1
Tracking your marketing investment:
Adoption and usage of marketing ROI within the
marketing department
– a case study for NS
By
Renate Enthoven
Final version
15 June 2018
Master Thesis Marketing Management
1e supervisor: Prof. dr. P.C. Verhoef (p.c.verhoef@rug.nl)
2nd supervisor: Dr. J.T. Bouma (j.t.bouma@rug.nl)
NS supervisors:
Mrs. J. Salomé (josien.salome@ns.nl) Mr. J.F. Traas (joey.traas@ns.nl)
Renate Enthoven S3226034
Van Julsinghastraat 33, Groningen r.a.enthoven@student.rug.nl
+31 6 38668371
University of Groningen Faculty of Economics & Business
2
Abstract
Marketing accountability can be measured by marketing performance metrics such as a return on marketing investment (ROMI). However, the actual adoption of these marketing performance metrics within the marketing department has been found difficult by organizations (Mintz & Currim, 2013; O’Sullivan & Abela, 2007; Wierenga & Oude Ophuis, 1997). Therefore, a case study was conducted at Nederlandse Spoorwegen (NS). Based on ten internal in-depth interviews and three external interviews (i.e. Schiphol Group, KPN, MIcompany), an efficient and effective way to adopt and use marketing performance metrics was found. The results of this research show several points of improvement. To have a successful marketing management support system adoption, the fit between the user and metric should be optimal. Besides, a lack in user involvement causes difficulties in the implementation of a metric (Wierenga & Oude Ophuis, 1997). The results show that this relationship is influenced by organizational culture, intrinsic and extrinsic user motivation, and user capabilities. Organizations should provide appropriate training and good communication between different marketing teams to help with a successful adoption. Furthermore, the results indicated that top management support is missing, while literature argues that top management support is a requirement of a successful adoption (Premkumar & Ramamurthy, 1995; Wierenga & Oude Ophuis, 1997). Therefore, top management should actively steer on a marketing performance metric (in this case ROI), in order for marketers to execute the marketing method. Additionally, the results found that an users’ predisposition (i.e. attitude and trust) towards the marketing metric needs to be sufficient for an adoption success. All in all, managers should be aware of the steps they need to undertake in order to adopt and use their marketing performance metrics more efficiently and effectively within their marketing department.
3
Preface
I’m very proud to present my Master Marketing thesis, which lays in front of you. My name is Renate Enthoven and this thesis is part of my graduation of the Master Marketing
Management at the University of Groningen. In association with NS, this thesis provides insight into the theme of adoption and usage of marketing performance metrics within a marketing department.
As this is my last official master contribution, I would like to take the opportunity to thank a few people. First of all, I would like to thank NS for giving me the opportunity to write my thesis for the organization and getting to know the business processes of the (marketing) department. Herein, a few people need to be put in the spotlight. A big shout out should be done to team K&M Marketing who were very supportive of me and made me feel part of the team. Above all, thanks to Joost Bosma for his trust and openness in giving me the chance to write my thesis at NS. Another special thanks goes out to Josien Salomé and Joey Traas for being my supervisors and supporters during my internship period at NS. Besides, Kerstin Unterlauf and Jaimy Kral helped a lot with providing feedback on (parts of) my thesis, thank you. Second, I would like to thank my RUG thesis supervisor, Peter Verhoef, for his guidance. With the help of his critical feedback, I was able to learn a lot during my thesis period.
I never expected to be finding myself in this position, finalizing my master thesis. It were two amazing years in Groningen (premaster and master). However, these years would not be the same without the support of my friends, thank you for everything. I would like to direct my final thank you words to my family, especially my parents for supporting me throughout the last two years. There were many ups and some downs, but with their support I always felt that anything was possible. Thank you so much for your trust!
Kind regards,
4
Table of content
ABSTRACT ... 2 PREFACE ... 3 TABLE OF CONTENT ... 4 1. INTRODUCTION ... 5 2. LITERATURE REVIEW ... 82.1 Financial metrics to measure marketing effectiveness ... 8
2.2 Marketing return on investment ... 11
2.3 Measuring online marketing performance ... 16
2.4 Usage of marketing metrics ... 17
2.5 Adoption models ... 17 2.6 Conceptual model ... 22 2.7 Literature summary ... 23 3. METHOD ... 24 3.1 Research type ... 24 3.2 Data collection ... 25
3.3 Sample method and size ... 28
4. CASE STUDY – NS ... 29
4.1 Current marketing effectiveness measurement at NS ... 29
4.2 The Braintrain ... 30
5. RESULTS ... 32
5.1 General ... 32
5.2 Stakeholders ... 33
5.3 Adoption and change processes ... 35
5.4 Marketing Management Support System (MMSS) ... 35
5.5 Effective and efficient ROI ... 38
5.6 External results ... 39
6. THEORETICAL MODEL ADJUSTMENTS ... 42
5
1. Introduction
“Half the money I spend on marketing is wasted; the trouble is, I don’t know which half” – John Wanamaker
In the past, there has been no clear view of the impact that a marketing department has on firm performance (Nath & Mahajan, 2008). One thing is undeniable, it is hard to define and measure the value of your market offering (Anderson & Narus, 1998). However, firms can benefit from having a Chief Marketing Officer (CMO) as part of their top management team to help with strategic decisions (Germann, Ebbes & Grewal, 2015). Contrary to this former statement, earlier research by Verhoef & Leeflang (2009) stated that the influence of marketing departments within organizations decreases due to the lack of accountability and innovativeness. Moreover, this lack of accountability threatens marketing credibility and, even worse, the position of marketing within an organization (Rust, Ambler, Carpenter, Kumar & Srivastava, 2004). Therefore, a marketing manager needs to focus on the creation of value for the business by improving the performance of marketing (Jeffery, 2010). In other words, marketing performance and effectiveness are not only relevant for the marketing department itself, but also for the rest of the organization (Fu, Phillips & Phillips, 2018).
6
To emphasize this, orientation interviews were held with a few NS stakeholders1, some positive
and negative feedback came forward on the current way to measure the marketing ROI (Appendix I). A commonly used positive point for the current ROI measurement is the underlying idea. Calculating the incremental revenue is innovative and outstanding from a lot of other companies. Furthermore, this incremental revenue can give direction and forces marketers to rethink their actions when this outcome is negative. Taking into account the negative point of the current ROI measurement at NS, it is argued that the external research on which the post-ROI’s are based, is not reliable. People might tend to respond in a desired and strategic way, leading to biased results. For example, customers of NS campaigns are asked to answer the following question: “Would you have made this trip anyway, without the offer of NS?”. People might say ‘no’, but actually they would have made the trip anyway because of an appointment. In this case, the customer is unaware of this and the answer would be invalid, while a big part of the ROI is based on this customer response. Another point raised is that the campaign researches are very expensive (in euros, but also capacity wise) and it is an unscalable process.
The orienting interviews not only reveal that the current ROI process is lacking input from NS marketers, it also shows that it is a very time-consuming process. Due to this lack of time, marketers are not able to learn more about the ROI process themselves, which results in postponing the making of a ROI for their campaign or asking team KMM for help. Although NS has a key role in this case study, one could argue that there are more companies with similar problems in similar situations. Companies develop a support system for their marketers to measure the effect of their marketing performance, however, the actual usage of these systems by the marketers is lacking.
To address this problem, this case study will focus on answering the following question:
“What is the most efficient and effective way to adopt and use the marketing performance metric (i.e. ROI) within a marketing department?”
In essence, the answer to this general question covers the problem stated above. Hereby, different perspectives should be looked upon. Arguably, the most important perspective is the one of the firms’ stakeholders. Which stakeholders are involved in the use of the ROI? And
1 The stakeholders are: the head of the K&M department, the head of the Marketing & Sales department, a sales
7
what do the stakeholders need, in order to be able to work with the ROI marketing model? To provide support for the general research question, several specific research questions were formulated:
1. Which metrics measure marketing performance?
2. What are the needs of stakeholders in using marketing performance?
3. What are the main points of improvements of the current ROI implementation within the marketing department?
4. What are the main barriers for an organization-wide adoption model?
Although NS is leading in this research, the results and implications will be implementable within any organization. Furthermore, this research will contribute to the existing literature in the way that it provides an accurate advice on how marketing measurements can be implemented throughout the whole marketing department. Earlier work on measuring the effectiveness of marketing stops at explaining the value of measuring marketing performance (Fu et al., 2018; Milichovsky & Simberova, 2015; Mintz & Currim, 2013; Rust et al., 2004a), and forget about the actual implementation of such a marketing performance metric (Wierenga & Oude Ophuis, 1997). Thus, this research will fill the gap of the missing literature on how to implement this in the most efficient and effective way.
8
2. Literature review
The measurement of marketing activities is a difficult task for many companies (Jeffery, 2010). Some companies do not have an appropriate culture of internal processes to support such measurements, other companies are not able to support these data driven marketing practices with their infrastructure. Thus, although marketing metrics are very useful, not every company has the right resources to measure their marketing effectiveness (Jeffery, 2010). Yet, finding the appropriate key metrics to measure marketing is very important and contributes to the evaluation of marketing effectiveness (Milichovsky & Simberova, 2015). In this section, different marketing metrics will be discussed, under which the commonly known, return on investment (ROI). Furthermore, adoption models will be deliberated.
2.1 Financial metrics to measure marketing effectiveness
Marketing should be financially accountable (Ambler & Roberts, 2008; Ganesh & Paswan, 2010; Rao & Bharadwaj, 2008). One way to make marketing accountable is by return on investment, which is discussed in the next section. However, there are also other financial metrics like Discounted Cash Flow (DCF) and Return On Customer (ROC) (Ambler & Roberts, 2008). These metrics are a basis for ROI calculation. Another useful metric is customer equity, based on the discounted lifetime value of a customer (Lemon, Rust & Zeithaml, 2001)
2.1.1 Discounted Cash Flow
According to Rao & Bharadwaj (2008) the discounted value of the cash flow contributes to the firm value. Therefore, these cash flows must be linked to the firms marketing activities (Doyle, 2000). Several techniques can be used as cash flow. Examples of this are net present value (NPV) and customer lifetime value (CLV) (Ambler & Roberts, 2008).
Net Present Value (NPV). The NPV is measured by subtracting the costs from the present value
9 Customer Lifetime Value (CLV). “Customer lifetime value refers to the monetary value of the
customer (or group of customers) during a time period” (Berger et al., 2002). Although the CLV is not always measured in a detailed and discrete way, this metric is rapidly acquiring acceptance (Venkatesan & Kumar, 2004). Usage of CLV supports the acquisition, growth and retention of the right customers. Furthermore, marketing decisions should be taken based on CLV by designing and budgeting the activities (Berger & Nasr, 1998). For these budgets, the CLV can be calculated over time by the net cash flow that a firm expects to get back from a customer.
Accounting for these techniques, the DCF metrics have some major drawbacks (Ambler & Roberts, 2008). First, there is a lack of independence between the employee who makes the DCF forecast and the marketer. Second, the forecasts are not guaranteed to be correct and reliable. Third, a forecast is not a good basis for performance, because it is not a benchmark. This could result in confounding forecasting errors. Fourth, since any performance variance needs to be minimized by the management, multiple forecasts are reconciled. Finally, it is hard to differentiate future improvements from future activities in contrast to past activities. All in all, it is important for a company to outweigh the benefits against the drawbacks and decide on which metric to focus (Ambler & Roberts, 2008)
2.1.2 Customer equity
10
However, customer equity also has three major drawbacks. First, a company must be able to understand how customer equity can grow and how it can be managed (Lemon, Rust & Zeithaml, 2001). This is a very complex process. Second, factors such as product sales, brand strength and short-term financial performance are seen as secondary indicators of a company’s success (Blattberg & Deighton, 1996). If a company focuses on the customer equity, they have to keep in mind that the financial aspect of the marketing is not the main key performance indicator anymore. Finally, customer equity has aggregate and disaggregate level approaches (Kumar & George, 2007). Both show the same emphasis on retention, but there are conceptual differences in acquisition and accounting for existing customers. Therefore, it would be best to measure customer equity from a hybrid approach (Kumar & George, 2007), which is very complex.
2.1.3 Return on Customer
Peppers and Rogers (2005) define return on customer as “a firm’s current-period [net] cash flow from its customers plus any changes in the underlying customer equity, divided by the total customer equity at the beginning of the period”. By maximizing the ROC, the current period and future profits will also be maximized. Hence, a positive ROC shows that the firm is doing better than expected (Ambler & Roberts, 2008). However, the missing information on the value (i.e. inaccurate forecasting or marketing performance) should be taken into consideration, especially when relevant for the performance evaluation.
11
2.2 Marketing return on investment
The most common way to measure marketing efforts is by the return on investment (ROI). Return on marketing investment (ROMI) is defined by Powell (2002) as “the revenue or margin generated by a marketing program divided by the costs of that program at a given risk level.” To simplify this, the ROMI is the marketing revenues minus the marketing costs. By this definition, return on marketing is not only a measure of brand equity, but also a measure of finance (Powell, 2002). Therefore, ROMI is a financial metric which provides a way to compare alternatives and gives insights in the accountability (Mintz & Currim, 2013; Stewart, 2009). Nonetheless, financial metrics like the ROMI must be supported by marketing metrics (e.g. market share, awareness, loyalty, etc.) in order to prevent marketing uncertainty (Mintz & Currim, 2013).
As a financial metric, the ROMI influences the firm profits (Abramson, Currim & Sarin, 2005) and shareholder value (Schulze, Skiera & Wiesel, 2012) with the retrieved information. More specifically, measuring and monitoring the ROMI can help a company to make good decisions before, during and after budgeting (Klein & Swartzendruber, 2003) and can assist in improving the effectiveness of marketing by setting thresholds for screening the investments (Danaher & Rust, 1995; Klein & Swartzendruber, 2003). A marketer can discover campaigns that are consistently returning less than the threshold that was set and therefore can improve the tactic effectiveness. Contrary to other statements, Joshi & Hanssens (2010) argue that ROMI is able to justify investments with a long-term effect.
12 Figure 1. Types of Return on Marketing Investment (Stewart, 2009)
Another research by Rust, Lemon & Zeithaml (2004) presents a framework for return on marketing which enables a company to link marketing actions to financial returns. In this case, as in many others, “marketing is viewed as an investment that produces an improvement in a driver of customer equity” (Srivastava, Shervani & Fahey, 1998). The model is presented in Figure 2. Deriving from the model, there are two ways a marketing investment can move towards the return on marketing investment. The simplest way is only taking the costs of the investment into account. The other way goes via ‘improved customer perceptions’ to an ‘increased CLV’, which might lead to an increase in customer equity and finally return on marketing investment. This second way can be linked to the previous discussed customer equity and CLV, which are DFC metrics. In this sense, customer acquisition and customer retention are the two main indicators for an increase in customer lifetime value, causing a higher ROMI. To put it another way, one could argue that this model by Rust et al. (2004b) uses the DCF metrics to increase the ROMI.
Figure 2. Return on marketing model (Rust et al., 2004b)
13
financial impact of particular marketing cases, this model addresses the impact of competing marketing expenditures with a new CLV model.
Powell (2012) describes in his book the marketing-effectiveness continuum. This continuum helps marketers in the short and long-term to benchmark the organizations accountability and develop a plan to improve the ROMI and marketing effectiveness. Thus, the main purpose of this continuum is to improve marketing ROI by lowering costs and risks on a continuous basis. As can be seen in Figure 3, the marketing effectiveness continuum is divided into five levels: activity trackers, campaign measures, mix modelers, consumer analyzers and brand optimizers. In order to move up the continuum, structural changes in the organization must occur. For example, by trying to improve the data-gathering process continuously. This model only shows successes in the long-term when consistent, repeatable small steps are made to improve the infrastructure of the organization.
Figure 3. Marketing effectiveness continuum (Powell, 2012)
2.2.1 Benefits and drawbacks of ROMI
14
In almost all literature on marketing ROI, authors state that it is important to measure the marketing performance and effectiveness (Fu et al., 2018; Milichovsky & Simberova, 2015; O’Sullivan & Abela, 2007; Powell, 2012). According to Powell (2002) ROMI helps with the (1) conceptualization, (2) planning and budgeting, (3) communication, (4) prioritization, (5) gaining approval, (6) execution and management, and (7) monitoring and measuring the marketing programs. These factors can be seen as the benefits of the return on marketing investment. Powell (2012) adds that using ROMI will pay-off in different ways: by reducing inefficient spending; reallocation of spending; having a shorter time to process marketing; and with growth and profit optimization. Eventually, the main purpose for calculating ROMI is knowing the monetary value of the marketing campaigns. As a final note, ROMI makes it possible to steer on specific campaign investments to measure the individual effects of a campaign. This is an advantage for marketers who want to evaluate their own campaign performance.
However, the ROMI also has some drawbacks. Amber & Roberts (2008) list five objections for using the ROMI. First of all, there is no general definition of the word ‘ROI’. This means that there is no consistent meaning of ROI. According to Farrington (2004), you can ask ten different people to define ROI, and you will get ten different answers. Second, ROMI is seen as an investment, but it should be seen as a maintenance. Marketing is an ongoing process which creates sales and profits continuously. Therefore, it should be maintained, rather than be seen as an investment. Third, continuing on previous point, the ROMI assesses mostly short-term marketing performance. However, marketing can also have a long-short-term effect, which is not accounted for. Fourth, most of the bottom-line performance measures (e.g. discounted cash flow and return on customer) subtract expenditures whereas ROMI divides expenditures. This could cause a conflict in the economic value added (EVA). Finally, as an effect of the law of diminishing returns, the maximum ROMI is reached before the point of maximum profit is reached. Subsequently, this creates suboptimal levels of activity (Ambler & Roberts, 2008).
Several financial metrics to measure marketing performance have been discussed. Table 1 provides an overview of all the pros and cons of the different metrics. In relation, one could argue that the DCF, customer equity and ROC are all a subset of the ROMI. All these metrics can contribute to a more positive outcome of the ROMI.
15 Discounted Cash Flow Customer Equity Return On Customer Return On Investment
Pros • NPV is a good financial
basis
• CLV is good customer basis
• DCF gives a good perspective in the value of the company because you take into account the future cash flows.
• Companies can steer their entire marketing program on customer equity.
• Balance between acquisition and retention.
• Quality of the relationship with the customer is most important. • Allows for a tradeoff between
customer value, brand equity and customer relationship
management.
• Long-term value – reliable source of future revenue and profit.
• Customer perspective – customers of a company are scarce resources.
• More strategical influence • With the ROC a company can
not only harvest current profits, but it can also keep and supplement shares of customer capital.
• Measures brand value and finance – ensures accountability • Influence on company profit and
shareholder value. • Good short-term effect
• Very feasible in implementation. • Usage of the ROI ensures
reducing inefficient spending; reallocation of expenditure; shorter time to process marketing; growth and profit optimization.
• Able to steer on specific campaign investments
Cons • Lack of independence
• No guarantee if outcome is correct and reliable • Prognosis is no benchmark • Multiple prognoses are
reconciled to minimize variance
• Hard to differentiate future improvements
• Knowledge and experience is needed – complex tool
• Product sales, brand strength and short-term financial performance are secondary indicators of success.
• Complex to understand how to grow and manage customer equity.
• Companies might misunderstand the purpose of this marketing function or lack authority to actually pursue this.
• Disaggregated and aggregated approaches – hard to measure
• Prognosis is not comparable to actual performance –
prognosis is adapted to benchmark
• Future cash is treated the same as current cash in hand • Prognoses are not comparable
to each other (same period, other year)
• No common definition. • Long-term effect is hard to
measure.
• Not clear if ROI is an investment (or an enforcement)
• Parts of the expenditures can be in conflict with the Economic Value Added (EVA)
• Maximum ROI is reached before the maximum profit is reached.
16
2.3 Measuring online marketing performance
The metrics that have been discussed previously are mostly focusing on offline marketing performance. Yet, online marketing is booming. The transparency and interactivity of the internet force an organization to adopt a proactive-reactive attitude towards their online communications (Gurău, 2008). For example, NS benefits from using both offline and online marketing campaigns. So far, one of the questions that remained unanswered is: How can NS measure their online marketing performances effectively?
As already mentioned, a lot of businesses are using Internet for their commercial activities (Teo, 2005). In fact, Internet made it possible to personalize marketing to a global audience. In this sense, marketers should not only measure their offline marketing performances, but also their online marketing performance. Online marketing can have a big impact on the campaign, which should be monitored just like offline marketing campaign (Goodwin, 1999). The key question is how. Figure 4 shows that traditional marketing mostly creates awareness, but online marketing could take a consumer through the whole sales process.
Figure 4. Online marketing and sales process (Goodwin, 1999)
Online marketing has several metrics to measure marketing effectiveness: (1) click-through rates, (2) cost per click, (3) cost per impression, and (4) return on investment (Goodwin, 1999). All these metrics measure and record the consumers’ actions quite well. However, they provide very numerical data and therefore might miss the actual consumer perception and attitude towards buying the product or service online. Eventually, it is best to measure online marketing in the form of an ROI instead of DCF or customer equity, since a ROI give the best financial insights (Goodwin, 1999).
17
2.4 Usage of marketing metrics
Different studies on marketing performance measures and marketing effectiveness suggest implications for marketing managers (Mintz & Currim, 2013; O’Sullivan & Abela, 2007). Yet, these articles do not show the real usage of these marketing metrics which managers need to actually be able to implement a new or better measure for marketing effectiveness. For example, the study by O’Sullivan & Abela (2007) proposes that marketing performance should be measured by both financial and nonfinancial metrics, because the more comprehensive metrics are used, the higher the CEO’s satisfaction for marketing will be. Nonetheless, the authors mentioned above, do not give clear examples of the metrics that could be used and how these should be used. In sum, one could argue that the need for implementation is acknowledged, but the actual managerial execution of these marketing performance metrics is still an issue.
Another example can be found in a research by Mintz & Currim (2013). They provide five strategies to increase metrics’ usage within an organization. Mintz & Currim (2013) suggest that managers should get training on how to use these metrics. Another point raised by these authors is that managers from different functions like finance and accounting should be involved in the marketing mix decision making to increase the use of financial marketing metrics. Furthermore, top management could appoint a CMO to engage in top management decisions and make sure that financial marketing metrics are used. Accordingly, top management could also tie a managerial compensation to specific metrics. All of these strategies mentioned by Mintz & Currim (2013) are only useful if you know how to implement it in your organization. Similarly, in the example by O’Sullivan & Abela (2007), these strategies are also lacking the actual managerial advice for execution and implementation.
2.5 Adoption models
18
According to early work on marketing decision support systems (MDSS), it can be defined as “a coordinated collection of data, systems, tools and techniques with supporting software and hardware by which an organization gathers and interprets relevant information from business and environment and turns it into a basis for marketing action” (Little, 1979). MDSS assists the decision maker in identifying important decision variables, which leads to increased effectiveness of their decision making (Little, 1979; Van Bruggen, Smids & Wierenga, 1998). The MDSS made by Little (1979) provides insight into a manager’s decision making (Figure 5). An earlier article of Little (1970) states that a (marketing) model should be (1) simple, (2) robust, (3) adaptive, (4) easy to control, (5) complete on important issues and (6) easy to communicate with. These are important characteristics to keep in mind when setting up a MDSS.
Figure 5. Marketing Decision Support System (Little, 1979)
19
Thirdly, it is important to note that the involvement of automation in the MDSS, does not necessarily increase ones’ satisfaction towards the process. Furthermore, developing a new MDSS from outside the company has relatively the same impact on the success of implementation as updating existing software within the company. Finally, the MDSS should be adopted to obtain new information instead of upgrading existing information. For the implementation of a MDSS to be successful, user factors (e.g. age, education, experience) and implementation factors (e.g. customer developed or commercial package, user involvement, sophisticated system) are most important.
Figure 6. Conceptual framework of factors that affect adoption, use and satisfaction (Wierenga & Oude Ophuis, 1997)
20
Another model that can be used as adoption model is the ‘marketing dashboard’. A dashboard “brings the firm’s key metrics into a single display” (LaPoint, 2005). Such a marketing dashboard can help to make sure that everyone in the marketing department is on the same page to see successes and failures of their marketing (Pauwels et al., 2009). With the request of CEO’s and CFO’s for more marketing accountability, the dashboard will provide good insights of growth and costs (Rust et al., 2004a). A dashboard includes a collection of interconnected key performance metrics which communicate underlying performance as reflection of short- and long-term interests (Clark, Abela & Ambler, 2006; Lehmann & Reibstein, 2006; O’Sullivan & Abela, 2007; Pauwels et al., 2009; Wind, 2005).
Pauwels et al. (2009) developed a framework for adoption and success. This framework is the adapted and extended version of the decision support system framework made by Wierenga et al. (1999). Pauwels et al. (2009) framework can be used as both a marketing management support system (MMSS) and dashboard to support implementation (Figure 7). A MMSS can be defined as “any device combining (1) information technology, (2) analytical capabilities, (3) marketing data, and (4) marketing knowledge, made available to one or more marketing decision maker(s) to improve the quality of marketing management” (Wierenga & Van Bruggen, 2000).
21
The success of the framework is depending on the fit between the demand side and the supply side (Pauwels et al., 2009). The demand side is the side that needs support in the decision process, whereas the supply side provides the employed functionality of the management support systems (MSS) (Wierenga et al., 1999). The fit between these two sides is critical. There should be a match between the type of information in combination with the decisions of the users (demand) and the metrics which are important and useful for the company (supply). Basically, all the factors of the demand side should match the factors of the supply side in order to accomplish a good fit which can result in a successful overall framework. The implementation stage involves a lot of people and processes, which might result in mistakes throughout the process (Pauwels et al., 2009). One of the factors improving the implementation is top management (TM) support. Active support from top management is required for a successful adoption (Premkumar & Ramamurthy, 1995; Wierenga & Oude Ophuis, 1997). Another factor that influences the implementation is training. Research conducted by Frambach & Schillewaert (2002) support this claim. They state that usage of a system depends on internal marketing variables such as trainings. Furthermore, managers should be careful when implementing the MMSS or dashboard (Pauwels et al., 2009). If the fit between demand and supply and the implementation is good, the predisposition among the users should be look upon. In this predisposition stage, three factors are important (Pauwels et al., 2009). First of all, the attitude of the decision maker will only change if they see the usefulness of the dashboard. Therefore, these decision makers should be convinced that the dashboard will help them perform better. Second, the dashboard should give trust to the decision maker in the way that the numbers are reliable. Additionally, Lippert & Davis (2006) argue that there are two types of trust that should be considered, namely interpersonal trust and technology trust. An individual needs both types of trust in order to be prepared to experience a change. In an adoption process, positive trust during the implementation process will lead to better organizational (economic) successes (Lippert & Davis, 2006). A final note is made on the management of expectations. High expectations are helpful in generating initial usage of the dashboard, but could backfire when the experience does not meet the expectations. Low expectations of the decision makers reduce acceptance (Pauwels et al., 2009). The predisposition stage accounts for the attitudinal, behavioral and cognitive changes in de decision making towards the dashboards. The last stage, adoption and success is measured on four criteria:
22
departments. Trials can help to take away unrealistic expectations of the decision makers.
2. Increased accountability – this is one of the key purposes of using a dashboard
3. Improved effectiveness and efficiency of marketing efforts – marketing efforts will be better monitored and deployed
4. Learning (explicit or implicit) – dashboards help to better understand the marketing processes which are useful for the business.
Eventually, the MMSS or dashboard should be judged on the additional profit it generates (Wierenga et al., 1999). A dashboard can fill the gap between theory and practice (LaPointe, 2005; Lehmann, 2005; Reibstein, Norton, Joshi & Farris, 2005).
2.6 Conceptual model
The research concept is derived from the MMSS framework by Pauwels et al. (2009). As can be seen in Figure 8, the fit between the user and metric has a direct effect on the adoption success of a model. If there is a good fit, the effect on adoption will be positive. Resulting in a successful adoption which leads to increased accountability and improved effectiveness and efficiency. This main effect can be influenced by internal support and by predisposition. In the case of internal support, a good support system will have a positive influence on the main effect. This internal support should not only come from top management, but it can also be integrated by, for example, training and user involvement. When looking at the predisposition, the company should take into account that good attitude, positive trust and reasonable expectations can complement the relationship between the user-metric fit and successful adoption. Predisposition is seen as a moderator effect since there is an interaction effect. This means that there is a direct, separate effect, in which the predisposition influences the main effect.
23
2.7 Literature summary
24
3. Method
3.1 Research type
The basis for this research is a case study. A case study research examines data within a specific context (Zainal, 2007). According to Yin (1984) a case study research method is defined as “an empirical inquiry that investigates a contemporary phenomenon within its real-life context; when the boundaries between phenomenon and context are not clearly evident; and in which multiple sources of evidence are used.” A case study has a few advantages (Yin, 1984; Zainal, 2007). First, the data is collected to support the context of its use. In other words, the collection of the data is based on the specific situation in which exploration is needed. Second, case studies have a broad range for their research and allow for both quantitative and qualitative data analysis. However, for this research only qualitative data is used. Finally, besides exploring or describing the situational data, case studies also help to explain more complex situation which might not be captured by experimental or survey research. Basically, case studies can capture the underlying explanations of particular situations. There are also some drawbacks, which have been taken into account to ensure the validity of the research (Yin, 1984; Zainal, 2007). Case studies are often biased, because they have a lack of strictness. Furthermore, due to the small number of subjects, scientific generalization is often not possible. Finally, case studies produce a lot of documentation which makes it harder to manage and organize in a systematical way. In consideration of all the advantages and disadvantages, a case study has been found the best way to explore the interests of the researcher in the conducted data (Zainal, 2007). In exploratory research, the situation under evaluation has no clear, single set of outcomes (Baxter & Jack, 2008; Yin, 2003). Most of the time, exploratory research has little or no previous research on the situation at hand (Brown, 2006), which is the case according to the literature review.
3.1.1 Population
25
approach. These companies have been selected based on the target description. Such an external company needed to meet two specific criteria: (1) active in conducting marketing communication campaigns, and (2) have some metric to measure the effect of these campaigns. External companies were meant to serve as additional input in order to compare the different strategies. For example, the marketing metric of NS (i.e. ROMI) was compared to the marketing metrics of Schiphol Group. Thus, two researches were examined: a NS case study (see next chapter) and a side study on a few external companies to benchmark NS’ marketing performance strategy against.
3.1.2 Representatively
As already mentioned, a case study has a bigger potential to be biased (Zainal, 2007). Especially when the researcher delivers sloppy work or interpret the data wrongly, which influences the direction of the findings and conclusions in a certain way. One could argue that this research is not representative for the entire population because it lacks reliability. However, the researcher took into account the design of the procedure to reduce the research method bias (Podsakoff, MacKenzie, Lee & Podsakoff, 2003), and therefore is representative. An example of this procedure design is protecting anonymity and reducing the evaluation apprehension, which means that the respondent should answer the questions as honest as possible, where there is no right or wrong. Another example is that the order of the questions should be counterbalanced in order to neutralize potential priming or bias effects (Podsakoff et al., 2003).
Furthermore, the several tips have been taken into account to control for method biases (Regoniel, 2013), and increase the validity and representativeness: random selection of the respondents; no manipulation of the original results; pre-interview with a non-respondent group validates the interview questions; respondents need to be willing to be interviewed; ask clear questions; prepare every single interview and make sure to keep the interviews within an one-hour limit. Most importantly, the researcher had a neutral-mindset and was objective during every interview.
3.2 Data collection
26
In-depth interviews are an useful qualitative technique. Besides, it is an appropriate data collection method when the researcher wants to ask open-ended questions which evokes the depth of information from a small number of respondents (Guion, Diehl & McDonald, 2001). An in-depth interview has a few key characteristics: (1) Open-ended questions, (2) semi-structured format, (3) seek understanding and interpretation, and (4) recording responses. Thus, an interview is not only about asking questions, but it is also about getting the deeper meaning of the answers the respondents gives (Guion et al., 2001).
Both internal interviews and external interviews had their own interview set-up with pre-planned key questions which were slightly different from each another (see Appendix II). The internal interviews were intended to give more background about the needs of the stakeholders and discuss the potential solutions. Part of the internal interviews was the framework of Pauwels et al. (2009), which provided support for a discussion about necessary improvements within the marketing department. The external interviews gave more insight in how other companies established their marketing metrics, which can be used as learnings in this case study. The semi-structured format of both types of interviews allowed the interviewer to make the interview more conversational by being able to specify certain questions.
According to an article by Guion et al. (2001) a researcher should take into account the following skills when conducting an interview: be open-minded, be flexible and responsive, be patient, be observant, and be a good listener. These skills can be attributed by full attention of the researcher, paraphrasing to confirm what the interviewee is saying and reflecting the emotions of the message (Guion et al., 2001). Since the researcher lacked initial interviewer experience, she took notes of points to keep in mind on every interview set up.
27
3.2.1 Stages for conducting an in-depth interview
Kvale (1996) describes seven stages which help to conduct an in-depth interview (Figure 9). The researcher went through all these stages to make sure the research had the optimal set up. In the first stage, thematizing clarified the general purpose of the interviews. The second stage determined the design to maintain consistency, stay on track during the interview and focus on exploring different topics. The interview questions were based on the literature about marketing performance metrics and adoption models. Moreover, since the MMSS framework by Pauwels et al. (2009) is used in this research as the theoretical basis for adoption, it was discussed in every interview. More general questions originated from input of NS to get more insight in the stakeholder’s point of view. Based on this, different interviews were made for every type of stakeholder (i.e. marketer, KMM analyst and manager). During the interviews (third stage), the researcher considered the before mentioned skills. Furthermore, at the beginning of every interview, the purpose of the interview was explained and permission was asked to record the session. All of the interviews were recorded with a recording app on a mobile phone. The fourth stage included the transcription of the interview. Before the study began, the researcher decided to transcribe the interviews within a week after the actual interview took place. The transcription was relatively extensive and coding was included before the data could be analyzed (fifth stage). Coding of the transcriptions was done with Atlas.ti. Initially, the four most relevant interviews were used as a starting point, in which most codes were made (via open coding). Subsequently, mainly axial coding was used. This meant that interviews only received a new code if it was relevant new information. In this way, codes could be easily grouped and a good overview could be maintained. A side effect is, however, that the codes might be more biased because the researcher could have interpreted the information given by the respondent in the wrong way. On average, each interview had 26 codes. The next stage verifies the information by triangulation. This meant that multiple perspectives are taken into consideration to examine the outcomes. These outcomes were supported by specific propositions, which increase the feasibility of the study (Baxter & Jack, 2008). In the final stage results were reported and shared with internal and external stakeholders.
28
3.2.2 Pros and cons of in-depth interviews
In-depth interviews have three big advantages (Stokes & Bergin, 2006), namely: (1) unique circumstances are applicable which are useful for sensitive or personal topics (Robson & Foster, 1989), (2) researchers have greater control over respondent selection and as a consequence, the process is more flexible (Cassell & Symon, 2004), and (3) comprehensiveness and depth of information resulting from the interviews account for a more preferential outcome (Hedges, 1985). On the contrary, there is one main disadvantage of in-depth interviews, only a small group of stakeholders or consumers is interviewed and therefore the researcher misses out on the opportunity to interact with other interesting consumers (Stokes & Bergin, 2006).
3.3 Sample method and size
Case studies in general have a small sample size, in which respondents are not selected randomly. This could lead to biased outcomes. However, different methods have been taken into consideration to reduce these biases and increase reliability. Therefore, there is argued that the small sample size is suitable for this research. The sample is divided into two groups: internal and external. Internally, different stakeholders with direct relationships to the subject of research were asked to participate in the interviews. The assignment was done by the researcher, in consultation with the KMM manager Josien Salomé and other KMM co-workers. In total, ten NS stakeholders (i.e. top management, KMM and marketers) participated in the in-depth interviews. These stakeholders were selected on their professional interest in marketing ROI metrics and their knowledge about it. Externally, the assignment for relevant companies was supposed to be done by MIcompany. However, since the interview with MIcompany (developer of the Braintrain) was postponed, other internal networks were consulted. Via other NS co-worker’s emails were send to potential external companies with an explanation of the research. The companies who replied positively had been invited for the in-depth interview. Overall, six companies were contacted2, of which four replied and eventually three companies took part in the in-depth interviews. The companies who participated were Schiphol Group, KPN and MIcompany.
29
4. Case study – NS
One could argue that marketing at NS is not necessary because the company has a social purpose. However, this statement is not true. The marketing campaigns of NS focus on attracting more people into traveling by train. The campaigns of NS are evaluated on the incremental revenue that the marketing returns. Therefore, the main question that NS is repeating every time is: “Without this marketing action, would our consumer still have made the same train ride?” In other words, did the marketing make an impact on the consumer? A good example of this is the round-trip ticket sold by Kruidvat, Etos and AH. These tickets are fairly cheap for a long train ride in comparison to buying a normal ticket for such a trip. If a consumer buys this ticket instead of the normal ticket, the marketing action obviously worked. Nevertheless, the main question still remains unanswered, is the trip additional or would the consumer have made the trip either way? NS seeks to capture these additional trips.
To start with, NS measures their performances in four different travel motive segments: social recreational consumers, business consumers, work consumers and residential-school consumers (i.e. students) (Salomé & Traas, 2017). The segments are reasonably self-explaining, however social recreational consumers are divided into two groups, low- and high frequent. Social recreational low are the consumers who travel without a NS subscription mostly without an OV chipcard. The consumers in the social recreational high segment do have a subscription on their NS card. Appendix III gives a broader description of the different segments, including their facts (e.g. number of customers in this group and revenue per customer).
4.1 Current marketing effectiveness measurement at NS
30 Figure 10. NS’ continuous learning process (Salomé & Traas, 2017)
4.2 The Braintrain
With the help of the ROI, NS is able to evaluate if a campaign generates profit and the amount of incremental revenue that the campaign effort returns. As already mentioned, this is measured in the ‘Braintrain’, a tool in which customer data is processed and is used for the ROI calculation. To put it in the simplest way, the calculation that the Braintrain uses is the following: ROI = (incremental revenue – marginal costs – commercial costs) / commercial costs
31
32
5. Results
The ten internal interviews with NS stakeholders (i.e. marketers of different marketing departments, KMM analysts and members of the management team) gave more insight in the thoughts about the adoption and usage of the ROMI (for transcripts and coding, see Appendix V, VII & VIII). The internal interviews were divided into five major sub categories, of which the results will be independently discussed in this section. Moreover, the results of the external interviews with Schiphol Group, KPN and MIcompany will be examined.
5.1 General
Each interview contained a preliminary question, which provided insights into the experience of the stakeholders in working with the ROI metric. Half of the respondents (n=5) does not work with the ROI on a regular base and are therefore grouped as ‘no experience’. Interestingly, this were mostly marketers. The respondents who work on a regular base with the ROI (n=5), are the KMM analysts and top management. Herein, KMM analysts mostly calculate the ROI for a marketer and top management has to monitor the pre-ROI and post-ROI.
33
Positive aspects current ROI Negative aspects current ROI
Measuring incremental revenues (n=6) Measure more than revenue minus costs – take incremental revenue into account
Biased (n=7) due to the way incremental revenue is determined (n=6)
Incremental revenue is determined on the basis of research which is not fact based
Gives an integral picture (n=5) Does not take customer satisfaction into account (n=4)
Takes into account several separate costs (n=3) (e.g. different kilometer costs for off-peak hours and rush-hours, withdrawal costs etc.)
Not workable for online marketing campaigns (n=2)
Other metrics are used to measure online marketing, not ROI
Learning effect (n=2) Complex method (n=2)
Gives control (n=1) Traditional, time consuming and laborious (n=1) Creates awareness (n=1) To little steering from top management (n=1) Translates costs to revenue (n=1) Calculation too simplistic (n=1)
Table 2. Positive and negative aspects of the current ROI calculation at NS
5.2 Stakeholders
As already mentioned, three different types of stakeholders were considered for this case study. Every stakeholder group will be discussed separately, since the different groups had specific respondent related questions.
5.2.1 Marketer
34
Remarkable is one of the comments by a marketer who said: “It sounds a bit weird, but I think that we might need a cultural change within NS if we want to motivate people to work more fact based” (Interview 8). This marketer explained that, in his opinion, there is not enough awareness within NS on how much money is spend on marketing and campaigns. The large budgets do not trigger the marketers to think about the amount of money they spend on a campaign, unsure of the profitability of that campaign. More awareness should be raised for tight control on revenue and more importantly costs.
5.2.2 KMM analysts
The three KMM analysts who deal with ROI calculations the most, were interviewed. KMM analysts described their ideal ROI situation as one in which the marketer performs the ROI and where KMM has a more advisory and controlling role (n=2). Marketers agreed with this setting if they are able to develop more ROI knowledge. They also noticed that KMM is missing capacity and ROI’s are low on their priority list. Hence, marketers have to wait a longer time to get their pre-ROI or post-ROI, which is not optimal. From the analysts’ point of view, a marketer needs more consensus of ROI’s and top management support in order to be able to work independently on a ROI (n=3). In other words, top management should steeron consensus within the marketing department for using the ROI. Another motivation for marketers to work independently on a ROI could be their intrinsic willingness to learn and to be more data driven (n=2). This can be externally supported with ‘learn by punishment’, as one of the analysts called it. It means that a marketer has to be more aware of the data and deliver upon this, otherwise the campaign could be cancelled. Thus, if they do not learn, they will be ‘punished’. KMM notified that some marketers miss the affinity and intrinsic drive to work with ROI’s, where others are more fact based. Analysts spoke out their concerns about this. In their opinion a marketer should always be interested in the results of their campaign. Besides, the future marketer needs to have more analytical skill to prevent a ‘talent gap’ (Leeflang, Verhoef, Dahlström & Freundt, 2014). Finally, simplified tooling should help marketers to make it easier to perform an ROI independently (n=2). In conclusion, one could argue that both intrinsic and extrinsic motivation is needed in order for marketers to start working with the ROI.
5.2.3 Top management (or MT)
35
management team is not focusing enough on the ROI. However, they know that if they want the ROI to have more impact, they should pay more attention to it. For example, with an official ‘go’ on an approved ROI.
5.3 Adoption and change processes
Inconsistently, respondents were questioned about how they usually change processes within their own team and within multiple teams. In general, the best way to adopt change is top-down (n=7), instead of bottom-up (n=3). It should be highlighted that adopting change in someone’s own team is fairly easy and works good with a bottom-up approach. But, if it involves multiple (disciplinary) teams, it is advised to work top-down. NS is a large company and therefore a stakeholder needs to discuss potential changes with top management first. However, it is also important to have supporters of your idea in your close environment to propagate this plan. The top-down approach is also one of the most difficult tasks in the adoption process. Respondents acknowledge that top management should support change in order to let the process run smoothly (n=2). Furthermore, they also indicate that expectation management of the different stakeholders who are involved is really hard and very time consuming (n=2). All the stakeholders have a different point of view, in which everyone wants the best for their own team and/or function. Finding common ground between all the stakeholders is one of the largest obstacles in an adoption process. One could argue that this holds stakeholders back from changing something in the marketing department. Eventually, large adoption processes are postponed because people lack the energy (n=2) or the time/capacity (n=2) to work on something impactful. In case of the former, one analyst stated: “if I’m truly convinced by the value of it, I will go for it, otherwise it will be too much of an energy drainer” (interview 3). Thus, someone will think twice before they start such a procedure and will only change something if it has value for themselves. This is partly caused by the expectation management of multiple stakeholders, which was discussed earlier. Besides all, most respondents are open for adoption of change (n=5). Yet, they might not want to implement the change themselves.
5.4 Marketing Management Support System (MMSS)
36
type of stakeholder. Due to this disagreement, one could argue that implications in both supply and demand need to be better aligned for a more optimal fit. Table 3 summarizes the feedback given by the respondents for the supply side and demand side. This also shows that the supply/demand fit is not optimal.
Demand side Supply side
Users – a lot of different marketers (n=5) Sophistication – make it more user friendly (n=4)
Metric – need to use more actual data (n=3) Metric – not intuitive enough (n=1)
Metric – create more awareness for metric/tool (n=1)
Table 3. Feedback demand side and supply side
37
case of the IT department, multiple respondents mentioned that if the ROI tool needs to be changed, then the IT department of NS is no option to go to because their schedule is completely full till next year. Summarizing, the implementation moderator needs optimization in order to have a positive effect on the success of adoption.
Needs to be improved Nothing has to change Not mentioned
Top management support (n=5) (n=5)
User involvement (n=8) (n=2)
Communication (n=3) (n=4) (n=3)
Introduction and training (n=4) (n=3) (n=3)
IT department (n=3) (n=7)
Table 4. Overview number of respondents per implementation topic
38
Attitude Trust Expectation
ROI is seen as something that needs to be done, but is not fun (n=5)
ROI is biased, needs to be taken out to create more trust (n=6)
- Could be done by working more fact based → with a control group for example (n=2)
Expectations are low (n=2)
Low affinity level with data (not data driven) (n=4)
Not sure if the methodology is the right one (n=3)
In general, is good, no change needed (n=2)
Missing direct steering from top management (n=3)
In general, is good, no change needed (n=3)
Truthful results – not always the case (n=1)
Needs more consciousness (n=1)
In need of a better explanation about the tool and the goals of using ROI (n=1)
Table 5. Feedback predisposition
Of the respondents who were asked about the most important factor of the MMSS framework to change (n=4), predisposition was named most (n=2), followed by implementation (n=1) and demand side (n=1). Eventually, one could argue that all the framework elements are related to each other and therefore all need to improve in order for the adoption to be successful.
5.5 Effective and efficient ROI
39
themselves will give them more knowledge about the revenues and costs of the (proposed) campaign.
Efficiency Effectivity
Marketer performs ROI, KMM supervises / final control (n=6)
More (intrinsic) motivation and urgency in using ROI’s with marketers (n=4)
More automated tool (n=4) ROMI as target imposed by top management (n=3)
More awareness of standardized processes (n=2) If you make ROI a standard process, make sure everyone knows it and uses it.
Less complex is better (n=2)
100% alignment of model with the targets (n=1) Table 6. Feedback ROI efficiency and effectivity
As a final question, respondents were asked to describe how they think the marketing department should implement the ROI. Half of the respondents cited to improve the user friendliness of the Braintrain and make it more trustworthy (n=5). Besides, more steering and control is needed from top management (n=5). According to a marketer, management is able to create more value for the person who needs to perform the ROI. If there is a strict ‘go or no go’ moment, people will be inclined to work with the ROI, otherwise they are not allowed to continue with the campaign (interview 7). Furthermore, creating more involvement (n=3) and more data driven marketers (n=3) will help with the department wide adoption of ROI. Another point raised, was about getting more insights in long-term effects (n=2). As an example, this could be done by focusing on customer satisfaction and CLV, which are based on the value of the customer, measured over a longer period. Finally, a more intuitive ROI tool could contribute to a better adoption (n=1). Summarizing, there are several points of improvements to help the department wide adoption of ROI.
5.6 External results
40
yourself what you are actually measuring.” Schiphol Group differentiates on customer segmentation, product segmentation, seasonality, diversity and their targets. For them, customer satisfaction is very important, but eventually every campaign is judged on their revenue. Hence, if the customer is happy, one will also see this back in the revenues.
The second external interview was an interview by phone with a wholesale marketer from KPN. Within the organization of KPN, ROI calculation is very important. This creates awareness for everyone. KPN wants to know about their revenues, but also the costs of a campaign are considered. Therefore, a marketer has to calculate a ROI for every campaign. The finance departments check these ROI’s. Interestingly, not every marketing department works with the same ROI calculation and implementation. For example, B2B marketing works with official ‘go’s’, in which different managers have to approve on a specific campaign based on the pre-ROI outcome. This is done via some sort of application form, which is send around to all the stakeholders that are involved. Alternatively, the B2C marketing department is less strict on approvals and only the manager of the team has to check the ROI. According to the wholesale marketer, it is best to implement an ROI top-down: “I think that it should be obligated at first by management, before you can become looser. Otherwise the marketers will not see the importance of working with the ROI.” Moreover, acquisition campaigns should have a short-term focus in measuring marketing performance, whereas retention and loyalty campaigns should focus on the long-term. In this, there is more attention for customer satisfaction and CLV.
41
implementation of a possible new adoption of ROI. MIcompany gave some final advice for NS to take into account in changing the ROI process:
1. Develop an easier ROI tool, to enable marketers to calculate their own ROI. This promotes marketers to think about the outcome of their campaign.
2. Set specific targets on the ROI (or apply it in the methodology)
3. Make use of use-cases to check for value and work on these cases in agile teams. This will account for a lasting solution and can provide an intensive training for users. 4. Challenge: replace data assumptions by fact based data
42
6. Theoretical model adjustments
Van Bruggen, Smidts & Wierenga (1998) state that a support system is more effective when factors are explicitly determined. Therefore, the MMSS framework by Pauwels et al. (2009) is adjusted according to the empirical data of this case study. The conceptual model (Figure 8, p.22) proposes that there should be a fit between users and metric, which is directly related to the adoption success. However, the results show that there is a distinction between data driven marketers and creative marketers. On the one hand, data driven marketers identify a fit with the usage and sophistication of the Braintrain. On the other hand, the metric is not aligned with the needs of the creative marketer. These marketers argue that the tool should be more user-friendly, which could be accomplished by automation of the Braintrain. Conversely, Wierenga & Oude Ophuis (1997) argue that automation does not inevitably increase the satisfaction towards a tool or process. In any case, the results indicate that there is no direct effect of the user-metric fit on the adoption success. Respondents declare that an user-metric fit does not necessarily result in an adoption success. Other factors are required to have a positive effect on the adoption. More precisely, this effect is mediated by the internal support and predisposition. Thus, in contrast to the conceptual model, the user-metric fit only has an indirect effect on the adoption success via internal support and/or predisposition. In this sense, there is another major conceptual model adjustment, in which internal support and predisposition were seen as moderators. In line with the framework by Pauwels et al. (2009), predisposition was already displayed as a mediator. Thus, current research can confirm this relationship. However, the effect of the internal support changed. It should be emphasized that empirical data shows a direct dependency of the internal support factors on adoption success, which results in a mediation effect.
43 Figure 11. New MMSS model for adoption success
Wierenga & Oude Ophuis (1997) argue that active top management support is required to adopt a marketing decision support system (MDSS) within an organization. Premkumar & Ramamurthy (1995) support this claim with the finding that top management support is a requirement for adoption success. This indicates that there is a direct effect of top management support on the success of adoption. The results of the NS case study confirm this relationship. An analyst reported: “There should be more top management support to work independently on a ROI”. However, the results also suggest that the effect is moderated by the organizations’ culture. Managers have to steer on working with the tool and set targets on the outcomes. But, if the organizational culture is not aligned with these targets (e.g. targets, organization KPI’s and vision should be intertwined), it will be hard to actively support a MDSS. Therefore, it is proposed that organizational culture influences the direct relationship between top management support and adoption success. Hence:
P1 – Top management has a direct effect on adoption success which is moderated by the organizational culture.
