“The past, the present, and the new payment future...”

(1)

“The past, the present, and the new payment future...”

An investigation into the optimal design of a mobile payment application, and its effect on

the usage intentions.

(2)

2

By

Merlinde Weijanne Warners

August 2013

University of Groningen

Faculty of Economics and Business

MSc Business Administration

Marketing Management & Research

Paterswoldseweg 89a 9718 BC Groningen 06-100 86 198 mwwarners@gmail.com Student number: s1615203 Supervisor: Dr. M.C. Non Second supervisor: Dr. H. Risselada

(3)

3

Management summary

Given the consumers’ need for convenient and timely payment, and the high penetration rate of smartphones in the Netherlands, mobile payment is expected to become an important potential channel for conducting point of sale payments (Sharma, 2011; Yang et al., 2012; Zhou, 2013). The adoption of mobile payment applications is critical for both service providers and investors to profit from such an innovation (Yang et al., 2012). However, limited literature has been published on the external variables for acceptance of mobile payment systems (Meharia, 2012). Therefore, this study attempts to identify the elements that affect the adoption of mobile payment applications. Understanding how these elements affect consumers’ adoption can help companies in creating and designing an optimal mobile payment application, which can give competitive advantage when entering the Dutch mobile payment market.

(4)

4 Although the derived utility gives insightful information on how the various attributes affect the consumers’ choice among different mobile payment applications, it does not tell how these attributes eventually encourage or discourage the intention to use a mobile payment application. Therefore, the second model estimates the intention to use a mobile payment application with the help of an ordered multinomial logit method. The findings indicate that direct payment, an extra level of authorization, an insurance against loss and fraud and charging no fees all significant positively affect the intention to use a mobile payment application. Especially, for high risk-averse consumers, risk relievers like an extra level of authorization and an insurance against loss and fraud are found to be effective for increasing the intention to use a mobile payment application.

This paper describes examples to illustrate how to best target the different segments. Furthermore, theoretical and practical implications of the outcomes are presented which are helpful for companies who are aiming to step into the mobile payment market in the Netherlands.

(5)

5

Preface

The master thesis that is lying in front of you is written to complete the Master Marketing Management and Marketing Research within the study Business Administration at the University of Groningen. While adding the finishing touches to this preface, I am realizing that by handing in this thesis the great time I had as a student finally comes to an end. When I started with the Bachelor Business Administration in 2006, I was unsure about what I liked most in business. Nevertheless, after an inspiring time at the University of Groningen, I eventually found what satisfied my interests most: Marketing, which is also the research domain of this thesis.

I would like to use this preface to thank some persons who made it possible for me to graduate. First of all I especially would like to thank my supervisor dr. Marielle C. Non for all her clear advices, helpful feedback on my writing, her excellent statistical insights and her enthusiasm while working together. Although her busy schedule, she always was willing to answer my questions, which I really appreciated. I would also like to thank my second supervisor dr. Hans Risselada for his suggestions on my thesis. In addition, I would like to thank my brother, family and friends for their help during this thesis, and my thesis group for making the long days at the University little easier and more fun!

However, above all, I would like to thank my parents for giving me the chance to study at the University of Groningen. They never lost faith in me, and their love and everlasting support helped me to become the person I am today. I am grateful for that!

Merlinde Weijanne Warners

(6)

6

Chapter 1 Introduction

“Think about the things you have in your pockets and purses… Probably your mobile phone will replace all these products in the near future: traveling with your mobile device, identifying yourself with your mobile device, wireless charging of your mobile device, measuring your health with your mobile device, starting your car with your mobile device, unlocking your front door with your mobile device, and even payment with your mobile device. But are these applications of mobile devices fiction or future reality? Probably the latter. Fifty years ago, the tools and devices we rely upon today were only science fiction. One thing is for sure: the technological developments are already far established, it is only the search for the right design and standards before the applications can actually be launched in the mobile market”.

1.1 Background

It is almost impossible to think of a world without Internet. Since the introduction of the Web 1.0, consumers all over the world were able to interact with both companies and each other on the multimedia platform we now know as the World Wide Web (Berthon et al, 2012). During the introduction of the Web 1.0, companies rushed to have an Internet presence, however the company websites contained not much more than just an online corporate brochure (Berthon et al., 2012). Over the last decade, the online presence on the Internet has evolved due to the advent of the Web 2.0, to online coordination, communication, and electronic commerce. Opposed to Web 1.0 content is spanning from entertainment and online shopping to education, social networking and business (Hawamdeh and Kim, 2011; Stephen and Galak, 2012; Berthon et al., 2012).

(10)

10 With the technological revolution of the Internet at its peak, the pace of technological developments also increases (Day, 2011). Marketers are currently being challenged to keep their strategies matched with the disruptive effects of these new technologies and the deluge of new channels to choose from. Simultaneously, marketers must understand the more technology-empowered customer (Day, 2011). For instance, the adoption of wireless and portable mobile devices, and the usage of mobile web functionalities had increased considerably the last couple of years (Troutman and Timpson, 2008). Mobile devices, and especially smartphones (mobile phones which have a mobile operating system and the ability to run applications), are more powerful than previous generation personal computers, and have become an important part of everyday life activities of consumers (Wilska, 2003; Troutman and Trimpson, 2008). Since mobile phones accompany consumers everywhere they go, mobile services can be delivered by a company anytime and anywhere, rather than to a fixed home or office as is the case with e-commerce (Henning-Thurau et al., 2010; Abdullah et al., 2011). Consequently, the ubiquitous presence of mobile devices and their expanding capabilities present innovative opportunities for marketers to engage in a new virtual marketplace, to attract (new) customers, improve shopping experiences and drive sales (Mort et al., 2000; Burt, 2013). Therefore, companies and service providers increasingly enter the field of mobile commerce with a variety of mobile services, such as mobile instant messaging, mobile games, mobile banking, location based services and other mobile applications which can facilitate day to day activities of users (Zhou, 2013). Mobile commerce, or m-commerce, refers to “any transaction, involving the transfer of ownership or rights to use goods and services, which is initiated and / or completed by using mobile access to computer-mediated networks (e.g. Internet) with the help of mobile devices” (Tiwari and Buse, 2007). It is challenging for companies to reach consumers on their most personal communication device and to recognize situations where mobile services and mobile usage complement, or substitute consumers’ existing channels and devices.

(11)

11 Given the users’ need for convenient and timely payment and the high penetration rate of smartphones, mobile payment is expected to become an important potential channel for conducting financial transactions (Sharma, 2011; Yang et al., 2012; Zhou, 2013). Nevertheless, research revealed that consumers use their mobile devices to gather data about their planned purchase, however when it comes to the actual transaction, they shift to other more preferred traditional payment methods (Quinton, 2012). In other words, even though companies put effort in promoting and offering mobile payment options, there is an absence of widespread consumer acceptance of using a mobile device to conduct payments, resulting in a lag of adoption for this new technology as an alternative form of payment (Sharma, 2011). Therefore, the big question for companies is (Burt, 2013): “How can we move consumers in viewing mobile payment as an essential and basic requirement in the shopping process?”

Mobile payment is rather uncharted territory, and limited literature has been published on the external variables for acceptance of mobile payment systems (Meharia, 2012). The scarcity of information on what consumers want from mobile payment keeps companies from making vital decisions, such as which technology standard to adopt and what type of mobile payment application to offer (Hayashi, 2012). In addition, from a consumer point of view, the first research results show that consumers are unsure about the value of mobile payment relative to traditional payment methods (Hayashi, 2012). Therefore, in order to overcome both supply side and demand side barriers, it is necessary to get a good view of consumer mobile payment preferences, and mobile payment functionalities, including the design of mobile payment applications (Heil et al., 2010; Burt, 2013). As McGuire (2001) stated: “How well, or poor, applications are designed will be critical to speeding this intentions”. Understanding how consumers are affected by mobile payment application elements can help companies in creating and designing an optimal mobile payment application, which can give competitive advantage when entering the mobile payment market.

1.2 Goal and relevance of research

(12)

12 create the design of a mobile payment application and that might influence the utility derived from a mobile payment, and eventually the intention to use a mobile payment application. At this moment, there is an abundance of initiatives in the field of mobile payment. For example Google is running a pilot in the United States with a mobile payment application that makes it possible to e-mail money, Apple and Ahold are developing its own mobile wallet application, and the German company Payleven created an external Bluetooth pin-reader that makes mobile payment possible (Zonneveld, 2013). However, the consumer benefits from just one or a few safe, simple, and relevant mobile payment applications (Meharia, 2012; TNS Nipo, 2013). By making a synthesis of existing findings and investigating the optimal design of a mobile payment application, this research contributes to the existing literature, and helps companies in their development of the optimal mobile payment application.

1.3 Problem definition and research questions

When looking at previous research, it becomes clear that mobile payment applications differ due to the offered payment system, payment option, payment fees, payment protection and payment insurance (Hayashi, 2012; Yang et al., 2012; Zhou, 2013). This research will investigate different levels of these mobile payment elements that create the design of a mobile payment application. The effect of these different levels will be measured on two dependent variables: I) the utility that will be derived from a mobile payment application and II) the intention to use a mobile payment application. Hence, the main question for this research thesis will be:

“What is the optimal design of a mobile payment application and which mobile payment attributes have the highest influence on the consumers’ utility and intention to use?”

In addition, the following sub-questions are proposed:

 What is the influence of the type of payment system on, A) the utility that will be derived from a mobile payment application, and B) the intention to use a mobile payment application?

 What is the influence of the type of payment option on, A) the utility that will be derived from a mobile payment application, and B) the intention to use a mobile payment application?

 What is the (negative) influence of payment fees on the utility that will be derived from a mobile payment application?

 What is the influence of payment protection on, A) the utility that will be derived from a mobile payment application, and B) the intention to use a mobile payment application?

(13)

13 Before launching a new mobile payment application in the market, it is important for companies to have a clear understanding of which consumers to target, to know what mobile payment preferences the consumers in this segment have, and how to target them (Kotler et al., 2008). To make this research complete, it will be explored whether different segments of consumers can be distinguished based on their different mobile payment application design preferences. This segmentation will be based on the principles of marketing (Kotler et al., 2008), which means that segmentation can be done on a basis of demographics, behavior and psychographics. From this the last sub-question for this research is formulated:

 Can different segments of consumers be distinguished with respect to their mobile payment design preferences?

1.4 Research boundaries

The focus of this research will be specifically on the Dutch mobile payment market. Some countries, like the Netherlands, do have great potential for mobile payment initiatives, but lag somewhat behind in the development of mobile payment. Although the three biggest Dutch bank companies (Rabobank, ING and ABN Amro) recently announced that they will run a mobile payment pilot to test mobile payment usage among their Dutch clients (Hoexum, 2013), payment with the help of a mobile device at the point-of-sale is not possible yet. According to TNS Nipo (2013), the percentage of smartphone users in the Netherlands is higher than the European average, making it one of the leading countries regarding the number of smartphone users. Without smartphones, there is no market for mobile payment applications, since a wireless connection to the Internet is a precondition for making a payment with a mobile device. The attractive smartphone market in the Netherlands creates a perfect place for introducing a new innovative mobile commerce service like mobile payment. To conclude, Jeroen Verrijdt, director of the payment and savings department of Rabobank mentioned: “We are on the eve of new payment technologies. Checkout with a mobile device will replace cash and debit cards quickly in the upcoming years” (Telegraaf, 2013). Therefore, the question is not “if” Dutch consumers will use mobile payment in the future, but especially, “when” and “how” consumers are going to conduct payments with their mobile device (TNS Nipo, 2013).

1.5 Structure of this research paper

(14)

(15)

(16)

16

Chapter 2 Literature review

This chapter gives an overview of the literature regarding the theoretical foundations of mobile payment. The chapter starts by a definition of mobile payment, and the dependent variables will be described in further detail. Afterwards, the independent and psychographic variables will be described in further detail. In order to build a conceptual model, the independent and psychographic variables of this study are concluded with hypotheses about what the expected influence on mobile payment utility and usage intention will be. Lastly, the chapter concludes with the demographical variables, and the conceptual model will be given.

2.1 Defining mobile payments

Dahlberg et al (2008), refer to mobile payment as: “a payment for goods, services, and bills with a mobile device by taking advantage of wireless and other communication technologies”. From this definition it can be concluded that mobile devices can be used in a broad range of payment scenarios (Dahlberg et al., 2000). Hayashi (2012) distinguish three main types of payments scenarios with a mobile device: First, person-to-person transfers initiated from a mobile device. Secondly, payments for services and goods (e.g. ringtones and tickets) purchased over the Internet with a mobile device. And lastly, mobile payments at a POS, which are payments at physical locations (e.g. grocery store) initiated from a mobile device (Hayashi, 2012). As POS purchases account for the vast majority of consumer payments (Hayashi, 2012), the main focus of this research is on this type of mobile payments.

2.2 Dependent variables

(17)

17 will be derived from a mobile payment application, and 2) the intention to use a mobile payment application (in the near future). The utility dimension measures the extent of relative benefit that users may derive when using mobile payment services over using other traditional payment services (Wallenius et al., 2008; Yang et al., 2012). The usage intention dimension, on the other hand, measures factors that are likely to encourage or discourage mobile payment usage intention (Yang et al., 2012).

2.3 Independent variables

(18)

18

2.3.1 Payment system

Research showed that system characteristics have the potential to directly affect the intention to use a system (Davis et al., 1989). Principally, creating a mobile payment application is an activity that occurs between several parties. Banks for example can develop their own mobile payment devices, but this would be very costly due to investment costs in R&D and production (Lim 2008). On the contrary, mobile payment device manufacturers cannot act as banks, as they are not recognized financial institutions (Lim, 2008). Therefore collaboration is needed, and together these parties can make a variety of applications that can make mobile payment possible at the point of sale (Lim, 2008). In the introduction several mobile payment initiatives were already mentioned like payment applications in which payment amounts can be emailed or applications in which payments can be made with the help of Bluetooth (Zonneveld, 2013). However, this research focuses on the two major systems for conducting mobile payments that the three biggest Dutch banks (ING, Rabobank and ABN Amro) are planning to use in their mobile payment applications.

(19)

19 At this moment, the biggest challenge in the Dutch mobile payment market is getting all the involved participants agree on which technology standard to adopt, a matter that has not yet been resolved (Hayashi, 2012). Without this agreement, incompatible technologies can be developed which leads to delayed consumer adoption. The main barrier that keeps investors from introducing mobile payment is uncertainty about the net benefit to themselves, and to consumers relative to traditional payment methods (Hayashi, 2012). From a merchant point of view, merchants need to invest in installing a new compatible payment terminal (Lin, 2007). Besides the costs of this payment terminal, uncertainty about which standard payment option will emerge as dominant appears to be keeping merchants from making this investment (Crowe et al., 2010). From a consumer point of view, the decision mainly depends on the payment convenience (Hayashi, 2012). Convenience can consist of different aspects, including flexibility, speed, and setting up and learning to use each payment method (Hayashi, 2012). Research of Weber and Darbellay (2010) showed that consumers adopt mobile services more easily when they are simple and rapid. In comparison to traditional payment methods, WAP applications have the advantage that several bank accounts can be coupled to the application, and consumers can choose a payment method that best suits the type of payment that needs to be made (Hayashi, 2012), which higher flexibility. For example every-day small payments can be paid directly from a bank account, whereas occasional larger payments can be debited from a credit card account (Hayashi, 2012). On the other hand, estimates show that NFC payment will be up to 15 to 30 seconds faster than traditional forms of payment (Hayashi, 2012). Research showed that perceived speed has a significant influence on consumers’ preferred payment method, especially in cases where consumers need to pay quickly (e.g. at highway toll gates) (Ching and Hayashi, 2010). Depending on system requirements of consumers, they can have a preference for one system over the other.

(20)

20 1997). Nevertheless, more novel products are also likely to be more complex, which can reduce consumer acceptance (Gielens and Steenkamp, 2007). When looking at mobile payment technologies, mobile payment systems with NFC can be seen as novel. NFC is a technology that does not exist in current payment products, which makes it a completely new payment method (Ruben, 2013). Consumers are initially unfamiliar with this technology, and therefore they cannot rely on existing knowledge to estimate the non-obvious features of the product (e.g. durability, convenience, risks, etc.) (Veryzer, 1998; Gielens and Steenkamp, 2007). This might negatively influence consumer acceptance. On the other hand, mobile payment applications with a WAP approach can be classified as new. Consumers are not previously encountered with WAP applications that have the specific characteristic to make payment at a POS possible to purchase offline goods, but WAP approaches are already used in many other mobile applications like mobile banking applications, ticket ordering applications, etc. Mobile payment with the help of a WAP application can therefore be seen as an extension of an existing technique (Veryzer, 1998). Given the fact that consumers are already familiar with WAP applications and feel comfortable when using them, it is predicted that a high novelty application - like a mobile payment application with NFC - is less acceptable to consumers than those with lower novelty - like the mobile payment applications with a WAP approach. From this, the following hypothesis is formed:

H1: Offering a mobile payment application with a WAP approach (opposed to a NFC approach) has a

positive effect on the consumers’ A) utility derived from a mobile payment application, and B) intention to use a mobile payment application.

2.3.1.1 Technology readiness

(21)

21 Technology readiness refers to: “peoples propensity to embrace and use new technologies for accomplishing goals in home life and at work” (Parasuraman, 2000). This construct can be viewed as an overall state of mind that determines consumers’ disposition to use new technologies (Ranaweera et al., 2008), and consists of four influential factors: optimism, innovativeness, discomfort, and insecurity (Parasuraman, 2000). The underpinning of the technology readiness construct is that the more technology ready consumers are, the more likely that they are the first adopters of the new innovation (Parasuraman, 2000). These consumers can be described as “explorers”, they are a relatively easy group to attract when a new technological system is introduced because they have no doubts about it (i.e. high on confidence and innovativeness) (Yousafzai and Yani-de Soriano, 2011). Yet, there has been little academic research on the impact of technology readiness on purchase intentions in an online buying context (Parasuraman, 2000; Jafaar et al., 2007; Ranaweera et al., 2008), and research on technology readiness in a mobile payment context is even more limited. Therefore, it is interesting to investigate consumers’ readiness to use this new innovation.

In addition, there is extensive evidence that shows a negative relationship between age and acceptance of new technological systems (Lee et al., 2002; Zhang, 2005; Yousafzai and Yani-de Soriano, 2011). The outcomes show that older consumers have are more anxiety in using new technological systems, often associated with negative attitudes and higher perceived usage difficulties. Younger consumers on the other hand are more technology savvy and are more attracted to new technological systems. They feel comfortable to accept new technological systems, and switch easily to new product offerings (Gielens and Steenkamp, 2007; Yousafzai and Yani-de Soriano, 2011). Besides, these younger consumers find it easier to learn and use new technological systems. Therefore, it is supposed that technology ready consumers will mainly be younger technology savvy consumers who like to try and use the latest technological systems available. The utility derived from mobile payment applications that consist of the latest NFC “swiping” technology system is therefore expected to be higher for technology ready consumers than for less technology ready consumers. In addition, the intention to use a mobile payment application which has this NFC technology system, is also expected to be higher for technology ready consumers. Overall, it indicates that there is possibly a moderating effect of technology readiness on the utility and the intention to use a mobile payment application when a company offers a mobile payment application with a swiping (NFC) system. Therefore, it is hypothesized that:

H2: The “more technology ready” consumers are, the higher A) the utility derived from a mobile

(22)

22

(NFC), and B) the higher the intention to use this mobile payment application, when a company offers a mobile payment application with swiping function (NFC).

2.3.2 Payment options

Before making a purchase, consumers first need to consider whether they can afford the purchase and if there are monetary resources to finance the purchase. If consumers do not have the monetary resources they can choose to start saving or take out a loan (Hahn et al., 2013). Saving means that the consumer waits for the desired product until the necessary funds are saved (Hahn et al., 2013). By taking out a loan, the consumer gets instant access to the desired product, but it also causes debt, sometimes even after the product is no longer functional (Hahn et al., 2013). Especially in times of financial stress – when customers have negative expectations about their financial situation - customers appear to have an underlying preference for spending from loan (Borzekowski et al., 2008). Another reason for customers to take out a loan is due to impatience (Hahn et al., 2013). A study of Berndsen and Van der Pligt (2001) showed that people generally prefer a positive outcome sooner rather than later.

(23)

23 than those who use direct payment (Soman, 2003). Hahn et al. (2013) explain this phenomenon a bit easier by stating that purchases involve costs and benefits - which are associated in the consumers’ mind. They explain that payment can be seen as negative (providing disutility) and the desired product or service as positive (providing utility). Assuming that consumers prefer a ‘happy ending’, payment by taking out a loan will constitute the opposite because it has a negative ending (Hahn et al., 2013).

Overall it can be concluded, that consumers might need to choose their payment type more consciously with regard to spending pattern and the future happiness with the product (Soman, 2003; Hahn et al., 2013). When thinking rational, offering direct payment (instead of monthly payment for example), will lower the pain associated with the payment, and the purchased product can be enjoyed as if they were free. When thinking emotional, monthly payment offers the opportunity to get instant access to a desired product that the consumer cannot afford at this moment, but it also causes debt, sometimes even after the product is no longer functional. Besides, monthly payment may cause extra costs for the consumer due to owed interest. It is expected that Dutch consumers – who are constantly warned by the government for entering into loans – are rational spenders. Therefore, from a rational point of view, direct payment is supposed to be the preferred payment option for mobile payments. From this, the following hypothesis is formulated:

H3: Offering direct payment (direct funding from a bank account) instead of monthly payment to

customers will positively affect the consumers’ A) utility derived from a mobile payment application, and B) intention to use a mobile payment application.

2.3.3 Payment fees

(24)

24 fixed per transaction (Wang, 2010). When comparing mobile payment to traditional payment with plastic cards (i.e. cash or credit card), the only difference is that the corresponding account number is now linked to a mobile payment application instead of a card (Hayashi, 2012). In this sense, it seems logical that the interchange fee will be the same for the merchant no matter what type of payment is used. Therefore, for both the merchant and the consumer, conducting mobile payments have neither an advantage nor a disadvantage over the traditional cash card or credit card payment methods (Hayashi, 2012).

Nevertheless, mobile payment deploys the possibility for reducing costs for both the merchant and the consumer (Weber and Darbellay, 2010; Thakur, 2013). Innovations like mobile payment systems are developed and supported by both banks and telecom service providers to increase the reach and capability to serve large masses of customers (Thakur, 2013). Therefore, the costs for reclaiming a payment amount can be divided by both the bank and the telecom service providers, instead of only the bank, causing lower interchange fees to the merchants (Thakur, 2013). Especially, from a micro point of view this offers opportunities for mobile payment. As mentioned previously, currently credit card transactions are not a viable payment option for - especially small - payments in many stores due to the relative high transaction fees being charged to customers. If mobile payment investors can indeed achieve to lower the transaction costs for cash card and credit card payments being made with a mobile device (i.e. lower interchange fees to the merchant), consumers can profit from making a payment with a mobile device instead of using traditional payment methods. In this case, merchants can charge no – or lower – fees to consumer no matter what amount needs to be paid and no matter what type of payment card - that is coupled to the mobile payment application - is used. Results from research by Borzekowski et al (2008), show that per-transaction fees for certain payment methods significantly lower the probability of using that method. Where 90 percent of the users use their cash card when no fees are being charged, this probability drops to 79 percent when fees are being charged (Borzekowski et al., 2008). Therefore, lower – and preferably no fees – being charged for mobile payments probably result in a higher willingness to use a mobile payment application as a future payment method (Weber and Darbellay, 2010).

(25)

25 extra service levels falls out of the range of this research. It should also be noted that the impact of payment fees on the intention to use a mobile payment application falls out of the range of this research as it makes the questionnaire too difficult to complete for respondents (more explanation can be found in chapter 3).

Research of Weber and Darbellay (2010) showed that consumers use mobile services more easily when they are cheap, so charging fees may have a negative effect on consumers’ utility that will be derived from a mobile payment application (Zhou, 2011; Yang et al., 2012; Chong et al., 2012), unless transaction fees are lower than the charged fees for traditional payment methods. It is interesting to investigate when interchange compensating fees are being charged, what type of fee (ad valorem or fixed) most negatively impacts the derived utility. From this, the following hypothesis is formulated:

H4: Charging transaction fees to consumers will negatively affect the consumers’ utility derived from a

mobile payment application. 2.3.4 Payment risks

(26)

26

2.3.4.1 Security of financial information

Research showed that security is the most important attribute for consumers when choosing a payment method, and a lack of security is the most frequent reason for refusal to use (Shin, 2009; Hayashi, 2012). Assurance about security relates to the extent to which mobile payment operators guarantee the safety of customers’ financial information (Zhou, 2011; Yang et al., 2012; Hayashi, 2012; Zhou, 2013). As many financial interactions involve some form of risk, it is not surprising that a substantial body of research has tried to understand how companies incorporate this risk in their security assurance choices (Charness and Gneezy, 2011). The security assurance of operators includes two duties: I) firstly, operators are not allowed to disclose personal data to third parties (Weber and Darbellay, 2010). Consumers need to transfer their personal (home address, phone number, etc.) and confidential information (e.g. payment amount, bank account information, identity of the payer/payee, etc.) through their mobile devices (Kim et al., 2008; Meharia, 2012). Therefore, companies need to protect this confidential information by insuring that it is restricted only to the parties involved in the transaction (bank, merchant, Credit Company, etc.) (Kim et al., 2008; Toma, 2012). II) Secondly, operators must prevent unauthorized people from fraudulent use of the system and customers’ personal financial information (Weber and Darbellay, 2010). Consumers may store large amounts of sensitive (payment) information on their mobile devices, and with such information stored in one single place; mobile devices become a bigger criminal target than cash and plastic cards (Hayashi, 2012). Therefore, the risk of becoming victim of fraud or identity theft is seen as a greater barrier for consumers’ to not use a mobile payment application. By know it becomes clear that companies need to invest in security mechanisms for mobile applications, to ensure payment security and information confidentiality.

Although the two duties do not vary from traditional online banking, consumers still perceive mobile payment as more risky (Dai et al., 2012). Some experts even said that it is not the immediate treat itself which holds consumers from using mobile payment, but the consumers’ perception about these treats (Schlosser et al., 2006). For example, for consumers it is quite unobservable when they are at risk when conducting mobile payments, as is the case for example with skimming of your cash card at the cash machine, and what companies do to shield consumers from these risks. How can payment operators then assure and communicate that they have the ability to protect their customers?

(27)

27 used for making fraudulent and unauthorized payments. Nevertheless, mobile payment security can be upgraded by companies when a chip is embedded in the mobile device. This chip makes the data unique to each transaction (i.e. dynamic authentication), and when intercepted by criminals impossible to make fraudulent mobile payment transactions (Hayashi, 2012). Another security mechanism that can be implemented more easily by companies is external confirmation (Weber and Darbellay, 2010). Thereby, transaction related data can be sent to the customer for example by email, so that the customer can confirm or decline the transaction (i.e. digital signature). Research into credit card payments found that in many cases, consumers identify irregularities too long after the fraud has been committed (Konidala et al., 2012). In the Netherlands, banks have contained rules that consumers must check their account (online) regularly (i.e. required login frequency). If consumers have not logged in often enough to check for irregularities, the consumer can be held responsible for the damage itself (Nijboer, 2013). Exposing payment details directly to customers has an advantage for customers because they can identify irregularities immediately. Besides, this extra confirmation can positively influence customer confidence in the offered service (Weber and Darbellay, 2010). However, these extra levels of authorization might cause the consumers some more verification handlings before/after usage. As mentioned before, consumers adopt mobile services more easily when they are simple and rapid (Weber and Darbellay, 2010). To some extent, extra protection mechanisms might conflict with these requirements and might therefore negatively influence mobile payment intention. Also, Hayashi (2012) found that though consumers’ security perception affects their decision to make online payments, once that decision has been made, safety does not affect consumers’ decisions about which payment method to use anymore, nor the number of online payments to be made.

Overall, security is mostly seen as the cornerstone of mobile payment, and it probably has the biggest influence on initial usage intention (Weber and Darbellay, 2010). Although consumers must sacrifice some payment speed to do some extra verification (i.e. protection) procedures, it is expected that these extra protection mechanisms will have a positive effect on the intention to use a mobile payment application, and will also positively influence the utility that will be derived from the mobile payment application. From this, the following hypothesis is formulated:

H5: Offering an extra level of authorization to customers will positively affect the consumers’ A) utility

(28)

28

2.3.4.2 Protection against loss and fraud

Under current laws and regulations, protection against loss and fraud depends on the payment instrument and not on the payment device being used (Hayashi, 2012). Meaning that when the mobile payment application of a bank for example is linked to a credit or cash card, the federal law for these payment methods (credit card or cash card) is valid, regardless the fact that the fraudulent payment is made with a mobile device. Nevertheless, when a consumer uses mobile payment applications of payment intermediaries such as PayPal, federal laws and regulations provide the consumer little or no protection against loss from fraud (Hayashi, 2012). So, depending on the vendor of the mobile payment application (i.e. an intermediary or a bank) consumers experience different levels of protection. In the offline buying context, the general rule in the Netherlands is that consumers fall under current laws and regulations unless there is gross negligence or intent of guilt (e.g. giving your PIN code to someone on purpose) (Nijboer, 2013). Nevertheless, the regulations regarding fraudulent payments on the Internet are a much grayer area, and will be even vaguer in the mobile Internet area. For example in the case of phishing mails and computer viruses, the allocation of responsibility for failure or loss may not always be clear (i.e. who to blame) (Bahli and Benslimane, 2004). It can be said that the consumer has responded to the phishing mail and gave personal information voluntarily, so is it fair to shift the responsibility for the loss to the bank? And in the case of computer viruses which encloses personal payment information: did the consumer took enough precautions? (i.e. antivirus software / regularly checking for viruses) (Nijboer, 2013). Of course, banks and other financial institutions have compensation policies; however, these may differ from company to company.

However, this possible lacking protection is not the only threat to consumers. There is also the chance that banks become victim of fraud. Especially, now Dutch banks have recently been under great attack - the so-called DDOS attacks (i.e. overloading online banking websites with visitors such that the website crashes (form of hacking)), consumers wonder to which extent banks can ensure the safety of their money themselves (Nijboer, 2013). Therefore, companies must demonstrate how their experience with fraud puts them in the best position to reduce this risk for customers (Uzureau, 2012).

(29)

29 insurance against loss and fraud will also positively affect the intention to use a mobile payment application, and also positively influence the utility that will be derived. From this, the following hypothesis is formulated:

H6: Offering an insurance against loss and fraud will positively affect the consumers’ A) utility derived

from a mobile payment application, and B) intention to use a mobile payment application. 2.3.4.3 Risk-aversion

Risk aversion is the term that is used to express “the consumers’ willingness to take on risky activities in pursuit of its goals and values” (Aven, 2013). This definition can be viewed as context specific, meaning that risk perception and the level of risk tolerance can differ among individuals (Davies, 2009). Nevertheless, still little is known about the role of risk aversion on the intention to use a mobile payment application. It seems logical that only those consumers with a higher degree of risk tolerance will be motivated to use a mobile payment application (Tan, 1999; Aven, 2013). However, this study is interested in examining the effectiveness of the risk reduction methods mentioned previously - offering an extra level of authorization (i.e. PIN code and (digital) signature), and an insurance against loss and fraud. Previous research already indicated that risk relievers in the Internet shopping context like money-back guarantees and warranties lower the consumers’ risk perception (Tan, 1999). Therefore, it is also expected that both mobile payment risk relievers positively affect the utility and usage intention for high risk-averse consumers (i.e. reduce their perceived risk). In other words, it is expected that there is possibly a moderating effect of risk-aversion on the utility and intention to use a mobile payment application when a company offers this extra level of authorization and an insurance against loss and fraud. From this, the following is hypothesized:

H7: The more risk-averse consumers are, the higher A) the utility derived from a mobile payment

application which offers an extra level of authorization, and B) intention to use a mobile payment application when a company offers this extra level of authorization

H8: The more risk-averse consumers are, the higher A) the utility derived from a mobile payment

(30)

30

2.4 Control variables

(31)

31

2.5 Conceptual model

From the literature review the conceptual model can be constructed. The constructs protection against loss and fraud, and the level of authorization relate positively to the dependent variables 1) utility derived from a mobile payment application, and 2) intention to use a mobile payment application. On the other hand, payment fees are expected to be a barrier to use a mobile payment application and will therefore be negatively related to the derived utility. For payment system and payment option, the influence is not specified since the different levels are not ordinal. Lastly, technology readiness and risk aversion are included as moderating effects. Figure 2.5.1 below gives a graphical representation of the expected relationships between the constructs.

Figure 2.5.1 Conceptual model

A. Utility derived from a mobile

payment application Level of authorization H1 H3 H4 (-) H5 (+) H6 (+) H7 +H8 : Risk aversion B. Intention to use a mobile payment application H2: Technology readiness Protection against loss and fraud

(32)

(33)

33

Chapter 3 Research design

This research design will be the blueprint for the rest of this research (Malhotra, 2007). First, a description of the needed information for this research will be given. Furthermore, the design of the questionnaire will be specified. This part includes the measurement and scaling procedures and the sampling technique that will be used. Finally, the plan of data analysis will be described, which contains the indicated measurement methods for this research.

3.1 Research goal

To give a valid answer to the research questions and to test the hypotheses that are formed in the previous chapter, an online choice questionnaire is made with the help of the program Sawtooth SSI Web and ThesisTools. An online questionnaire facilitates the gathering of a large amount of data in a relatively short time span, and because the responses are limited to the alternatives stated, a reliable dataset can be obtained which can be analyzed in a uniform and coherent manner (Malhotra, 2007). The questionnaire aims to measure the following issues:

- The utility that will be derived from the attributes of a payment application: payment system, payment option, payment fees, payment protection and payment insurance.

- The influence of these payment applications’ attributes on the intention to use a mobile payment application (in the near future).

- The extent to which technology readiness and risk aversion moderate the utility derived from a mobile payment application and the intention to use a mobile payment application (in the near future).

- A segmentation of consumers with regard to their responsiveness to the mobile payment attributes, described by their demographics and attitude towards mobile payment.

3.2 Design of the questionnaire

(34)

34

3.2.1 Part 1 - Choice based conjoint

In traditional conjoint analysis, respondents evaluate products (i.e. profiles), constructed with selected levels from each attribute, one at a time (Hair et al., 2010). In real market situations, however, consumers choose a product out of different alternatives. Therefore, it is chosen to use a choice based conjoint approach, which employs a unique form of presenting products (i.e. profiles) in sets, rather than one by one (Hair et al., 2010). This approach gives a better, more realistic representation of the behavior of consumers. Probably, when a consumer decides to use mobile payment in the near future, the consumer can choose from a range of available mobile payment applications (for example from different vendors).

In addition, in real market situations, consumers have the option to use other payment types instead of mobile payment (e.g. traditional forms of payment like cash and cards or with their mobile device in the future). Therefore, it seems realistic to include a none-option in the set of profiles, indicating that the consumer prefers traditional forms of payment over payment with a mobile payment application. Still, it is chosen to not include a none-option, as this research investigates a hypothetical situation in which the respondent has already decided to install a mobile payment application on their mobile device, but now needs to choose which type of mobile payment application is most suited to conduct future mobile payments. In addition, the drawback of including a none-option might be that respondents are more inclined to choose the familiar, traditional ways of payment and not take mobile payment applications into consideration as they are not familiar with this type of payment yet, and also have no prior experience with it. Therefore, it is expected that respondents will choose the none-option too often.

(35)

35 short description (one or a few words). The full description can be found in the questionnaire in appendix 1. More explanation about the attributes and the attribute levels will be given later on in this chapter.

3.2.2 Part 2 - Usage intention

The usage intentions of the respondents are measured in the second part of the questionnaire. In this part, the respondent will see several promotional offers, which include all the different product attributes again. The product attribute payment fees is left out in these promotional offers, as it otherwise resulted in a extreme large number of questions that needed to be evaluated (48 questions = (2 attribute levels ^ 4 attributes) * 3 attribute levels). By leaving payment fees out, the respondent only has to evaluate 16 (2 attribute levels ^ 4 attributes) promotional offerings related to their usage intentions. Still, the respondent has to fill in several other questions in the last part, therefore rating another 16 product profiles is probably too difficult. In order to make sure that the respondents will not be confused and overwhelmed from all the survey questions, it is chosen to reduce the number of promotional offers to 5 product profiles that need to be evaluated (4 attributes + 1). The usage intention is measured on a 3-point Likert scale, by asking the respondent how likely it is they will use the mobile application in the promotional offering in the near future: 1) Not likely at all, 2) Maybe, or 3) Very likely.

3.2.3 Part 3 - Moderators Technology readiness and Risk aversion

(36)

36

Variable Questions Scale Reference for question/scale

Technological skills/ capabilities (TS) Part of technological readiness

TS1: The use of a mobile payment application would be easy to learn for me

5-p Likert scale Crespo et al. (2009) and Shih and Fang (2006), and adapted for this research

TS2: A mobile payment application would be easy to use for me

5-p Likert scale Crespo et al. (2009) and Shih and Fang (2006), and adapted for this research

Behavior towards new innovations (AI) Part of technological readiness

AI1: I always prefer to use the most advanced technology available

5-p Likert scale Ranaweera et al. (2008), and adapted for this research

AI2: Most of the time, I am among the first in my circle of friends to start using new technology

AI3: I feel comfortable with the use of new technologies 5-p Likert scale Made myself

Risk Aversion (RA)

RA1: I am careful/cautious in trying new/different products 5-p Likert scale Ranaweera et al. (2008), and adapted for this research

RA2: I never buy something I don’t know about at the risk of making a mistake

RA3: I usually try than buy something I am not very sure of 5-p Likert scale Ranaweera et al. (2008), and adapted for this research

RA4: The idea of buying a product through a mobile device gives me worry / gives me a feeling of anxiety.

5-p Likert scale Gutiérrez et al. (2010), and adapted for this research

RA5: I don’t think it is safe to make a payment though a mobile device

Table 3.2.3.1 Research questions and academic foundation

3.2.4 Part 4 - Demographics

At the end of the questionnaire several demographical background questions will be asked, which afterwards can be used for describing the different segments. These are all multiple choice questions (gender, education, status, income, household size, geography, etc.), except for attitude and age. Age will be measured by means of an open question, and attitude will be measured by asking 4 questions which are measured on a 5-point Likert scale. Table 3.2.4.1 shows an overview of the research questions and the academic foundations. Again it will be checked whether the attitude scales show sufficient reliability and consistency to take the mean from the attitude questions as a score to measure attitude (see paragraph 3.5).

Variable Questions Scale Reference for question/scale

Demographics Age, gender, income, education, status, household size, geography Multiple choice/ open question

Plumer (1974), and adapted for this research

Attitude (AT)

AT1: The use of a mobile payment application is: unfavorable – favorable

5-p Likert scale Peslak et al. (2010) and adapted for this research

AT2: The use of a mobile payment application is: useless – useful 5-p Likert scale Peslak et al. (2010) and adapted for this research AT3: The use of a mobile payment application is: worthless –

worthwhile

AT4: The use of a mobile payment application is: unhelpful - helpful

(37)

37

3.3 Data collection and sampling

A minimal sample size of 200 respondents is needed to conduct the choice based conjoint analysis and the latent class analysis. Besides, this number of respondents has been found to provide an acceptable margin of error (Hair et al., 2010). This research concentrates on Dutch mobile phone users, with different demographics, such that a representative sample is obtained. This means that everyone is allowed to participate, but the respondent must own a mobile phone of must be planning to buy one in the near future. Therefore, the questionnaire will start with a question to explore whether the respondent indeed meets this precondition. If it turns out that the respondent does not own a mobile phone and is also not planning to buy one in the near future, the respondent will be excluded from further participation. In order to get 200 respondents, the sampling technique of snowball sampling is used. This is a non-probability sampling technique in which an initial group of respondents is selected, and afterwards is asked to find others who belong to the target population of interest (Malhotra, 2007). For this research, the initial group consists of family, friends and colleagues of the researcher. This group is reached by email and social media like Facebook and LinkedIn, and is asked to forward / share the questionnaire to their contacts.

3.4 Plan of analysis

This paragraph clarifies the research methods that will be used to test the hypotheses of the previous chapter. First, the choice based conjoint analysis by Latent Gold will be described. This method will be used to measure the utility. Besides, Latent Gold will be used to create and describe segments. Secondly, the Ordered Multinomial Logit approach will be explained. This method will be used to measure the usage intentions

3.4.1 Choice based conjoint analysis

(38)

38 attributes, except payment fees, are measured at two levels. Therefore, it does not matter whether a part-worth function of linear vector model will be used. Still, all five attributes are non-metric (i.e. nominal), therefore a part-worth model is chosen.

Attributes Attribute levels Pref. function Academic source

Payment System 1) Swiping Part-worth Lin, (2007) Borzekowski, (2008); Wang, (2010); Hayashi, (2012)

2) WAP application

Payment Option 1) Direct payment Part-worth Soman, (2003); Song and Sahedi (2005); Hayashi, (2012)

2) Monthly payment

Payment Fees

1) Free

Part-worth

Verdier (2009); Yang et al, (2012), Hayashi, (2012), Zhou, (2013); (amounts adopted from: Waris et al., (2006); Wang, (2010)

2) Fixed costs (€0.10)

3) Flexible costs (2% payment amount)

Payment Protection 1) PIN code Part-worth Hayashi, (2012); Weber and Darbellay, (2012)

2) PIN code + (digital) signature

Payment Insurance 1) NO Part-worth Shin, (2010); Hayashi, (2012)

2) YES

Table 3.4.1.1 conjoint task and academic sources

The number of attributes included in the analysis directly affects the statistical efficiency and reliability of the results (Hair et al., 2010). Two limits come into play when considering the number of attributes to be included. First, Green and Srinivasan (1978) recommend including a maximum of six attributes to create enough difference between the profiles, and to make them clear enough for the respondent. This research includes five attributes, which falls within the recommended ranges. Secondly, adding more attributes to a conjoint study increases the minimum number of profiles in the study (Hair et al., 2010). Therefore, it is evaluated how many profiles are sufficient for this research.

(39)

39 The program Sawtooth Software SSI Web makes it possible to test the whole design for efficiency. The rule of thumb is that an efficiency of zero indicates bad efficiency, where an efficiency of one indicates the best possible efficiency. In appendix 2 the results of this test can be found. For the conjoint design of this research the efficiencies are one - or very close to one - indicating that the choice questionnaire has an efficient design.

In order to determine the relative importance consumers attach to salient attributes and the utilities they attach to the levels of these attributes first a conjoint analysis on aggregate level will be performed, to test for general results. To eventually find the most preferred mobile payment application, the utility values associated with each of the product attributes can be summed. The mobile payment applications with higher utilities are assumed to be more preferred and have a better chance of being chosen by the respondents. This can be depicted in the following formula for determining utility (see figure 3.4.1.1):

Figure 3.4.1.1 Utility formula

Afterwards, a conjoint analysis on segment level will be performed. The estimated conjoint part-worth utilities, in combination with other variables (demographics and the moderators), will be used to group respondents with similar importance/preference values and part-worths to identify segments (Hair et al., 2010).

3.4.2 Ordered multinomial logit

To measure the mobile payment usage intentions of consumers, an ordered multinomial logit approach will be used. The assumption of this approach is that consumers will choose the alternative that gives the maximal utility (Leeflang et al., 2000). As mentioned previously, the usage intentions of the consumers will be measured on a 3-point Likert scale, ranging from “not likely to use” till “likely to use”. This dependent variable is characterized as discrete, instead of continuous; therefore a logit model should be used. In addition, it follows an ordered pattern (i.e. ordinal), meaning that the

U

j

= ß

1, j

X

1

+ ß

2, j

X

2 +

ß

3, j

X

3

+ ß

4, j

X

4

+ ß

5, j

X

5

Where,

U = Utility X3 = Payment Fees

j = Segment 1,…..,n X4 = Payment Protection

X1 = Payment System X5 = Payment insurance

(40)

40 respondent faces a ranked variable. Given the fact that the dependent variable is ordered in a meaningful way, and the respondents face more than two options to choose from in the questionnaire, an ordered multinomial logit model is the most appropriate model to use for this research (Franses and Paap, 2001). In the ordered multinomial logit model, all attributes except payment fees will be included. In addition, the moderators – technology readiness and risk aversion – will be included as interaction effects. The promotional offerings are designed manually, and tested for multicollineairity (i.e. VIF > 10) to make sure that the independent variables are not correlating. The results of this test shows VIF values of respectively: Payment system (1.2), Payment option (2.4), Payment protection (2.4), and Payment insurance (1.2), indicating that there is no multicollineairity and the design is sufficient to use.

3.5 Data preparation

The moderating questions and the questions related to attitude are checked to see if the scales show sufficient reliability and consistency to take the mean of these questions outcomes. Given the fact that multiple questions are asked, most of which are correlated, they must be reduced to a manageable level (Malhotra, 2007). As a starting point, it is preferred to do a factor analysis to check if the questions really relate to the a priori made dimensions 1) technological skills/capabilities, 2) behavior towards new innovations, 3) risk aversion, and 4) attitude towards mobile payment. Before the analysis can be executed, it is necessary to examine the appropriateness of factor analysis. The rule of thumb is that the Kaiser-Meyer-Oklin (KMO) measure of sampling adequacy should be between 0.5 and 1.0 for factor analysis to be appropriate (Malhotra, 2007). For this research the KMO measure was 0.835, which perfectly falls within this range. In addition, the Bartlett’s test of sphericity tells us with a significance level of 0.000 that the variables correlate perfectly with itself, but has no correlation with the other variables (Malhotra, 2007). The outcomes of the factor analysis can be found in appendix 3. The results show a factor solution of 3 factors to be best for this research. To interpret the factors, a Varimax procedure is used, which minimizes the number of variables with high loadings on the same factor (Malhotra, 2007). A large value indicates that the variable (in this case the question) is closely related to the factor (Malhotra, 2007). The results showed that the two parts explaining technology readiness were taking together in one factor, and both risk aversion and attitude are in a separate factor.

(41)

41 0 to 1. A value of .60 to .70 is deemed to be the lower limit of acceptability (Hair et al., 2010). Below in table 3.5.1 the Cronbach’s alphas are shown. As can be seen, all Cronbach’s alphas are far above .70, indicating that the mean of the items can be used for measuring moderating effects later on in this research.

Reliability statistics

Dimension Number of items Cronbach's Alpha

Attitude 4 0.934

Technology readiness 5 0.779

Risk aversion 5 0.863

(42)

(43)

43

Chapter 4 Analysis and results

Within this chapter, the outcomes of the performed analyses will be discussed. First, the descriptive statistics of the research sample will be discussed. Afterwards, the results of the aggregate conjoint analysis will be described, followed by the results of the latent class analysis. With the help of these results, the segments of consumers can be explained and evaluated. The chapter concludes with the outcomes of the ordered multinomial logit, which tests the role of the mobile payment attributes and the moderators on the usage intentions of consumers.

4.1 Sample descriptive statistics

The online questionnaire was open to respondents for about 2.5 weeks, and in total 314 respondents participated. However, after the first questions about the mobile phone usage, about 60 of these respondents quit the questionnaire. Furthermore, 4 respondents did not own a mobile phone and were not planning to buy one in the near future. Therefore, these respondents were excluded from further participation. After the conjoint part, only 214 respondents were left in the questionnaire. Fortunately, all these 214 respondents completely filled in the questionnaire and therefore the minimal number of 200 respondents is achieved. This results in a total sample of 214 respondents that is used in the analyses of this research.

(44)

44 respondents have an income less than the average of €33.000 annually (59.3%), which is probably due to the fact that about half of the respondents have a status as being student. Many of the respondents have a household consisting of 1 or 2 persons (no children) (82.5%) and 137 persons live in the city (64%), whereas 77 persons live in a rural area (36%). A complete overview of the descriptive statistics can be found in appendix 4.

Within the sample, every person owns a smartphone, and most of these smartphones have Internet options (86.5%). If we take a look at the main purposes for which the sample uses their mobile device (see figure 4.1.2), it can be seen that making a phone call is still the main purpose for using their phone (17%). Tough, it is surprising that this is closely followed by using the device to surf the web (14%) and for making pictures and texting (both 12%). Also, it is interesting to see that mobile devices are quite often used for mobile banking (9%). As explained in the literature part, some research showed that experience with other similar mobile services may influence the consumers’ intention to use new services. Thus, a mobile payment application might enjoy positive externalities from the success of these mobile banking services.

4.2 Choice based conjoint analysis

This part of the results gives insight in the results of the choice based conjoint analysis which is generated with the help of the program Latent Gold. First the aggregate choice based model will be discussed to specify a utility function for all respondents in the sample (i.e. based on dummy coding). This means that this model assumes that all respondents within the sample have the same preferences. Afterwards, a choice based model on segment level will be presented. From these results the number of segments can be chosen and described in further detail.

4.2.1 Aggregate solution

“The past, the present, and the new payment future...”