent off to Microsoft in order to gain an initial understanding of the context and its social actors. As one group member already worked at Microsoft, he was able to conduct inside research through observations acting as a complete observer. As he was able to “conceal his identity” while observing the social actors, it was possible to gain an initial understanding of the context without changing or affecting the behavior of the social actors.

In the beginning the observed actors were unaware of the observations, hence, we gained a closer-to-reality impression of the actual context. Afterwards, when we had gained enough knowledge about the actors, while having a concealed identity, the researcher revealed his identity before continuing his observations. This made it possible to have a more open and direct dialogue about the phenomenon. Now as the purpose of the conversation were even clearer, it was possible to ask more straightforward and clarifying questions about the implementation process.

4.2.2 Informant Interview

After we had conducted our observations, we did an unstructured in-depth informant interview. The purpose of this was, firstly, to confirm our observations and, secondly, to further understand the strategic considerations and choices prevalent to the implementation of marketing automation. The gained knowledge was also useful to specify and narrow down our problem statement which we were continuously refining. The informant interview gave us a more technical insight on the implementation of marketing automation, as well as a more holistic view on the whole implementation process. Furthermore, the informant interview helped us acquire knowledge of the overall structure of Microsoft, and Microsoft Denmark’s role therein, as explained in 2.1 Case description.  

Lastly, the informant interview also laid the foundation for the interview guide we later used in our more structured follow-up interviews.

4.2.3 Interviews

After the exploratory informant interview, and an unstructured analysis of our findings herein, we progressed forward to a more explanatory phase. The learnings from the early observations and in-depth informant interview were processed and compiled into a semi-structured interview guide for later use. The choice of semi-structured interviews for the follow-up interviews were based on a need for a more structured approach without sacrificing the desired flexibility. According to Saunders et al., (2009) explanatory studies are well supported by semi-structured interviews as they excel at increasing understanding of the relationship between variables. One could argue that in-depth interviews share the same benefits, however, to explain the phenomenon in question thoroughly, there was a clear need to conduct semi-structured interviews where questions could be divided into different segments in an attempt to understand how Microsoft had first planned the implementation process and later to question their own evaluation of the process.  

 

4.2.4 Data Sources

Our study is an embedded case study, as it primarily takes place in the Marketing Communication department, who are in charge of implementing marketing automation for the Danish market. Hence, we had a very natural limitation in terms of what employees we could pick-and-choose for interviews. In an effort answer our research question we were seeking nuanced and diverse insights into the employee’s reality.

 

Firstly, in order to gain access to our respondents and resources, we had to straighten out permissions of conducting our research at Microsoft. As one member of the group members were employed at Microsoft at that particular time, we gained access quite easily. As the group member worked in the Central Marketing Organization of Microsoft Denmark, this was where we also reached our limitations in terms of how extensive we could gain access to resources. If time and resources allowed it would have been interesting to study different departments or subsidiaries from different markets.

Our observations within the Microsoft organization led us to determine who were the best respondents for our interviews. In order for us to gain the best possible results, we chose to go with someone from the management angle and then someone who works hands on with the implementation of marketing automation.

 

• Signe Toftegaard Møller – Central Marketing Organization Lead (Management)
• Christian Christoffersen – Product Marketing Manager of Azure and Enterprise Mobility

 

When planning the interviews, we went with a semi-structured format, with an underlying hybrid design, which is part appreciative and part laddering (Schultze & Avital, 2011).
This methodological approach assisted in generating deeply contextual and authentic data from the semi-structured interview method.

For the setting of the interview, we followed the appreciative design method and made sure we were as polite as possible and set a positive tone for the following interview. Throughout the interview we stayed clear of negative aspects and made sure not to put the respondents on the spot, so that they did not feel the need to defend themselves against the questions that were being asked, because their involvement, good or bad, was not the topic.
The appreciative interviewing method “… purposefully seeks to turn attention away from the prevailing investigation and treatment of dysfunctions, anomalies and deviant behaviors toward a relentless search for evidence that demonstrates strengths, possibilities, goodwill, and the grace of the human spirit.” (Schultze & Avital, 2011, p. 5). Furthermore, the appreciative method allows respondents to share their most outstanding personal experiences, thereby, assisting the research in unearthing the best practises.

On top of this, we wanted to utilize the strength of depth in the inquiry of the laddering interview method, and chose to deliberately asks follow-up questions, based on the answers of the respondents. This method made us able to react to the data gathered and investigate the answers we encountered, which facilitated the collection of rich data from the interviews.

Furthermore, our interview designs were planned with an emphasis on establishing a personal contact with the interviewee. The emphasis on personal contact with the interviewee was to increase the validity of the data and to contain any potential bias. According to Saunders et al. (2009), establishing a personal contact with the interviewee is beneficial for the data collection, as the interviewee is more likely to be honest and open about personal and sensitive information. Moreover, a personal connection between the interviewer and interviewee is more likely to increase trust, thereby, reducing potential response bias (Saunders et al., 2009).

4.2.5 Triangulation

To achieve a more nuanced and valid set of empirical data, we have throughout our research made use of different sorts of triangulation. Even though triangulation can be more time consuming, we strongly believe that the gained effect of triangulation have added value to our research. In this paper we have used triangulation in the following ways:

 

As mentioned earlier, we have both used observations, informant interviews and semi-structured interviews as parts of our data collection method. By triangulating our data collection methods a more nuanced view on our phenomenon has presented itself, as well as using the different methods progressively each one building upon the previous step. We used our literature review to discover potential pitfalls before conducting the observations. Building upon that, we used our observations to prepare for the informant interview, and finally we used the response from the informant interview to prepare for the semi-structured interviews.

 

To increase validity and minimize errors in the transcription, not only were the two different interviews transcribed by two different group members, but all three group members have listened to the interviews and read through the transcriptions. By doing so, we have maximized the likelihood of identifying important points from the interviews.

 

We fully acknowledge that even more use of triangulation would ensure better quality in our research, but due to limitations in time, we have chosen not to triangulate our research approaches. We could also have chosen to collect qualitative data we could use to combine with our qualitative empirical data. Not only would this have been very time consuming, it would also have have leaned towards a more pragmatic approach that would have been in contrast to interpretivism as our paradigmatic base assumption.

5.0 Measurements

5.1 Data Preparation

Collecting and analysing data through the method of semi-structured qualitative studies is very much an interactive process (Saunders et al. 2011). Based on our conclusions from the informant interview, our concept matrix gathered from the literature, and the propositions encountered through the interviews themselves, the analysis of the data was a continual process throughout the research. This is an important part of the inductive research method and allows for us to explore and follow-up on the concepts and issues that emerge. This also allowed for a relative flexibility in terms of the chronology of the research project.

However, due to the complex nature of qualitative studies, once the interviews had been conducted, we faced the challenge of extracting and condensing the essence of the interview, which we could use to conclude on our research question. According to Saunders et al. (2009), this is done by summarizing the answers, putting them into categories in a conceptual framework in order for meaningful analysis to take place, and will be developed on further in section 6.0 Results.

 

In order to do so, we have to be able to work with the qualitative data gathered from the interviews, in constructive manner. This is done by transcribing the audio recordings of the interviews, which is a written reproduction of the accounts. Though transcribing interviews is a lengthy process, it is indeed needed to prepare the collected data for the analysis ahead.

 

See Appendix 1. and 2. for transcriptions of interviews.

5.2 Data Quality Issues

Data quality issues are inherent in any research where empirical data is collected, analysed and interpreted. These data quality issues can be segmented into four main categories to isolate and process each issue individually, namely, reliability, generalizability, bias, and validity (Saunders et al. 2011). In order to accommodate for any potential criticism relating the quality of the data, we have analyzed any potential data quality issues in the following section.

 

The most prominent issue with the qualitative data collection method is a lack of generalizability and reliability. Mainly, due to the unrepresentative nature of semi-structured and in-depth interviews difficulties regarding statistical generalizations are innate in any collection of qualitative data (Saunders et al. 2009). The small sample size associated with qualitative studies, compared to the larger sample size expected when conducting quantitative studies, will also influence the transferability of the research. However, when conducting a single case study relationships are more dynamic and complex, thus, modifying the argument that a quantitative data collection method would increase the transferability of the findings (Saunders et al. 2011). One could also argue that an in-depth study of a single case is not meant to replicable, however, since the exploratory and explanatory nature of such a study aims at reflecting the current reality in the current situation and extract learning to deepen understanding of a dynamic phenomenon. According to Saunders et al. (2009) significant efforts to account for the lack of generalisability could inhibit the qualitative data collection method, as attempts to make the research replicable by other researchers can potentially undermine the strengths of the dynamic qualitative approach. We have been conscious of these data quality issues throughout the research process and are therefore solely aiming at constructing guidelines based on the experience of a selected segment, thus, focusing on inspiring the choices of future decision-makers or researchers, rather than large scale generalizations and definite transferability.

 

Bias, stemming both from the interviewer or the interviewee, can be another potential issue when trying to maintain the integrity of the data quality, this is known as interviewer bias and/or response bias (Saunders et al. 2011). As the interview situation is a dialogue between two or more individuals, prior perceptions affect relationships, views can be imposed upon the other part, questions can be framed, and so forth. Bias has been extensively covered in other research disciplines such as psychology and communication, and will not be handled further in this paper. However, it is important for us to point out that we are aware of potential bias emerging through our data collection, and therefore have accommodated for possible bias when planning our interviews by using the interview techniques explained in 4.2.4 Data Sources.   

 

Validity is also a major concern when dealing with any type of collection of data. However, qualitative data collection methods have some advantages over their quantitative counterparts due to their flexibility and the dynamic nature of a dialogue. Thus, in an interview setting misinterpreted questions can be clarified, responses can be probed for further explanations, and elaborations in general will help increase the validity of the collected data and help avoid misunderstandings.       

5.3 Analytical Procedures

 

To further the preparation of the data for analysis we will need to summarize and derive categories for our data in order for a conceptual framework to emerge. This method of analysis is also known as the grounded approach, whereby the exploratory research has begun without a fixed theoretical framework, also explained in section 4.0 Methodology.

Instead we see developing relationships between the emerging themes and concepts, to be tested in order to conclude on the findings in the data. The theory will reveal itself through the data collection and analysis (Saunders et al. 2009).  

 

To gain a more condensed overview of the findings in our data, we chose to gather the key data points and sum up our interviews in a summary. This process also prepares the data for the next phase of analysis, whereby the key findings are categorized.

The process of categorizing our data means that we are engaging with the analysis part and moving towards recognizing the mentioned relationships between our data points. This enables us to develop testable propositions that emerge inductively from our data and to conclude on the actual relationships between them (Saunders et al., 2009). “Alternative explanations frequently exist, and only by testing the propositions that you identify will you be able to move towards formulating valid conclusions… ” (Saunders et al. 2011, p. 495).

The conceptual framework emerging from the stated procedure, will be the foundation our results and the main part, where our ongoing analysis is summed up and can be tested upon. This will enable us to conclude on our research and answer the research question.

6.0 Results

Based on our data collection we will, in this section, analyse and discuss the results we have gotten from our interviews. As we have chosen to conduct unstructured interviews, our results are meanings expressed through words, summarised and condensed into categorized which are subsequently analyzed through the use of conceptualisation. (Saunders et al., 2009)

6.1 Summaries

The output of the interviews were audio files we then transcribed. In the following section summaries of the two semi-structured interviews will follow, as described in the 5.1 Data Preparation section, which will provide the material for the categories of our conceptual framework used to conclude on our findings.

 

6.1.1 Summary of the interview with Signe Toftegaard Møller, Central Marketing Organization Lead (Appendix 1).

Signe starts out by informing us, that she is the current marketing director, but that she has only been in this role since the 1st of April. However, she has previously held a position relevant to this study, as she worked hands on with the implementation of marketing automation in Microsoft Denmark, furthermore, she has been 12 years in the company.

She begins her insights on marketing automation with telling us, that today there are many other software suppliers in the market and therefore Microsoft needs to be a lot more personal and specific when talking to the individuals online. She goes on to make some comparisons to other large players within marketing automation, namely Amazon and Salesforce and told us that before the implementation of marketing automation in Microsoft, a decision was made to take their efforts of direct sales to the next level, and with that, the Marketo platform was chosen and how it is becoming more and more significant for marketing in Microsoft to drive traffic to the demand centers (marketing automation lead nurture flows).

 

One of the important aspects was that the potential of Microsoft’s cloud offerings are so vast, that they do not have the manpower to handle all the requests manual marketing, and the only way of relevantly reaching everyone is with automation.

She goes on to explain how Marketo actually works, with lead nurturing through nurture streams, and how content, actions and sign ups to gated content or events, give users a lead score. Once they reach a certain threshold of points, they are sent to an appropriate department for sales execution, either by direct mail, telemarketing or similar.

So the lead scoring system and the inner workings of the Marketo platform is managed by Corp, but the content get localized by the subsidiaries, as well as customizing the nurture flows for the local markets, she says.

 

A surprising find in the beginning of the implementation process, was to figure out that content like webinars didn’t perform as expected and people wanted to find something physical that they could download to their machine, like white papers. The knowledge sharing between the subsidiaries was crucial in this phase as it helped them learn best practises from different markets with more experience.

She mentions that corp decided to roll out marketing automation as a pilot on the Azure product line for the B2B segments, but that she feels that it should be a priority to prove the success of the system and roll it out on all other products as well.

 

There have been some organizational pushback, especially from the sales departments who think it remains to be seen, if the system will bring leads that leads to actual sales. And that they argue that the automation system falls short on big clients. She thinks it’s natural for the sales departments to fear a system that could in time replace them and lead to big organizational changes.

However, where the system is working really well is for the “unmanaged space”, the space in the market where they are too small for personal account managers and too big for consumer products, and normally wouldn’t be targeted by the sales department.

 

In her opinion, “content is most important”  and it has to be audience specific and relevant for the end user in order for marketing automation to be successful, which poses a myriad of other challenges both with localization and the richness of contextual content for the individual.

She believes that marketing automation has had a profound success so far and that “it’s not a matter if we should do marketing automation it’s a matter of how fast we should do it” and she feels that the are missed opportunities with not having the system in place for the consumer market and says that it feels like they are missing the customer centric focal point, because of it.

6.1.2 Summary of the interview with Christian Christoffersen, Product Marketing Manager of Azure and Enterprise Mobility (Appendix 2).

Summary: With this interview, we gained insights from a person, who is deeply engaged with the deployment and technical side of the implementation. He was a significant part of the marketing automation project from the very beginning, with the Azure pilot, underlining his strategic position within the company and the depth of his knowledge. He quickly goes on from the introduction to talk about the content management system, and how corp manages the nurture streams and the subsidiaries are in charge of localization of the content used in the said nurture streams. He emphasises that the main reason why it is important to get on board with Marketing Automation, is that companies both big and small, are rapidly changing their software procurement methods, which relies more and more on online software as a service deals, where no contact with the sales team is required. This means that Microsoft needs to be on top of their online marketing, which not only encompasses marketing automation and content marketing, but also on SEM (Search Engine Marketing) and SEO (Search Engine Optimization) and underlines the crucial need for Microsoft to master each discipline.
He compares Microsoft to other big players in the marketplace like Salesforce and Amazon and talks about how they are also on the edges of the online marketing trends.

In regards to this, he thinks that the way Microsoft have traditionally been conducting their business is very old school, which is why it is important Microsoft adapts to the new landscape.

He talks about decision making process of marketing automation and that it was chosen by Microsoft Corp to go with the Azure product line for the initial pilot. Then goes on to the learnings gained from this pilot. Here he mentions that it is very much a learning by doing process, and what they thought to be the optimal way forward, with specific gated content around webinars, was actually not as successful as other methods, like white paper downloads and physical events, to be used to get people’s information into the nurture streams. He argues that webinars is in general was too far down the buying process to spark the interest of potential buyers browsing the content.

 

Towards the end he talks about how the clear success of marketing automation is yet to be perceived by the rest of the company, especially the sales department, who is somewhat scared of the change marketing automation could bring. He concludes that a priority on the success of the pilot and a keen eye on content excellence is the most important and that “content is king”. On his final notes he touches upon how bots and artificial intelligence will have a profound impact on the field of marketing automation in the future.

6.2 Conceptual Framework

Based on the summaries above, the conceptual framework outlines the key areas and will provide the categorization needed to conclude on our research findings.

 

From the interviews we quickly uncovered that the implementation of marketing automation was considered a success (Appendix 1). Furthermore, we identified the following five different categories of considerations and choices that were prevalent to the successful implementation and adoption of marketing automation in Microsoft, through extensive analysis:

 

1. Prioritization
2. Knowledge sharing
3. Content creation
4. Change management
5. Organizational structure

 

In the following we will go into depth with these five concept categories.

 

6.2.1 Prioritization

Even though Signe Toftegaard Møller expressed a wish to roll out marketing automation for all products across all markets, we identified that one of the key factors in Microsoft’s success were the prioritization made. It was decided early in the process to only roll out marketing automation for Azure in the different markets. With a clear ambition to fully implement marketing automation within the whole product portfolio, Microsoft believes that this stepwise approach have given them the opportunity and possibility to learn while implementing. Christian Christoffersen says that it is important “to succeed with the things that we have been in a learning process with, for the last couple of years, before we begin to do other things.” (Appendix 2).

6.2.2 Knowledge Sharing

Another factor prevalent to Microsoft’s successful implementation of marketing automation is knowledge sharing across all markets. By establishing the necessary platforms for sharing knowledge, Microsoft Corp have made it possible for the different markets to share the different learnings gained locally in the markets. Sharing these learnings with the rest of the markets, increases the chances for a collective success for all parties involved in the implementation. To further assist in the lack of specialist knowledge, Microsoft Corp have also facilitated webinars, workshops and more in order to teach the different subs about all aspects of marketing automation: “They did a workshop where they came around into the markets which has been very good, and we have contacts in Corp and they have Sharepoints and websites and loads of information” (Appendix 1). Christian Christoffersen also agrees and confirms when we ask about if he sees knowledge sharing as an important part of the implementation process (Appendix 2).

6.2.3 Content Creation

Traffic and visits are crucial to collect data for marketing automation systems. And when we asked Signe Toftegaard Møller about what is the most important for their implementation and their ability to drive traffic, the answer is simple:  “I think content is the most important” (Appendix 1). Also Christian Christoffersen mentions that “is content is king” (Appendix 2). Signe Toftegaard Møller emphasises that it is not only important to create content – the content simultaneously needs to be relevant and audience-specific: “But I would be very audience specific in order to make the content as relevant as possible.” (Appendix 1). She also touches upon the challenges of making it audience-specific, as efforts to segment the relevant content and make it and audience-specific also requires a lot of resources.

6.2.4 Change Management

The common conception is that increased automation of work will replace jobs and decrease labor-costs. Signe Toftegaard Møller also acknowledges that fact: “I am afraid that it will replace the need for people in the subs over time both in sales and marketing” (Appendix 1). Therefore there is a strong need for change management. She explains how they have tried to involve sales and marketing in the process to prepare them for that particular change and thereby minimize potential resistance. Christian Christoffersen adds that he believes that “reactions to this kind of change are still to come” (Appendix 2). Besides the change management efforts made to reduce resistance to change, Microsoft are also aware of the cultural change that follows the implementation: “So it is a cultural change also where sales have to let go and they have to trust that it’s actually for the better.” (Appendix 1).

6.2.5 Organizational Structure

As mentioned in 4. Change Management, it is no secret that the implementation of marketing automation will replace jobs over time, meaning that Microsoft are facing an organizational change. Furthermore, Signe Toftegaard Møller explains how Holland, as one of the leading subsidiaries, can attribute a lot of their success to fact that they have allocated three times as many resources for their implementation of marketing automation. Finally she enhances that statement saying, “they have been digital for a long time they also got a CMO leader three years ago who’s all digital”. So not only is the right organizational structure important, but resource allocation also needs to be factored in.  

7.0 Conclusion

As supported  by the previously written literature, particularly in relation to the concepts of IT Business Value and Organizational Performance, many factors play a role when implementing new IT into a large company.  Based on the results achieved through this study, the factors have been narrowed down and compiled into a list of five concepts that require extra attention when attempting to successfully implement marketing automation technology in a large company, namely, Prioritizations, Knowledge Sharing, Content Creation, Change Management, and Organizational Structure. These conclusions have been reached by trying to understand the implementation process undergone by Microsoft Denmark, with particular focus on their prior considerations, their choices during the implementation process, and the following evaluations.

 

This study is burdened by the limitations of the inductive research method and the challenges of producing reliable results therefrom. However, this study supports previous research on the difficulties of implementing IT in large corporations, and produces dynamic guidelines to assist similar companies with the strategic aspects of implementing marketing automation, based on privileged information that discloses a glimpse of the inner workings of a multinational corporation, rarely accessible to the public. Future research is needed and encouraged, more specifically, a deductive study testing the theory constructed through this research could assist in solidifying the reliability of the results.

 

8.0 References

Websites:

• What is Azure (2016). Retrieved from https://azure.microsoft.com/da-dk/overview/what-is-azure/

 

Blog Articles:

• Watson, Z. (2015, April 22). How to Use Marketing Automation as Lead Generation Software. Technology Advice, Retrieved from  

http://technologyadvice.com/blog/marketing/how-to-use-marketing-automation-as-lead-generation-software/

• Wilson, C. (2015, December 1). Will SMEs find any Value in Marketing Automation? Smart Insights, Retrieved from

http://www.smartinsights.com/lead-generation/marketing-automation/will-smes-find-value-marketing-automation/

Books:

• Guba, E. and Lincoln, Y. (1994) Competing paradigms in qualitative research, in N.K. Denzin and Y.S. Lincoln (eds). Handbook of Qualitative Research. London: Sage, pp. 105–17.

• Patzer, G.L. (1996) Using Secondary Data in Market Research: United States and World-wide. Westport, CT: Quorum Books.
• Robson, C. (2002). Real World Research (2nd edn). Oxford: Blackwell.
• Saunders, M., Lewis, P. & Thornhill, A. (2009). Research Methods for Business Students (5th edn.). London: Pearson.
• Saunders, M., Lewis, P. & Thornhill, A. (2011). Research Methods for Business Students (6th edn.). London: Pearson.
• Yin, R.K. (2003). Case Study Research: Design and Method (3rd edn). London: Sage.

Journal Articles:

• Biegel, B. (2009). The current view and outlook for the future of marketing automation. Journal of Direct Data and Digital Marketing Practice, 01/2009; 10(3), pp. 201–213.

DOI: 10.1057/dddmp.2008.37

• Chae, H. Chang, K. E. & Prybutok, V. R. (2014). Information Technology Capability and Firm Performance: Contradictory Findings and Their Possible Causes. MIS Quarterly Vol. 38 No. 1 pp. 305-326. Retrieved from  http://aisel.aisnet.org/cgi/viewcontent.cgi?article=3168&context=misq
• Chang, W., Park, J. E.& Chaiy, S.(2010). How does CRM technology transform into organizational performance? A mediating role of marketing capability. Journal of Business Research, Volume 63, Issue 8, pp. 849–855.

DOI: 10.1016/j.jbusres.2009.07.003

• Hair Jr, J. F. (2007). Knowledge creation in marketing: the role of predictive analytics. European Business Review, Vol. 19 Iss: 4, pp. 303 – 315.

DOI: 10.1108/09555340710760134

• Melville N., Kraemer, K. & Gurbaxani, V. (2004). Review: Information Technology and Organizational Performance: An Integrative Model of IT Business Value. MIS Quarterly Vol. 28, No. 2, pp. 283-322. Retrieved from

http://www.jstor.org/stable/25148636

• Schultze, U. & Avital, M. (2011). Designing interviews to generate rich data for information systems research, Information and Organization Vol. 21, pp. 1–16.

DOI: 10.1016/j.infoandorg.2010.11.001

• Wierenga, B. & Oude Ophuis, P. (1997). Marketing Decision Support Systems: Adoption, use, and satisfaction. International Journal of Research in Marketing, Volume 14, Issue 3, pp. 275–290.

DOI: 10.1016/S0167-8116(97)00009-8

• Xu, Y., Yen, D. C., Lin, B. & Chou, D. C. (2002). Adopting customer relationship management technology. Industrial Management & Data Systems, Vol. 102 Iss: 8, pp.442 – 452.

DOI:  http://dx.doi.org/10.1108/02635570210445871

 

Appendices: