How safe do people feel in safe apps?
The Importance of Security Features Compared to other
Features in Mobile Banking for Product Adoption by
Consumers and the Role of Security Concerns
- Master thesis -
MSc Marketing Intelligence
Author: Lotte Bons
Student Number: S2609231
University of Groningen
Faculty of Economics and Business
First supervisor: R.P. Hars
2
Abstract
3
Table of contents
Introduction 4
Theoretical framework 8
- The technology acceptance model (TAM) (Davis 1989) 8
- Security features: biometric verification methods 10
- Ease of use and usefulness in biometrics 11
- Non-security features 12 - Security concerns 15 - Conceptual model 18 Methodology 19 - Sample 19 - Conjoint set up 20 - Model type 21 - Model validation 23 Results 24 - Data collection 24 - Sample 24 - Model estimation 25 - Moderation effects 30
Discussion and conclusion 34
- Limitations and recommendations for further research 36
4
Introduction
5 Besides that, banks increasingly embrace new features for more precise targeting, customer service, and relationship management (IAB, 2015). In order to provide high quality, on-demand, and context-aware services, apps contain features and functionalities that make use of personal preferences, account users' location, interests, and other personal data (Pentina et al., 2016). But the permissions to access private content required by various apps have exploded and led to greater security and privacy concerns for customers (Olmstead, 2014). Banks and other financial institutions carry much information about the customers and customers are aware of that. Especially given the fact that the customers are increasingly being used by the general public to manage and carry-out sensitive processes, such as mobile banking.
Although implementing new features in bank apps is used to create more value, it can lead to less value at the same time which seems contradictory. Possible consequences of adding innovative features are: breaches of personal identity, financial risks and other possible consequences of personal information becoming public. Mobile phones that are lost or stolen, can create an extra avenue for attackers to perform malicious activities, such as using this phone as a start point for an attack. As it is often the case when such tech innovations like new features come along, it has given rise to new security threats with vulnerabilities that can be exploited by malicious hackers. Even though new mobile banking features like biometric authentication are more reliable as compared to password or token-based systems, the customer's perception can be different which comes along with security concerns. Therefore, security and privacy concerns among consumers arise in mobile banking and there is a growing importance for keeping the financial information secure. Although concepts of security and privacy are tangled, it is possible to have security without privacy, but impossible to have privacy without security. Hence, the moderating role of security concerns will be examined.
6 confidence to accept it as the norm. If customers do not feel confident in mobile banking and have the feeling their transactions could be risky, they will not be loyal of staying with that bank (Gerrard and Cunningham, 2003; Lee, Kwon and Schumann, 2005). Focusing on innovating and adding features might lead to the risk of losing customers to competitors where they feel more secure. Besides, it’s interesting to find out if consumers have a preference in the different types of biometrics that banks are offering. Therefore, there is a need to understand how banks and other financial institutions can create the perfect secure experience for customers. In other words, do security features lead to a more secure feeling among app users. The moderating role of security concerns have to deserve more attention in the existing the literature, since this is an important area of study and only very few studies have done research about the moderating role of these security concerns in mobile banking (San Martin and Camarero, 2009). In other words, do security features become more important in product adoption when people have more security concerns?
At the moment it is not clear to what extent people associate features like biometrics with security, and therefore how important these features are for the product adoption of these mobile banking features. Subsequently it’s unclear whether the implementation is positive or negative from the perspective of creating value to the customer. It’s also interesting to look at which type of biometric is most preferred. Furthermore, it’s questionable to what extent the value for the customer changes when security concerns are greater. Therefore, the aim of this study is to fill the literature gap and contribute to the existing literature by examining the following research question:
“How important are security features compared to other features in mobile banking for the product adoption and to what extent do security concerns have a moderating role? “
Answering this research question is not only relevant for the banking sector, but probably for many different businesses, becoming more and more dependent on technology. Technology is also a matter of the people using it and investigating what these users want. Besides, this research could help banks develop a security strategy in online banking and create more security in general for users and themselves.
8
Theoretical Framework
This chapter provides an overview of relevant literature and theories for the research. The main concepts are defined and the relationships between the concepts will be explained using existing literature. Subsequently, the hypotheses are built on this.
The Technology Acceptance Model (TAM) (Davis 1989)
9 to continue using a bank app. For this reason, TAM constructs would be applicable for measuring continued product adoption of new features in mobile banking.
Figure 1: TAM Davis (1989)
Perceived usefulness and perceived ease of use are constructs that can be applied to many different contexts and the interpretations of these constructs can differ per situation. Perceived ease of use (PEOU) is defined as the individual’s perception that using the new technology and will not takes a lot of effort (Davis, 1989, 1993). Many consumers might find that it is hard to use the new technology system (Davis, 1989). It can be seen as the opinion of an individual’s assessment of the effort utilized on account of using a technology (Davis, 1989). People would not need to allocate much of their time and efforts while using the technology. According to Rauniar et al., 2014) PEOU influences the viewpoint of an individual towards using a technology and besides it predicts the behavioral intention to use the new technology (Nysveen et al., 2005a, 2005b, Davis, 1993). The more clear it is to use a technology, the greater the expected benefits from the technology with regard to its performance. Besides, this relation has been found in online technology context in other studies (Gefen et al., 2003; McCloskey, 2006).
10 using a password. In general, consumers don’t like the idea of using identity traits and think it's a violation of their privacy. According to Fortes and Rita (2016) security concerns about identity fraud are crucial for consumers to download and continue using an app. So, to what extent certain features are evaluated as useful or simpler, and to what extent security features versus other features are important in customer evaluations of usefulness. There seems to be ambiguity among users regarding the benefits of using these security features.
Security Features: Biometric Verification Methods
Biometrics has revolutionized the way identification is performed, it’s becoming more popular in security systems, especially in access control and financial services. Biometrics refers to the use of physiological and/or behavioural characteristics to identify an individual. Because biometric identification is dependent on the person itself, it’s more reliable than traditional systems like passwords. Basically, biometric identification is based on who the user is or what the user is doing. In fact, these characteristics are related to the user himself and are linked to the user. It is impossible to copy or transfer biometric traits of someone else to be used instead. For this reason, biometrics can be seen as a reliable system to identify individuals. In order to identify a person properly, different biometric traits are used: face, iris, voice and fingerprint are the most common ones. In this paper, the focus is on facial recognition and fingerprints recognition since these verification methods are most often used (Bhatia, 2013). However, the benefits of these methods are still not very clear, and some users don’t feel comfortable using this new system.
11 Facial verification is a way of recognizing a human face through technology. Next to face recognition, fingerprint is also used for authentication. Fingerprint is the oldest and also a very popular method in biometric identification thanks to its wide user’s acceptability, accuracy, security as well as to its relatively inexpensive cost. The use of biometrics as a verification method could increase the ease of use and therefore be more preferred than authentication based on memorization or PIN codes. Thus, biometrics could be a proper choice for authentication in universal access systems (Mayron et al. 2013).
Ease of Use and Usefulness in Biometrics
TAM explains users' intentions to use a technology based on the way they behave toward that technology, which are determined by their perceptions of usefulness and ease of use of that technology (Morosan, 2011). Usefulness relates to the practicality of the system and whether it allows the user to achieve their tasks effectively
.
One study about the usefulness in biometrics shows that participants find biometrics faster with better performance and more secure than traditional password techniques: only 31% think that password is faster than biometrics; and only 21% do not believe that biometrics will improve the performance in their lives (Rashed & Henrique, 2013). In order to increase the ease of use in biometric systems, there are multiple factors that are implemented. Sensors are smaller, more reliable, and more ergonomic, biometric algorithms are better, feedback can be provided during use, and they are being integrated to provide a more seamless use and environment (Patrick 2004). The use of biometric data for authentication purposes typically results in a greater comfort and ease of use, with respect to traditional approaches like password or tokens, which can be easily lost, forgotten or stolen.12 authentication. Furthermore, fingerprint as authentication methods also had the highest satisfaction levels by reason of its ease of use. As a result, biometric technologies having an impact the way users look at security. With the changing needs in the financial and banking sector, there is a big chance the use of biometrics will be more used in the future. The perceived usefulness has been argued that influences user intention to adopt the technology. An individual may find it more useful to unlock their phone with fingerprint biometrics or face biometrics than punching a four-digit number. Hence, both fingerprint and face biometrics as a verification method will be perceived as useful and practical for the task, more so than other traditional verification methods. Therefore, it is hypothesized that the use of biometrics in mobile banking will also increase the product adoption.
Hypothesis 1: Face verification in mobile banking as compared to mobile banking without face verification will have a positive effect on the product adoption.
Hypothesis 2: Fingerprint verification in mobile banking as compared to mobile banking without fingerprint verification will have a positive effect on the product adoption.
Non-Security Features
Non-security features are used to embrace banking on mobile devices and be up to date of the newest trends in the banking industry. Keeping the focus on developing mobile banking technologies will help banks to improve their competitive advantages. In this research three different features are considered as non-security features. It’s chosen to use push notifications, qr-payments and geolocations services as non-security features. These features are all innovative and recently added mobile banking features (Streeter, 2019).
Push notifications
13 reminds users to make use of the app instead of relying on the consumer to use the app. On the other hand, receiving push notifications without providing opportunities to customize the app, can be experienced as disturbing and oppressive (Westermann et. al., 2015). The idea of having control over computer-mediated features is an important aspect for users (Sundar, Xu, & Bellur, 2010). Allowing app users to have customization-control over their notifications can increase the actual usage of the app. Based on different research about push notifications it’s important for consumers to: give the user control, offer actionable notifications, don’t message too much, and give the opportunity to customize messages and timing of the messages (Babich, 2016; Kahuna, 2014; Vizard, 2016). Furthermore, linking this feature to the TAM model, Awad et al. (2014) assessed the effect of consumers’ attitudes toward push notifications and shopping intentions. This research argued the positive effect of push notifications to the behaviour characteristics of perceived usefulness and ease of use (Awad et al., 2014).
Besides having control about the push notifications, another study found out that familiarity with the app positively impacted user experience (Bowen & Pistilli, 2012). On the other hand, some researchers note that the ease of use of a website has a significant impact on the user experience, irrespective of how experienced the users are (Flavián et. al., 2006; Davis, 1993). Therefore, having familiarity with an app is important for the user experience of the user, but the ease of use is influential as well. Furthermore, market researchers have found that in the presence of push notifications, users access apps more frequently, either by directly clicking on the push notification banner or by separately opening the app in response to a recent push notification (Clor-Proell et. al., 2019).
Hypothesis 3: Push notifications in mobile banking as compared to mobile banking without push notifications will have a positive effect on the product adoption
QR-payments
14 (Harris et. al., 2019). Another reason why qr is gaining popularity is because merchants only need to print their qr-code and not invest in expensive technology. As qr-payment is an innovative technology in mobile banking there is a big chance that this technology is going to be a worldwide trend in a few years. Adding qr-payments in mobile banking leads to faster payments and thus can affect the perceived ease of use and usefulness from the TAM positively. Based on this assumption, qr- payments increase the product adoption of qr-payments as well. These predictions are based on previous research (Yang et al. 2012; Wang et al. 2014), which can predict the intention to use a new technology and frames positive perceptions about innovation in terms of advantage and ease of use (Lu et.al., 2005).
Hypothesis 4: The possibility to make qr-payments in mobile banking as compared to mobile banking without the possibility to make qr-payments will have a positive effect on the product adoption
Geolocation Services
Geolocation is an advanced technology that helps in identifying the location of any device. The process of geolocation gives the developers of thee apps the opportunity to identify the physical location of the users. The trend of geolocation has densely entrenched in the mobile application market. Nowadays, there are hundreds of location-based apps that detect the location of users to provide the desired information based on their location (Promatics Technologies, 2018). For example, apps like Uber use geolocation to know where the customer is and match it with a nearby driver. Another app using geolocation are weather apps, these apps can give a notification when it’s going to rain, based upon your location. Geolocation can also be used as a feature in mobile banking. With this mobile banking app feature, the customer can see where the nearest ATM or psychical bank is located. Besides there are also banks that gives the actual information on available cash in the nearest ATMs.
15 create successful marketing efforts, location-based services have to be applicable and useful (Zhao, Lu et al. 2012). Offering features like geolocation services in mobile banking is not the first priority in mobile banking, but it can certainly add value to the customer. As a result, the value created by this service has to be much higher to compensate for the negative feelings consumers might experience with their location being tracked. Even though geolocation can increase the ease of use in mobile banking because it enriches our mobility experiences, it creates privacy concerns, when others have the possibility of tracking the location of a user (Chen & Tsai, 2019). For this reason, location-based services can be considered as not useful. This paper hypothesizes that geobased services will not have a positive effect on the product adoption of the app because location-based services will be considered to be too intrusive.
Hypothesis 5: Geolocation services in mobile banking as compared to mobile banking without geolocation services will have a negative effect on the product adoption
Security concerns
16 construct in the behavioral stream of research on security because it affects user intentions and behaviors (Balapour et al., 2020). So, the subjective security users perceive in mobile banking can affect their attitude towards production adoption. Research about security concerns in online environments demonstrated that an individual’s perception of security is the building block of trust towards any form of electronic transaction, which can fuel users’ behavioral intentions and thus the product adoption process (Bansal, 2017). Another study done by Johnston and Warkentin (2010) revealed that security perceptions can indirectly lead to organizational security compliance, which could be an issue in many companies. Thus, when security concerns arise, the features in mobile banking become weaker.
Researchers have found that security plays an important role in studying user attitude towards online banking (Thakur et al., 2014). Security of mobile banking services is a big issue and it will always be a big challenge. Customers want to have the convenience: they want to have the possibility to access their bank account on their mobile phones everywhere. However, the security of the apps still crucial, since many of these transactions are done through the internet.
17 where financial information of customers is used, it’s expected that security concerns decrease the probability of adopting innovative features in mobile banking.
Hypothesis 6a: The extent to which a person holds security concerns moderates the effect of face verification on the product adoption negatively
Hypothesis 6b: The extent to which a person holds security concerns moderates the effect of fingerprint verification on the product adoption negatively
Hypothesis 6c: The extent to which a person holds security concerns moderates the effect of push notifications on the product adoption negatively
Hypothesis 6d: The extent to which a person holds security concerns moderates the effect of qr-payments on the product adoption negatively
Hypothesis 6e: The extent to which a person holds security concerns moderates the effect of geolocation services on the product adoption negatively
18
Conceptual model
19
Methodology
Sample
The data used for answering the research question was collected by asking the participants questions about security concerns and their preference for mobile banking in app features in several conjoint choice sets. Five different types of features were tested in this study. Face verification and fingerprint verification are considered as security features. Push notifications, qr-payments and geolocation services are considered as non-security features. The different weights that this study found can be interpreted as the relative importance of the different types of features and thus the importance of security in mobile banking. The choice based conjoint has five attributes which are the independent variables with two levels (face verification, fingerprint verification, push notifications, qr-payments and geolocation services). The dependent variable in this study is the product adoption.
Conjoint set up
20 Table 1: Attribute and levels for conjoint analysis
ATTRIBUTES LEVELS
FACEVERIFICATION ● No face verification
● Authentication based on face verification
FINGERPRINTVERIFICATION ● No fingerprint verification
● Authentication based on fingerprint verification
PUSHNOTIFICATIONS ● No push notifications
● The app will send you push notifications
QR-PAYMENTS ● No option for QR-payments
● You can make payments based on QR codes
GEOLOCATIONSERVICES ● No geolocation services
● Geolocation services
Table 2: 5-point scale questions measuring general security concerns
General security concerns (Dinev, Hart 2005)
● I am concerned over the security of personal information on mobile applications
● I am concerned that my personal information may be shared with businesses without my consent
21 The design of this conjoint study is based on a random selection of stimuli. To keep the participants engaged and to avoid hypothetical bias the participants were exposed to only ten choice sets. According to Eggers et al. (2018) choice sets should have minimal overlap since alternatives that have the same level of an alternative provide no information on the preference for that attribute. So, the sets are going to be ordered to ensure there is a minimal overlap. Furthermore, another element that increases the realism is to include a so-called no-choice option, which can be chosen if none of the alternatives are acceptable.
Figure 1: Example conjoint choice-set pictured as seen in the survey
Model type
22 added in order to test for moderation effects. Because there are only two levels in the dependent variable and in the independent variables, using a binary logistic regression sounds most likely. However, a multinomial logistic regression is used to estimate the models. Because this depends on how many options the respondent had to choose from in the choice set. A binomial logit regression is used when there are two alternatives in the choice set. In this survey there were 4 alternatives to choose from, so a multinomial logit model is used. The multinomial logit model is mostly used for analysing conjoint choice data. The greatest asset of this model is its simple form for the choice probabilities The binary choice data that results from such conjoint choice experiments is normally analysed with the multinomial logit model (e.g., Louviere & Woodworth 1983; Elrod et al., 1992) using weighted regression techniques or maximum likelihood. Moreover, all choices made by a respondent, within a multinomial logit, are considered as independent observations. Hence, this estimation brings the assumption that a respondent starts each new choice set with an objective mind and that there are no correlations between the choice sets (Haaijer,1999).
Because of the fact that there are only two levels for each attribute, a partworth model is used to examine the different utilities. This means there is no assumed functional relationship between the attribute levels. In order to get the most efficient design, the conjoint is based on a balanced and orthogonal design. A balanced design means that each level is displayed an equal number of times and orthogonal means each level combination appears an equal number of times to prevent correlation between attributes. Furthermore, the full factorial design is used in the survey. All the analysis will be done in R Studio.
Table 3: Variables
Variable
Selection Dummy Binary
Face verification Categorical (effect coded)
23
Push Notifications Categorical (effect coded)
Qr-Payments Categorical (effect coded)
Geolocation services Categorical (effect coded)
Security concerns Ordinal (likert 1-5)
Model validation
A likelihood ratio test will be done and the McFadden R2 will be calculated to assess the goodness of fit. The estimated model will be compared to the NULL-model in order to see if there is a difference between the models. The H0 states that there is no difference between the two models. When there is a difference, the H0 can be rejected and this means that the parameters in the estimated model are significantly different from zero and the model fit is good. The chi-square test statistic is used to test the hypothesis with the following formula:
24
Results
Data collection
The data used in the analysis was collected through a survey from Preference Lab. The survey was designed with questions to gather demographic information and questions for the conjoint analysis. Most people were approached by social media or at work. Studies about conjoint analysis for commercial purposes advise to gain between 100 to 1000 respondents. Research with a lower number of attributes, requires between 100 to 150 participants (Cattin And Wittink, 1989). Sample
The total number of respondents to the survey was 191. Five respondents were removed because they had an unrealistic completion time of less than 2 minutes. After deleting these respondents, the model was estimated with (n=186) respondents. The sample contained 83 men and 103 women. Furthermore, almost everyone downloaded a mobile banking app except for two. The highest level of education of the respondents is well spread. Most respondents in the sample have completed a HBO degree (77) or a university degree (74). Furthermore, 28 people had an MBO degree and 7 people had only a high school degree. From figure 2 it can be seen that exactly 50 percent (93 respondents) have an age between 25 and 40 years. From the sample 34 were between 41 and 55 years and 48 respondents were between 16 and 24 years old. Only 11 people were between 56 and 70 years. Besides age and education, the income of the respondents had been asked. Most people of the sample (116) said they had an average income, 22 respondents a low and 27 respondents a high income. 21 people didn’t want to tell their income.
25
Variable Frequency Percentage
Gender - Male - Female 103 83 44.6 % 55.4 % Age - 16-24 - 25-40 - 41-55 - 55-70 48 93 34 11 25.8 % 50.0 % 18.3 % 5.9 % Education Level
- High school degree - MBO - HBO - University degree 7 28 77 74 3.8 % 15.1 % 41.4 % 39.8 % Income Level - Low - Medium - High
- Don’t want to tell
22 116 27 21 11.8 % 62.4 % 14.5 % 14.5% Table 4: Descriptives Model estimation
In order to determine the direct effects, a model was estimated without taking any moderators into consideration. The response variable in all models was a dummy variable, which indicated for each alternative in a choice set whether it was selected or not. Based on this, a multinomial logistic regression model was applied throughout the analysis. This model type accounted for the fact the response variable was not continuous. In the estimated model the selection dummy was the independent variable and face verification, fingerprint verification, push notifications, qr-payments and geolocation services as dependent variables.
26 Chi-square of 393.36. The Chi-square table gives a p-value of p=0.001 (df=10, critical value=29.29). This means that the parameters in the model are significantly different from zero and the main model outperforms the NULL model. Furthermore, the (adjusted) McFadden-R2 is (0.1928) which indicates a good model fit.
Attribute importance main effects model Now that it is proven that the estimated model is an appropriate model, we can start interpreting the utilities of the different attributes. Since utilities can only be interpreted relative to the other utilities of the same attribute, and not across attributes, only an indication of their effects can be given. In table 5 the outcomes of the attribute importance are shown. The relative importance of the attributes is calculated by dividing the range of the attribute coefficients by the sum of all the ranges (Eggers et al., 2018). The outcomes of these calculations are the measures of how much influence each attribute has on people’s choices. In figure 3 it can be seen that fingerprint verification had the most influence on people’s choices in the conjoint setup. Followed by geolocation services and face verification.
Table 5: Importance of variables
27 Figure 3: Importance of variables
Conjoint analysis main effects model
28 terms of the sign of its coefficient relative to the reference level, rather than the absolute value. The results in table 5 reveal that consumers have a high preference for the none option (𝛽= 1.71), which is remarkable. The none option can be compared to the sum of part-worth utilities across the other attributes. However, that total utility of the attributes doesn’t exceed the utility of the none option, thus the none option is more likely to be chosen.
Table 6: Estimation results
Beta Standard Error
Face Verification 0.3981** 0.062 Fingerprint Verification 0.7660** 0.063 Push Notifications -0.0612 0.061 QR-Payments 0.3078** 0.062 Geolocation Services -0.6089** 0.063 None Option 1.7130** 0.225 Significance codes: 0 ’***’ 0.001 ’**’ 0.01 ’*’ 0.05 ’.’ 0.1 ‘ ‘ 1
Hypothesis 1: Face verification in mobile banking as compared to mobile banking without face verification will have a positive effect on the product adoption
Based on the results, it can be seen that there is a positive estimate for face verification. Compared to the other attributes, face verification has the second biggest positive effect (𝛽=0.3981) on the product adoption with a highly significant result (p<0.001). Hence, the positive and significant estimate implicates that face verification increases the ease of use and usefulness. Furthermore, face verification shows to be an important variable as well looking at the variable importance (22.60 %). Therefore, H1 is supported.
Hypothesis 2: Fingerprint verification in mobile banking as compared to mobile banking without
29 As expected from the literature, the results show a positive and relative strong effect (𝛽=0.7760) of fingerprint verification on production adoption Besides, fingerprint verification comes along with a high significant effect (p<0.001). Hence, the probability of product adoption in mobile banking becomes bigger when fingerprint verification is offered compared to mobile banking without fingerprint verification. Taking all other variables into account, fingerprint verification has the strongest effect because it’s found to be the most important predictor of product adoption (35.7%). Although this was expected, it’s still interesting to find out why fingerprint verification has a stronger effect than face verification. Linking these outcomes to the TAM-model, it’s plausible that fingerprint verification is easier to use than face verification. Which can lead to more ease of use and usefulness. Thus, hypothesis 2 is supported as well.
Hypothesis 3:Push notifications in mobile banking as compared to mobile banking without push notifications will have a positive effect on the product adoption
No significant effect (𝛽=0.3079, p=0.318) is found between push notifications and product adoption. This means we can reject hypothesis 3. Besides, push notification has the smallest variable importance compared to the other variables (2%). Further research will be necessary in order to confirm this hypothesis. The results from the variable importance indicate that the benefits provided by push-notifications in mobile banking are not considered to be very valuable by the consumer.
Hypothesis 4: The possibility to make QR-payments in mobile banking as compared to mobile banking without the possibility to make QR-payments will have a positive effect on the product adoption
30 Hypothesis 5: Geolocation services in mobile banking as compared to mobile banking without
geolocation services will have a negative effect on the product adoption
Geolocation services show a significant (p<0.001) but negative effect (𝛽=-0.6089) on the product adoption. This variable is the only variable compared to the other variables with a negative effect instead of positive. However, geolocation services has the second biggest variable importance with 28.4 %. Since geolocation has a negative effect, these outcomes indicate that geolocation services can be considered to be too intrusive. However, other research showed that using the customers’ locations can be advantageous for firms, but it is important to keep in mind the customers experience this advantage as well. Thus, hypothesis 4 can be confirmed.
Estimating moderation effects
Afterwards, the influence of general security concerns was investigated. As mentioned, these security concerns were predicted to act as a moderator in the analysis that was conducted. In order to determine the effect of security concerns properly, it was examined whether the items on a 5-item scale were internally consistent. Furthermore, Cronbach’s alpha was calculated to check whether these items could be combined in one construct and was assessed to determine the internal reliability of the constructs. The test results showed a value of 0.813, indicating high internal consistency. The threshold for Cronbach’s alpha is 0.7 which means the average of these three items can be combined in one construct. This new construct is later used in the multinomial logit models. Furthermore, in table 7 the correlation between the different questions can be seen. Table 7: Correlation matrix Security Concerns
Sec1 Sec2 Sec3
Sec1 1 0.6242 0.5931
Sec2 0.6242 1 0.6145
31 Table 8: Model with interaction of moderator
Beta Standard Error
Security Concerns * Face Verification 0.2108*** 0.061 Security Concerns * Fingerprint Verification -0.1586*** 0.059 Security Concerns * Push Notifications 0.1487 0.064 Security Concerns * QR-Payments -0.072** 0.055 Security Concerns * Geolocation Services -0.1811*** 0.062 Security Concerns * None Option 0.3690*** 0.061 Significance codes: 0 ’***’ 0.001 ’**’ 0.01 ’*’ 0.05 ’.’ 0.1 ‘ ‘ 1
32 features. These results indicate face verification (security feature) can be useful when people have a higher degree of security concerns. Contradictory fingerprint verification (security feature) reveals to have a decrease in preference when consumers have security concerns.
Table 9: Summary results
Hypothesis
H1
Face verification in mobile banking as compared to mobile banking without face verification will have a positive effect on the product adoption.
Supported
H2 Fingerprint verification in mobile banking as compared to mobile
banking without fingerprint verification will have a positive effect on the product adoption.
Supported
H3 Push notifications in mobile banking as compared to mobile banking without push notifications will have a positive effect on the product adoption
Rejected
H4 The possibility to make QR-payments in mobile banking as compared to mobile banking without the possibility to make QR-payments will have a positive effect on the product adoption
Supported
H5 Geolocation services in mobile banking as compared to mobile banking without geolocation services will have a negative effect on the product adoption
Supported
H6a The extent to which a person holds security concerns moderates the effect of face verification on the product adoption negatively
Rejected
H6b The extent to which a person holds security concerns moderates the effect of fingerprint verification on the product adoption negatively
33 H6c The extent to which a person holds security concerns moderates the effect
of push notifications on the product adoption negatively
Rejected
H6d The extent to which a person holds security concerns moderates the effect of qr-payments on the product adoption negatively
Supported
H6e The extent to which a person holds security concerns moderates the effect of geolocation services on the product adoption negatively
34
Conclusion and Discussion
The increasingly rapid digitalization in our world in the last few years in the financial service industry has made it harder to retain our security. This study tried to shed some light on the acceptance of new features and the reaction of consumers to these new changes. This study also highlighted that TAM is the model considered when it comes to user perception towards the product adoption of this technology. Managers should have better knowledge on the consequences of their security strategy and, in particular, the impact of the type of features they offer. The aim of this analysis was to add knowledge to the literature regarding the product adoption of different mobile banking features with security concerns taken into account and understand the user’s perception towards the acceptance of biometrics in mobile banking. The goal was to find how consumers react to different types of features which are divided into security features and non-security features. In this study the so-called biometric verification methods, like face verification and fingerprint verification are considered as security features. The other three features added in this study are considered as non-security features. The results from the conjoint study give an overview of the consumers’ preferences towards mobile banking. Furthermore, general security concerns were introduced as a moderation variable.
35 watched and don't feel comfortable with that idea. Hence, banks and other companies should highlight the advantages of this feature in order to create more convenience.
The fact that in this research the security features are preferred over the non-security features does not mean that the non-security features itself are not valuable. Especially in the setting that was used in the conjoint set up, where the respondent had to choose his most preferred option in the choice set. This is not the case in real life and customers don’t have to make these decisions. It is possible that respondents of the survey selected face verification and fingerprint verification because they compared the features by considering the ease of use. Choosing the features which increases ease of use and usability sounds reasonable for this outcome.
Besides the most preferred feature in mobile banking, we were also interested in the product adoption when considering security concerns of the customer. Results show that for fingerprint verification, qr-payments and geolocation services a negative moderating effect of security concerns on product adoption. On the other hand, face verification shows a positive moderating effect of security concerns. Meaning that customers who have more security concerns, seem to value the security feature face verification as more important, and perceive the other security feature fingerprint verification as totally unimportant. Based on these outcomes, it is plausible to say face verification in mobile banking increases the perceived value of security without compromising it. Possibly, face verification has a higher acceptability rate than fingerprints among customers. Face verification is universal in nature as every person has facial features but not everyone has fingerprints or may lose them due to an accident. Moreover, when it comes to the trade-off between ease of use and security. Facial verification could eliminate this trade-off. It is a fast and secure method. It allows many businesses and other financial institutions to verify their customers within seconds and can eliminate the threat of identity fraud.
36 are still reserved to use this technology. Especially in a setting where it comes to their money. Understanding the antecedents that are the determinants of customers’ intention to use and adopt such technology is crucial for banks and other financial services. Using TAM based on (Davis, 1989), this study addresses the issue of the relation between different features to product adoption combined with the role of security concerns. Banks provide these security features in their drive to better identify new customers, secretly authenticate existing customers and protect high-value transactions and combat fraud. However, the perceived security is something subjective for the customers. The real issue is customers’ lack of understanding of how biometrics work. While people are fed up with passwords and want an easier, more convenient solution to protect their bank account information, the human tendency is to stick with things they are familiar with when the alternative is unknown and feels risky. This means banks have to convince the customers even more to adopt these new features and give them confidence that using these apps are really secure. By doing this, banks do not only increase objective security, but subjective security as well.
Limitations and recommendations for further research
37
References
Ahmed, I. U. (2017), Smartphone Authentication: User experience, expectation and satisfaction, Master’s thesis, Lappeenranta University of Technology. Available at: http://lutpub.lut.
fi/bitstream/handle/10024/143330/mastersthesis_ahmed_imtiaz
Albashrawi, M., & Motiwalla, L. (2019). Privacy and personalization in continued usage intention of mobile banking: An integrative perspective. Information Systems Frontiers, 21(5), 1031-1043.
Allen, I. E., & Seaman, C. A. (2007). Likert scales and data analyses. Quality progress, 40(7), 64-65.
Al-Jabri, I. M., & Sohail, M. S. (2012). Mobile banking adoption: Application of diffusion of innovation theory. Journal of Electronic Commerce Research, 13(4), 379-391.
Babich, N. (2016). Mobile UX design: What makes a good notification? UX Planet. Available at: https://uxplanet.org/how-to-craft-mobile-notifications-that-users-actually-want-
7b585e0e1fa1#.m06hb78gp
Balapour, A., Nikkhah, H. R., & Sabherwal, R. (2020). Mobile application security: Role of perceived privacy as the predictor of security perceptions. International Journal of Information
Management, 102063.
Bansal, G. (2017). Distinguishing between privacy and security concerns: An empirical examination and scale validation. Journal of Computer Information Systems, 57(4), 330–343. https://doi.org/10.1080/08874417.2016.1232981.
Bhagavatula, C., Ur, B., Iacovino, K., Mon Kywe, S., Cranor, L. & Savvides, M. (2015), Biometric authentication on iphone and android: Usability, perceptions, and influences on adoption, in ‘Workshop on Usable Security at USEC ’15’. doi: 10.14722/usec.2015.23003.
Bhatia, R. (2013). Biometrics and face recognition techniques. International Journal of Advanced Research in Computer Science and Software Engineering, 3(5).
Bowen, K. & Pistilli, M. (2012). Student preferences for mobile app usage. Center for applied research, 1-12.
Bustard, J. The impact of eu privacy legislation on biometric system deployment: Protecting citizens but con- straining applications. IEEE Signal Processing Magazine, 32(5):101–108, 2015. 6
Cao, K., & Jain, A. K. (2016). Hacking mobile phones using 2D printed fingerprints. Department of Computer Science and Engineering, Michigan State University.
Capgemini & Efma. (2016) World retail banking report. Available at:
38 Chen, C. C., & Tsai, J. L. (2019). Determinants of behavioral intention to use the Personalized Location-based Mobile Tourism Application: An empirical study by integrating TAM with ISSM. Future Generation Computer Systems, 96, 628-638.
Clor-Proell, S., Guggenmos, R., & Rennekamp, K. M. (2019). Mobile Devices and Investment Apps: The Effects of Information Release, Push Notification and the Fear of Missing Out. Push Notification and the Fear of Missing Out (August 22, 2019).
Cooper, I., & Yon, J. (2019). Ethical Issues in Biometrics. Sci Insigt, 30(2), 63-69.
Dalcher, I., & Shine, J. (2003). Extending the new technology acceptance model to measure the end user information systems satisfaction in a mandatory environment: A bank's treasury. Technology Analysis & Strategic Management, 15(4), 441-455.
Davis, F.D. (1989) ‘Perceived usefulness, perceived ease of use, and user acceptance of information technology’, MIS Quarterly, Vol. 35, pp.319–340.
Davis, F.D. (1993) ‘User acceptance of information technology: system characteristics, user perceptions and behavioral impacts’, International Journal of Man-Machine Studies, Vol. 38, No.3, pp. 475–487.
DINEV, T. and HART, P., 2005. Internet Privacy Concerns and Social Awareness as Determinants of Intention to Transact.International Journal of Electronic Commerce, 10(2), pp. 7-29.
Dootson, P., Beatson, A. & Drennan, J. (2016). Financial Institutions using Social Media – Do Consumers Perceive Value? International Journal of Bank Marketing, 34(1), 9-36.
Eggers, F., & Sattler, H. (2011). Preference measurement with conjoint analysis. Overview of state-of-the-art approaches and recent developments. GfK Marketing Intelligence Review, 3(1), 36-47.
Elrod, T., J.J. Louviere and K.S. Davey (1992), “An Empirical Comparison of Rating-Based and Choice-Based Conjoint Models,” Journal of Marketing Research, 29(3), 368-377.
Eggers, Felix, Sattler, Henrik, Teichert, Thorsten, Völckner, Franziska (2018). Choice-Based Conjoint Analysis. Handbook of Market Research, Springer.
Flavián, C., Guinalíu, M., & Gurrea, R. (2006). The influence of familiarity and usability on loyalty to online journalistic services: The role of user experience. Journal Of Retailing And Consumer Services, 13(5), 363-375.
39 Friedman, B., Khan, P. H., & Howe, D., C. (2000). Trust on line. Communications of the ACM, 43 (12), 34-40.
Gefen, D., Karahanna, E. and Straub, D. (2003), “Trust and TAM in online shopping: an integrated model”, MIS Quarterly, Vol. 27 No. 1, pp. 51-90
Gerrard, P., & Cunningham, J. B. (2003). The diffusion of Internet banking among Singapore consumers. International Journal of Bank Marketing, 21(1), 16-28.
Haaijer, M. (1999). Modeling conjoint choice experiments with the probit model (Doctoral dissertation, University of Groningen).
Harris, M., Chin, A., & Beasley, J. (2019). Mobile Payment Adoption: An Empirical Review and Opportunities for Future Research.
IAB. (2015). Apps and mobile web: Understanding the two sides of the mobile coin. Available at:http://www.iab.com/wp-content/uploads/2015/07/IAB_Apps_and_Mobile_Web_Final.pdf
Johnson, M., & Verdegaal, M. (2016). How traditional banks are innovating the basics to provide customers with an Uber-like mobile banking experience. Journal of Digital Banking, 1(1), 33-44.
Johnston, A. C., & Warkentin, M. (2010). Fear appeals and information security behaviors: An empirical study. MIS Quarterly, 34(3), 549–566. https://doi.org/10.2307/ 25750691.
Jun, M., & Palacios, S. (2016). Examining the key dimensions of mobile banking service quality: an exploratory study. International Journal of Bank Marketing.
Kahuna. (2014). 5 ways push notifications protect user experience. Kahuna. Available at: https://www.kahuna.com
Karma, N. G., Ibrahim, S. B., & Ali, A. H. (2014). Key factors affecting mobile banking adoption among banks’ customers in Sudan. International Journal of Liberal Arts and Social Science, 2(6), 112-122.
Kim, C., Oh, E., Shin, N., & Chae, M. (2009). An empirical investigation of factors affecting ubiquitous computing use and U-business value. International Journal of Information Management, 29(6), 436-448.
Lee, E.-J., Kwon, K.-N., & Schumann, D. W. (2005). Segmenting the non-adopter category in the diffusion of Internet banking. International Journal of Bank Marketing, 23(5), 414-437.
40 Louviere, J.J., & Woodworth, G. (1983). Design and Analysis of Simulat- ed Consumer Choice or Allocation Experiments. An Approach Based on Aggregated Data. Journal of Marketing Research, 20(4), 350-367.
Lu, J., J. E. Yao, and C. Yu. 2005. “Personal Innovativeness, Social Influences and Adoption of Wireless Internet Services via Mobile Technology.” The Journal of Strategic Information Systems 14 (3): 245–268.
Marangunić, N., & Granić, A. (2015). Technology acceptance model: A literature review from 1986 to 2013.Universal Access in the Information Society, 14(1), 81–95.
Martin-Flatin, J. (1999). Push vs. pull in web-based network management. In Sixth IFIP/IEEE International Symposium on Integrated Network Management. (pp. 957 - 958). Available at: http://ieeexplore.ieee.org/document/770671/
Martin, C., Snabe, J. H., & Nanterme, P. (2017). Digital Transformation Initiative In collaboration with Accenture for Business and Society. World Economic Forum.
Mayron, L., Hausawi, Y. & Bahr, G. (2013), Secure, usable biometric authentication systems, in ‘UAHCI 2013: Universal Access in Human-Computer Interaction. Design Methods, Tools, and Interaction Techniques for eInclusion’, pp. 195–204. doi: 10.1007/978-3-642-39188-0_21.
McCloskey, D.W. (2006), “The importance of ease of use, usefulness, and trust to online consumers: an examination of the technology acceptance model with older consumers”, Journal of Organizational and End User Computing, Vol. 18 No. 3, pp. 47-65.
McCole, P., Ramsey, E., & Williams, J. (2010). Trust considerations on attitudes towards online purchasing: The moderating effect of privacy and security concerns. Journal of Business Research, 63(9-10), 1018-1024.
McFadden, D. (1981). Econometric Models of Probabilistic Choice. In C. Manski, & McFadden, D. (eds.), Structural Analysis of Discrete Data (pp. 198-272). Cambridge: MIT-Press
Morosan, C. and Jeong, M. (2008), “Users’ perceptions of two types of hotel reservation web sites”, International Journal of Hospitality Management, Vol. 27 No. 2, pp. 284-92.
North, E., & De Vos, R. (2002). The use of conjoint analysis to determine consumer buying preferences: A literature review. Journal of Consumer Sciences, 30(1).
Nysveen, H., Pedersen, P.E. and Thorbjørnsen, H. (2005a) ‘Explaining intention to use mobile chat services: moderating effects of gender’, Journal of Consumer Marketing, Vol. 22, No. 5, pp.247–256.
41 Olmstead, K. (2014). Mobile apps collect information about users, with wide range of permissions. Pew Research Center, Press Release, 29.
Patrick, A. (2008), Fingerprint Concerns: Performance, Usability, and Acceptance of Fin-
gerprint Biometric Systems. Unpublished. Available at: https://tinyurl.com/patrick-fingerprintconcerns.
Pentina, I., Zhang, L., Bata, H., & Chen, Y. (2016). Exploring privacy paradox in information-sensitive mobile app adoption: A cross-cultural comparison. Computers in Human Behavior, 65, 409-419.
Promatics Technologies, (2018). Geolocation in mobile apps: How brands are using it for marketing? Available at: https://medium.com/promatics
Rogers, E. M. (2002). Diffusion of preventive innovations. Addictive behaviors, 27(6), 989-993.
Safeena, R., Date, H., Kammani, A., & Hundewale, N. (2012). Technology adoption and Indian consumers: study on mobile banking. International Journal of Computer Theory and Engineering, 4(6), 1020.
San Martin, S., Camarero, C. (2009). How perceived risk affects on line buying. On line information review, 33 (4), pp 629-654.
Saxena, D. (2019). Review of the literature on Opportunities and constraints in adoption of electronic banking. International Journal of Management, IT and Engineering, 7(12), 154-165.
Shin, D.-H. (2009). Towards an understanding of the consumer acceptance of mobile wallet. Computers in Human Behavior, 25(6), 1343–1354.
Sundar, S. S., Xu, Q., & Bellur S. (2010). Designing interactivity in media interfaces: A communications perspective. Proceedings from CHI 2010 Perspectives on Design.Atlanta, GA, April 10-15.
Svendsen, G. B., Johnsen, J. A., Almås-Sørensen, L., & Vittersø, J. (2013). Personality and technology acceptance: the influence of personality factors on the core constructs of the Technology Acceptance Model. Behaviour & Information Technology, 32(4), 323-334.
Statista, (2020). Number of smartphone users worldwide 2016-2020. Available at: https://www.statista.com
Streeter, B. (2019). 16 Must-Have Mobile Banking Features that Raise the CX Bar. Available at: https://thefinancialbrand.com/91274/mobile-banking-app-features-cx-alert-biometric-savings-payment/
42 Thakur, R., & Srivastava, M. (2014). Adoption readiness, personal innovativeness, perceived risk and usage intention across customer groups for mobile payment services in India. Internet Research.
Urban, J. and Hoofnagle, C., 2014. The privacy pragmatic as privacy vulnerable.
Vizard, L. (2016). Notification overload: Best practices for designing notifications with respect for users | Creative Cloud blog by Adobe. Adobe Creative Cloud. Available at: https://blogs.adobe.com/creativecloud
Walker J., Ben-Akiva, M. (2002). Generalized Random Utility Model.Mathematical Social Sciences, 43(3), 303-343.
Wang, T., C. H. Jung, M. H. Kang, and Y. S. Chung. 2014. “Exploring Determinants of Adoption Intentions Towards Enterprise 2.0 Applications: An Empirical Study.” Behaviour and Information Technology. 33 (10): 1048–1064.
Weis, E. (2019). Grand View Research Predicts $59.31 Billion Biometrics Market by 2025. Available at: https://findbiometrics.com
Westermann, T., Möller, S., & Wechsung, I. (2015). Assessing the relationship between technical affinity, stress and notifications on smartphones. Proceedings in 17th International Conference on Human-Computer Interaction with Mobile Devices and Services Adjunct, 652-659. Copenhagen, Denmark, August 24 – 27. New York: ACM.
Wittink, D. R., & Cattin, P. (1989). Commercial use of conjoint analysis: An update. Journal of marketing, 53(3), 91-96.
Yang, S., Y. Lu, S. Gupta, Y. Cao, and R. Zhang. (2012). “Mobile Payment Services Adoption Across Time: An Empirical Study of the Effects of Behavioral Beliefs, Social Influences, and Personal Traits.” Computers in Human Behavior 28 (1): 129–142.
Zhou, T. (2010). The impact of privacy concern on user adoption of location-based services. Industrial Management & Data Systems, 111(2), 212–226.
ZHAO, L., LU, Y. and GUPTA, S., 2012. Disclosure intention of location-related information in location-based social network services. International Journal of Electronic Commerce, 16(4), pp. 53-90.
