A Customer-Centric View of E-Commerce Security and Privacy

Abstract

Business organizations have huge potential to increase their customer base by offering e-commerce services, especially in the post-pandemic era. Ensuring secure e-commerce applications plays an important role in increasing customer base. To develop appropriate policies and secure technological infrastructures, business organizations first need to establish an understanding of the reservations of their customers toward e-commerce, as well as their perception of security and privacy of e-commerce applications. In this paper, we present the results of an empirical study of e-commerce customers conducted in Pakistan to gain an insight into their mindset on using e-commerce applications. An online questionnaire was set up to collect data, which were analyzed using the partial least squares method with SmartPLS software. The empirical findings highlight that customers’ concerns about credit card usage, concerns over information security, motivational factors for shopping offered by business organizations, customer trustworthiness, and user’s feelings about the reputation of e-commerce impact their perception of security of online data and trust in an e-commerce application. The results of this study can help organizations in Pakistan to develop policies and improve technological infrastructures by adopting emerging technologies and digital forensics.

1. Introduction

E-commerce is the use of computing devices and communication platforms to conduct business online. It has proven to be an important enabler for business organizations by expanding their reach to global customers, cost-effectively and efficiently. This technological revolution is termed disruptive technology [1] as it has changed traditional business models and brought about new business approaches. This extensive digital transformation process requires changes in organizational processes, human resources, and technological infrastructure. To be successful, business organizations need to carry out a detailed analysis of their marketplace to develop an effective business strategy, which requires financial and human resources [2]. Furthermore, e-commerce adoption is also dependent on social, legal, economic, political, and technological factors [3]. The difference in technology accessibility, level of skills, security concerns, shopping behavior, purchasing power, cultural differences, and legislation also affect the acceptance of e-commerce [4]. As a result, the adoption of technological infrastructure by business organizations varies considerably across the globe.

Recently, COVID-19 has resulted in a rapid digital transformation in every sector and similarly, e-commerce has also gained enormous popularity and become a replacement for conventional brick-and-mortar stores during lockdown [5,6]. Akram et al. [7] highlighted that ease of use, trust, mobility, and customer involvement have resulted in an exponential increase in the use of mobile commerce during the pandemic. Gu et al. [8] highlighted that, due to COVID-19, customers have become more experienced with online shopping, and it has influenced their buying behavior. Guthrie et al. [9] observed that the purchase intentions of the customers regarding basic needs and other categories of products have seen a significant shift, as compared to pre-COVID-19 buying patterns. The ease of online shopping also brought significant security challenges for customers, as well as business organizations. Business organizations have been taking a variety of cybersecurity measures to secure their e-commerce infrastructure; however, Shu and Liu [10] highlighted that there is a need to further explore privacy-enhancing technologies and associated strategies to better understand the security implications. Customers are an important stakeholder for e-commerce business organizations, and, if they are not confident that a certain e-commerce platform is secure, they will not indulge in a business transaction. Therefore, customers’ perception regarding the security and trust in e-commerce is of paramount importance. Customer perception of security and trust is dependent on the national culture, which is built upon factors such as individualism, risk-taking behavior, and power structures [11]. Mohammed and Tejay [12] found that e-commerce acceptance, online safety perception, personal preferences, and privacy concerns are important factors for e-commerce adoption in a society. Clemons et al. [13] found that the reputation of business organizations affects the trust of customers differently across geographical regions.

Motivation

Differences in the banking system, legal framework, and user sensitivity toward security can impact the customer’s perception of security and privacy. Most studies conducted to understand customer perception originate from developed countries, such as the USA, the UK, Singapore, China, and Poland. These countries have a lot more exposure to technology, and they have been blessed with a more reliable banking systems and robust legal infrastructures [14,15,16,17]. Therefore, it is scientifically interesting to explore the factors determining customer perception regarding the security of e-commerce applications in a developing economy. Consequently, we conducted an empirical study in Pakistan, a developing nation, to understand how e-commerce customers in Pakistan perceive security related to e-commerce and trust in online shopping. Positive perception can help to foster secure behavior in e-commerce usage whereas a negative perception may lead to limited usage of e-commerce by consumers. We used the protection motivation theory (PMT) as a theoretical model for our study.

PMT was developed by Rogers [18] to better understand fear and explain how people react to it. Perceived vulnerability, severity, response efficacy, self-efficacy, and response cost are the key constructs of PMT, and their joint impact helps in determining intentions to protect against such a threat [18]. Perceived vulnerability refers to the likelihood of being affected by the threat, perceived severity highlights the severity of the threat in case of impact, response efficacy refers to the usefulness of mitigation strategies, self-efficacy refers to the ability to carry out mitigation strategies, and response cost refers to the cost incurred to execute mitigation strategies. This theory has been applied in many domains to explain the behavior of people in different situations such as health risks [19,20,21], flooding, and typhoons [22]. Similarly, PMT has also been used in different cybersecurity contexts [23,24,25]. We, thus, used this theory in our work to understand the customer’s perception regarding the security of e-commerce websites in Pakistan. We used perceived vulnerability, perceived severity, response efficacy, and response cost to understand the customer’s perception of security and privacy in Pakistan. The contribution of our study presents empirical insights into customers’ perceptions and trust in e-commerce in Pakistan. On the basis of the empirical data, we present a model to improve the trust and security perception of e-commerce among Pakistani customers.

The remainder of the paper is structured as follows: related work is discussed in Section 2 and research design is presented in Section 3; Section 4 presents the results, followed by a discussion in Section 5 and a conclusion in Section 6.

2. Related Work

Many studies have focused on the security aspects of e-commerce technological infrastructure, such as securing communication [26,27,28], secure payment mechanisms [29,30], and fraud protection [31,32]. There has, however, been an enhanced interest in customer perception of security and trust in e-commerce. Shah et al. [33] highlighted that buying intention of customers is dependent on trust and perceived benefits of the product, whereas the potential risk decreases customer trust. Liu and Li [34] identified that trust and social aspects influence customers while engaging in mobile commerce transactions. López Jiménez et al. [35] highlighted that the adoption of security seals on company websites results in increased sales, as well as increased web traffic. Dhende and Meshram [36] highlighted that intra-organization, inter-organization, and customer-to-organization data exchanges using advanced technological infrastructure result in security challenges. Tan [37] noticed that despite the advancements in e-commerce, a lack of trust for customers in business is a constant impediment to the adoption and realization of electronic commerce. E-commerce vendors need to develop trust because repairing the trust of customers is a very difficult and expensive activity. Zhang et al. [38] highlighted that, to repair customer trust, single relational strategies are not enough; rather, an integration of single transactional and integrative strategies should be deployed.

Turner et al. [39] highlighted that users with limited technical knowledge rely on the reputation of the websites, their interaction with the website, and third-party recommendations to keep them secure on the internet. Liebermann and Stashevsky [40] highlighted that demographic traits and behavior of use affect risks of using the internet, and they developed a detailed perceived risk map to make the users aware with potential risks. Belanger et al. [41] highlighted that customer perception of a merchant’s trustworthiness is a key factor that encourages customers to indulge in online or conventional business transactions. Furthermore, security, privacy, and usability are the factors that affect shopping intention. Yenisey et al. [42] experimented with simulated e-commerce applications to determine the factors that cultivate positive feelings among customers. McDonald and Cranor [43] highlighted that investing time to read privacy policies of websites is a form of investment made by users and, in return, they should receive payments as they are revealing information. Huang et al. [44] analyzed how knowledge, impact, and severity of risk, probability of occurrence, and awareness affect information security behavior. They further discussed computer experience and potential loss affect user behavior. Beldad et al. [45] found that potential risks influence users to read online privacy statements, and that older people are more inclined to read privacy statements than younger people. Furthermore, people with low education levels and internet experience also have a higher probability of reading privacy statements as compared to users having higher education and more internet experience. Bonera [46] highlighted that intention to make an online purchase is affected by usable design, customer’s perception of security, and usefulness of e-commerce websites. Chowdhury and Chowdhury [47] highlighted some other key factors that influence online shopping behavior including complex online shopping processes, lack of reliability, weak payment settlement systems, and traditional mentality of customers. Kasuma et al. [48] emphasized some motivations that lead to online shopping including convenience of placing orders online, timesaving, website design, and security. Zahrani [49] stressed that social influence, perceived usefulness ease of use, and trust significantly contribute to perceived risk and security, which ultimately affects consumer intention to use a credit card online. Alotaibi and Alshehri [50] conducted a study and found that men exhibit more secure behavior online as compared to women. McCormac et al. [51], however, argued that gender and age are not significant in information security behavior, but emotions and risk acceptance of users affect the information security behavior. Tang et al. [52] found in their research that governmental social media accounts can influence information security behavior among users.

Yazdanifard et al. [53] discussed that security and privacy concerns can result in negative e-commerce growth; thus, they conducted a literature review to document common reasons for privacy and security concerns. Jensen [54] argued that there is a need to include trust as an important component of security models to secure computer networks by developing quantitative and qualitative trust models. Saeed [55] carried out an empirical study on e-business adoption by expatriates in Saudi Arabia and highlighted that cultural differences between expatriate communities should be considered during the technical design stage to make systems more usable. Gupta and Nage [56] outlined a set of security concerns and associated techniques to minimize their impact in web-based business environments.

In addition to other benefits, digital transformation by business organizations also results in many security risks, not only for the organizations themselves but also for their customers. Mason [57] identified four core issues about personal information ownership: privacy, accuracy, property, and accessibility, which are critical for customers while engaging in business transactions online. Gull et al. [58] found that mobile commerce customers in Saudi Arabia expect more strict security mechanisms for a positive security perception of these platforms.

Khan et al. [59] discussed that blockchain can be used to enhance security in e-commerce by using smart contracts which ensure the privacy of transactions. Gouthier et al. [60] found the use of data analytics by e-commerce companies is dependent on customers’ willingness to disclose personal information. They stressed the importance of transparent data management policies which prove critical for the customers to willingly share personal data. Mashatan et al. [61] investigated privacy and security concerns of customers in the adoption of crypto payments and concluded that there is an innate need to enhance customers’ understanding of potential problems and security issues to increase crypto payments. Vuță et al. [62] conducted a systematic literature study and discussed that human factors are the main risk along with fear of cyber-attacks, and that artificial intelligence, blockchain, and machine learning technologies play a critical role in minimizing these risks.

Masyhuri [63] studied Amazon and e-Bay to identify essential elements to improve customer satisfaction with the e-retailing business. Alkis and Kose [64] carried out an empirical study across 29 European countries and presented their findings on customers with more protection against online privacy are more likely to make online purchases and companies providing a higher level of privacy attract more customers. Anshori et al. [65] carried out an empirical study with the users of Tokopedia application and found that security and privacy have a significant impact on the trust and perceived value of users, which in turn have a positive impact on reuse intention. Cebeci et al. [66] developed a secure protocol that was evaluated on electronic transaction 3D secure systems, and the results showed a reduction in security concerns, as well as low communication cost. Novita and Budiarti [67] carried out a study on “Shopee” application users and found that, even for satisfied customers, perceived security, trust, and privacy positively influence customer retention.

2.1. Protection Motivation Theory in Cybersecurity

More recently, protection motivation theory has been used in the cybersecurity literature to relate to how users respond to cybersecurity threats while working with information technology applications. Marett et al. [68] explored the behavior of social media users inline with the PMT and found that it explains the behavior of both types of users, those with serious security concerns and casual. Meso et al. [69] carried out an empirical study with PMT on college students to understand how knowledge gained regarding cybersecurity during lectures is applied in practice. Dang-Pham and Pittayachawan [23] used PMT in their empirical study at an Australian university to understand user intention while using their personal devices at home as compared to the university. They found differences in user behaviors toward malware avoidance within both contexts. Hanus and Wu [70] explored the impact of information security awareness on desktop security behavior and found that security awareness affects PMT constructs, perceived security, response efficacy, self-efficacy, and response cost. Verkijika [24] expanded on the PMT and included anticipated regret as a mediator among PMT constructs and intended behavior. He found that perceived security and perceived vulnerability have a positive influence on the anticipated targets. Chiu et al. [25] used PMT to examine the employees’ incentive to observe organizational information security policies. Similarly, Erhart [71] used PMT constructs to explain employees’ intention to comply with organizational security policies. Ganesh et al. [72] developed a game to increase user aware of mobile phone risks inline with the PMT, and they found that their game helped users to better tackle security issues while they used their mobiles. Mou et al. [73] contributed to information security discourse by showing how PMT and structural equation modeling can be used to verify different theories. These contributions highlight that PMT has been a central part in cybersecurity literature. The protection motivation theory examines how people respond to stressful situations against potential threats; on this basis, it could be forecasted and key indicators could be identified which result in changed behavior of people. Therefore, it is interesting to apply PMT in studying which factors contribute to users’ perception of security and privacy, so that appropriate control systems can be designed.

2.2. E-Commerce Research in Pakistan

Within the context of e-commerce in Pakistan, several studies have been published over the years. Khan et al. [74] stated that internet legislation, poverty, and the lack of knowledge and infrastructure are all major barriers for e-commerce in Pakistan. Arshad and Zaidi [75] found that consumer behavior and organizational growth positively correlate with e-commerce in Pakistan, and that limited integration with organizational objectives and weak internal capabilities can be a major threat to Pakistani business organizations. Ghouri et al. [76] identified that privacy, security, and trust are important success factors for e-commerce adoption by Pakistani businesses. Anjum and Chai [77] found that the ease of use of cash on delivery helps in improving the security perception of e-commerce consumers in Pakistan. Tanveer [78] discussed how honesty and fulfillment are important factors for customer relationship building for e-commerce vendors in Pakistan. Saeed et al. [79] carried out a usability evaluation of e-commerce portals in Pakistan and highlighted the dire need for major improvements in user interface and experience. Agren and Barbutiu [80] reported that women face more cultural barriers in e-commerce adoption than men in Pakistan. Ahmed [81] raised the point that the design and implementation of an effective e-commerce policy can help e-commerce growth in Pakistan. Amjad et al. [82] discussed that the main entrepreneurial challenges for e-commerce in Pakistan include a change-resistant culture, poor law enforcement, and unprofessionalism of vendors. Imtiaz et al. [83] argued that e-business adoption in Pakistan is relatively slow, and that privacy, security, and trust factors are prerequisites for improved e-business services in Pakistan.

2.3. Hypothesis Formulation

Even though there have been several studies carried out on e-commerce and related issues in Pakistan, there is no empirical study that focused on Pakistani customers’ security perception of e-commerce. National culture is an important factor that affects technology adoption practices; therefore, in this paper, we explore how customers in Pakistan perceive security and privacy issues relevant to e-commerce websites in Pakistan. The findings can help in better understanding the customer’s security concerns and in the subsequent design of appropriate controls. The hypotheses were rooted in the previous literature where we identified demographic factors gender [H1], age [H2], and computer proficiency [H3], threat awareness [H4], threat perception [H5, H6, H7, and H8], security concerns [H9], motivational factors [H10], trustworthiness [H11], and reputation [H12] as key parameters for understanding security perception. As a result, our questionnaire features dedicated questions focusing on these aspects. Therefore, 12 hypotheses were developed for our study, which are listed in

Table 1. List of hypotheses and associated constructs.

Nr.	Hypothesis	Relevant Study
H1	Gender has an impact on the customers’ perception of the security of online data and trust in e-commerce applications.	[50,51]
H2	The age of the customer has an impact on the customers’ perception of the security of online data and trust in e-commerce applications.	[40,51]
H3	Computer proficiency level has an impact on the customers’ perception of the security of online data and trust in e-commerce applications.	[51]
H4	Customers who read privacy policy statements of e-commerce websites and review security certificates have a positive perception of the security of online data and trust in e-commerce applications.	[43]
H5	Customers who have a positive perception of the security of online data and trust in e-commerce applications are less prone to hacking attacks.	[44]
H6	Customers who read privacy policy statements of e-commerce websites and review security certificates are less prone to hacking attacks.	[45]
H7	Customers’ rating of security aspects of e-commerce websites reflects the customer’s perception of the security of online data and trust in an e-commerce application.	[11,39]
	Authentication of users by e-commerce websites	[26,27,28]
	Authentication of vendors by customers	[11]
	Fraud protection provisions	[31,32]
	Provision of secure communications by e-commerce websites	[26,27,28]
H8	Customers’ credit card usage concerns reflect the customer’s perception of the security of online data and trust in an e-commerce application.	[49]
	Network security issues	[46,49]
	Operating systems’ security issues	[46,49]
	Applications’ security issues	[49]
	Unauthorized transaction issues	[29,30]
	Fraudulent transaction issues	[29,30]
	Disputed vendors	[33]
	Product quality issues	[33]
H9	Customers’ e-commerce usage concerns reflect the customer’s perception of the security of online data and trust in an e-commerce application.	[40,46]
	Concerns over private information provision	[57]
	Concerns over information misuse	[46]
	Concerns over lack of control on submitted information	[46]
	Concerns over unintended use of information by intruders	[46]
H10	Incentives proposed by business organizations for shopping have an impact on the customer’s perception of the security of online data and trust in an e-commerce application.	[47]
	Data transparency	[60]
	Encryption security	[60]
	Product return policy	[33]
	Dispute guarantee policy	[47]
	An understanding of the terms of conditions	[39]
	E-commerce platform’s reputation	[39]
H11	Customers’ trustworthiness of e-commerce platforms impacts the customer’s perception of the security of online data and trust in an e-commerce application.	[35,41,54]
	The availability of product pictures	[41,55]
	The provision of a privacy seal on the website	[41,42]
	The provision of search feature on the website	[41,42]
	The provision of security seal on the website	[35]
	Presence of branded product logos	[41,42]
H12	Users’ feelings about the reputation of e-commerce impact their perception of security of online data and trust in an e-commerce application.	[42,67]
	Users’ feelings about location disclosure	[42,51]
	Users’ feelings about browsing history disclosure	[42]
	Users’ feelings about personal data misuse	[42]

, along with the relevant literature. Some of the hypotheses comprised additional constructs, which are mentioned in the rows after the relevant hypothesis. The mapping of these constructs is provided in

Table 2. Parameters of the model.

PMT Constructs	Model Construct	Parameters
Perceived vulnerability	Consumer rating	CR1: Authentication of users by e-commerce websites CR2: Authentication of vendors by customers CR3: Fraud protection provisions CR4: Provision of secure communications by e-commerce websites
Response efficacy	Trustworthiness	T1: The availability of product pictures T2: The provision of a privacy seal on the website T3: The provision of search feature on the website T4: The provision of security seal on the website T5: Presence of branded product logos
Perceived severity	Credit card usage concerns	CC1: Network security issues CC2: Operating system’s security issues CC3: Applications’ security issues CC4: Unauthorized transaction issues CC5: Fraudulent transaction issues CC6: Disputed vendors CC7: Product quality issues
Response efficacy	Motivation factors	MF1: Data transparency MF2: Encryption security MF3: Product return policy MF4: Dispute guarantee policy MF5: An understanding of the terms of conditions MF6: E-commerce platform’s reputation
Response cost	Customer worries	CW1: Concerns over private information provision CW2: Concerns over information misuse CW3: Concerns over lack of control on submitted information CW4: Concerns over unintended use of information by intruders
Response cost	Customer feelings	CF1: Users’ feelings about location disclosure CF2: Users’ feelings about browsing history disclosure CF3: Users’ feelings about personal data misuse

.

3. Materials and Methods

The reason for selecting Pakistan as a case setting was based on the huge potential for e-commerce growth due to a large population. We were specifically interested in developing an understanding about how people perceive privacy and security aspects in a developing country, and how these challenges can be addressed by businesses and increase their online dealings. Pakistan is the fifth biggest nation in the world, with a population of more than 225.2 million; by geographical area, it is the 33rd largest country [84]. Pakistan is the 46th largest e-commerce market in the world, with annual sales of 4 billion USD in 2020; for the next 4 years, the annual growth rate is expected to increase by another 18% [85]. According to a survey based on data obtained in 2019, the provincial shares of the net e-commerce sales were as follows: the province of Punjab had a 55% share, followed by Sindh with a 36% share, whereas other regions, Khyber Pakhtunkhwa, Balochistan, and Azad Kashmir, contributed to the remaining 9% of total e-commerce sales [86].

Our study is based on quantitative data, which are available in Appendix A, and we used PMT as a theoretical framework [18]; hence, four constructs of PMT, namely, perceived vulnerability, perceived severity, response efficacy, and response cost, were used in our questionnaire. For each of the constructs, we designed closed-ended questions on different aspects of user perception, some of which were rooted in earlier studies [87,88,89]. Once the questionnaire was complete, it was reviewed by two colleagues to ensure content validity. After validation, the survey was posted on Google forms. To recruit respondents, we used the nonprobability sampling method and used the snowball sampling technique by contacting our peers; the qualification criterion for the respondent was the use of e-commerce websites for online purchasing. By the end of the survey period, we received 187 responses, of which we discarded one response as the concerned respondent reported not having experience with online shopping. There were 135 male respondents and 51 female respondents, which were divided into four age groups. There were 62 respondents in the age group of 19–25 years, 51 respondents in the age group of 26–35 years, 58 respondents in the age group of 36–50 years, and 15 respondents were above 50 years old. While looking at computer knowledge, among the 186 respondents who participated in the study, 22 were beginner-level users, 63 were middle-level users, and 101 rated themselves as advanced users. The individual questions were mainly based on a five-point Likert scale (strongly disagree, disagree, neutral, disagree, and strongly disagree). Different questions were mapped to relevant constructs. Since SmartPLS only processes numeric data, the complete data file was coded into numeric data and then processed using SmartPLS [90] to identify the relevance of constructs in the model outcome. During our analysis, we applied the bootstrapping procedure, where we used the partial least squares method with 5000 iterations. The hypotheses were approved and rejected according to p-values.

4. Results

In this section, we discuss the data gathered during our study and the analysis we ran on it. We created a SmartPLS model for hypothesis testing, which is shown in Figure 1. In this basic model, we had latent variables of gender, age, computer proficiency, affected by the hacking, and customer’s habit of checking security certificates and reading privacy certificates to understand how they affect online data security and trust in e-commerce. After the application of blindfolding, we obtained Q² values of 0.046 for online data security and trust in e-commerce and a value of 0.049 for the hacking victim variable; as proposed by Haier et al. [91], a value above zero indicates the model’s predictive relevance.

To understand the detailed factors having an impact on customer perception of security and privacy, we developed a comprehensive model, as shown in Figure 2. There were six main constructs of the models, namely, consumer rating, trustworthiness, credit card usage concerns, motivation factors, customer’s concerns, and customer feelings, while using the e-commerce infrastructure. The constructs were developed on the basis of different sections of our questionnaires, where each section contained a set of related questions. Each of these constructs included additional parameters highlighted in Table 2. These parameters were based on individual questions within the questionnaire. After running the analysis with the partial least squares function on SmartPLS, we found that T5, CC4, CC5, CC6, and CC7 had values less than 0.7, which means that users gave lower ratings for these factors and they were not significant factors; hence, we removed these parameters and constructed a modified model as shown in Figure 3.

Table 3. Construct consistency and reliability values.

	Mean	Std Deviation	Cronbach’s Alpha	rho_A	Composite Reliability	Average Variance Extracted (AVE)	Q2 (1 − SSE/SSO)
Consumer rating	0.128	0.085	0.942	0.994	0.958	0.851	0.728
Trustworthiness	0.265	0.041	0.832	0.832	0.889	0.668	0.44
Credit card usage concerns	0.268	0.033	0.784	0.801	0.873	0.696	0.383
Motivation factors	0.305	0.039	0.912	0.915	0.933	0.699	0.576
Customer worries	0.42	0.048	0.899	0.91	0.93	0.768	0.6
Customer feelings	0.191	0.056	0.894	0.902	0.935	0.827	0.618

highlights the mean and standard deviation values for the constructs used in the model, which range from 0–1, showing the contribution of individual constructs to the model outcome. As shown in Table 3, the construct values were acceptable in terms of internal consistency and reliability. Cronbach’s alpha value was higher than 0.7 for all constructs, showing a good internal consistency [91]; similarly, a rho value above 0.8 for all constructs confirmed the internal consistency [92]. A composite reliability above 0.7 shows good reliability, and all of our constructs achieved these values. In the case of average variance extracted (AVE), all our constructs scored greater than 0.5, which highlights the reliability of our model [93]. Q² values greater than 0 indicate that the model has good predictive relevance [94]. After blindfolding, we obtained a Q² value of 0.313 for the customer perception of security and privacy; thus, as per Hair et al., the model has predictive relevance.

The results of the path coefficient analysis are shown in

Table 4. Hypothesis testing results.

Hypothesis	p-Value	Result
H1	0.006	Approved
H2	0.016	Approved
H3	0.999	Rejected
H4	0.829	Rejected
H5	0.000	Approved
H6	0.389	Rejected
H7	0.113	Rejected
H8	0.000	Approved
H9	0.000	Approved
H10	0.000	Approved
H11	0.000	Approved
H12	0.000	Approved

. A p-value of less than 0.05 is normally considered as strong evidence against the null hypothesis; therefore, in our case, hypotheses 1, 2, and 5 were approved by our analysis. This implies that gender and age have an impact on the customer perception of the security and online data and their trust in e-commerce applications. Furthermore, customers having a positive perception of the security of online data and trust in e-commerce applications are less prone to hacking attacks. On the other hand, the p-value for H3 was 0.999, for H4 was 0.829, for H6 was 0.389, and H7 was 0.113, all of which are greater than 0.05. These higher p-values highlight that there was no significant evidence that computer proficiency level and reading privacy policy statements of e-commerce websites and reviewing security certificates of websites before engaging in transactions have any impact on the customers’ perception of the security of online data and trust in e-commerce applications. Furthermore, there was also no evidence found that customers who read privacy policy statements of e-commerce websites and review security certificates are less prone to hacking attacks. Similarly, in the case of H7, customers’ rating of the security aspects of e-commerce websites and customers’ perception of the security of online data and trust in e-commerce applications were insignificant. In the cases of H8, H9, H10, H11, and H12, the p-values were 0, which is less than 0.05; thus, the relationship was significant, and these hypotheses were approved. Therefore, customers’ credit card usage concerns, e-commerce usage concerns, shopping motivational factors by business organizations, customer trustworthiness, and user feelings about the reputation of e-commerce impact the customers’ perception of the security of online data and trust in e-commerce applications.

5. Discussion

Customer trust and positive security perception are vital for e-commerce growth in Pakistan. Our case provides emphasis on improving security aspects in e-commerce infrastructures in Pakistan. There is a need for a collective effort from all stakeholders to enhance customers’ trust in e-commerce. The empirical data show that there is a need to improve customers’ trust and security perception in e-commerce for Pakistani customers. Cybercrimes are considered a major factor affecting customer confidence [95]; hence, the usage of advanced digital forensic techniques and tools can help in restricting e-commerce fraud and substantially increase customer confidence [96,97]. In the case of gender, our findings support the results of Alotaibi and Alshehri [50] that gender has an impact on information security behavior, whereas our findings are contrary to the findings of McCormac et al., who found no impact of gender on information security behavior [51]. Furthermore, our results are contrary to Huang et al. [44], as we did not see any correlation between computer proficiency and customer perception. In the e-commerce context, the user and e-commerce vendor/banking infrastructure [98] are vulnerable to cybercrimes; thus, we propose to use a three-layer adoption of digital forensic tools to curb e-commerce-related crimes. Customers are vulnerable to cyberattacks; hence, periodic self-digital forensics of customers’ computers and mobile devices can help to identify the presence of any malware and enhance personal security [99]. On the other hand, the adoption of digital forensic tools by e-commerce vendors and their banking partners can help them to identify, respond to, and investigate cybercrimes at the organizational level [100,101]. On the third level of government, law enforcement agencies can use digital forensic tools to investigate and respond to e-commerce frauds, and deal with criminals at the governmental level [102].

5.1. Theoretical Implications

Since cybersecurity has become very critical for the continuity of technological infrastructures, human factors are considered as issues [103]. In order to understand the security behavior of end-users, several information system theories have been employed by researchers. In the context of e-commerce, understanding user behavior is also very critical to foster a secure shopping experience. Customer motivation and security behavior are very critical for the growth of e-commerce, and our study highlights the perception of security implications of e-commerce portals in Pakistan. The findings from the study will help to perceive customers’ behavior aligned with PMT constructs [23,24]; as future work, the behavior of users can be studied in line with the theory of planned behavior [104]. Such an enhanced understanding of customer behavior can help in building effective controls for e-commerce websites.

5.2. Practical Implications

On the basis of our findings, we propose a model for improving customer trust and privacy in E-commerce in Pakistan. As shown in Figure 4, we identified four important actors: customers, vendors, technology, and the government. Different actions were identified which need to be carried out by each of the actors to enhance trust and privacy of end-users. Although product quality and price competitiveness are direct contributors to customers’ trust, it has been found that customers are sometimes willing to compromise their security if they find the quality and price reasonable [17]. Therefore, we recommend that vendors should provide quality products along with a competitive price as an indirect contributor to enhancing the trust of customers. Market analysis modeling tools can help in setting a price aligned with the pricing strategy of vendors [105,106]. Furthermore, if vendors have partnerships with known brands and advertise them on their website, then it also serves as a trust booster for customers [107]. Another important aspect that vendors need to pay more attention to is developing and displaying a use of information policy for all the information collected by them. If customers are aware of which information about themselves is being collected and how it will be used, it provides them with confidence about their security and privacy. Well-implemented opt-in and opt-out strategies by the e-commerce vendors show customers their sensitivity in preserving their customers’ privacy. Another factor that can enhance customer trust in e-commerce websites is setting up a well-defined dispute resolution policy. Many respondents highlighted their fears that, in cases they end up with credit card fraud [49], the e-commerce vendors will not help them [108].

Technological infrastructure is another important factor in fostering trust and security feeling among customers. The technological infrastructure deployed by the e-commerce vendors needs to adopt data encryption algorithms and advanced technologies such as blockchain [16] to secure communication [109]. To ensure that legitimate users can use their accounts, multifactor authentication mechanisms can be helpful; hence, the e-commerce business should implement multifactor authentication for user account management. Security and privacy badges provide users with a sense of satisfaction that the e-commerce website having such badges satisfies minimum security criteria [35]. E-commerce technical infrastructure needs to be fraud-resistant so that fraudulent transactions can be proactively responded to, and the payment mechanisms are fully secure [110]. Usability is another important contributor to enhancing customer interests [46,48,79]; a well-designed e-commerce platform provides customers confidence that the website is trustworthy. Features such as enhanced product displays, detailed information, and search features provide customers with a sense of security toward the e-commerce platform. A user-centric design approach to e-commerce portals can help in improving customer usage [111,112,113]. Adoption of advanced forensic tools can help in real-time monitoring of the network traffic, and in extracting and preserving digital evidence against e-commerce crimes.

Customers are another important actor; to increase customer confidence, there are some activities required on part of the customers. Customers need to prescreen the e-commerce platforms before purchasing by reading reviews on different horizontal/vertical search engines and directories [114]. An enhanced presence on different intermediaries and media platforms gives users confidence about the authenticity and credibility of that e-commerce platform. Secondly, customers need to ensure that their devices, the vicinities where they connect to e-commerce websites from, and their network connection need to be secure. Empirical data showed that a large majority of customers do not read the privacy policies and security certificates of e-commerce platforms. To make their shopping experience more secure, customers should regularly read and review privacy policies and security certificates, and this practice can minimize their chances of security violations.

In the context of Pakistan, the government also needs to carry out some actions to enhance the customers’ trust in e-commerce [14,15,115]. The empirical data provide evidence on customers’ reservations thar, in cases of financial fraud and misuse of their data, they expect little support from banks and other relevant authorities [116,117]. Therefore, the government needs to develop cybersecurity regulations to reassure the customers that banks and other institutions will support them in cases of banking or other legal violations during their shopping experience.

At the customer level, the personal security of the technological infrastructure is critical, and digital forensic tools can help in self-diagnostics of the devices used for engaging in e-commerce activities, and computer, network, email, and malware forensic tools can be beneficial. At the e-commerce vendor level, the focus should be on organizational security to protect customer and company data. Here, computer, network, analysis, malware, email, malware, and memory tools can be critical to secure the infrastructure from unauthorized access. At the governmental level, all types of forensic tools can be used to investigate and preserve digital evidence for strict actions against e-commerce crimes.

6. Conclusions

E-commerce has affected the buying behavior of customers; however, in developing countries such as Pakistan, there is still a huge potential for its growth. Recent COVID-19-induced lockdowns and closures have provided a significant push for digital transformation in business organizations; however, customers’ perception of trust and security of these websites is a vital factor for engaging in electronic business transactions. In this paper, we documented the results of an empirical study, suggesting that the age and gender of customers have an impact on the customers’ perception of the security of online data and trust in e-commerce applications. Furthermore, customers who have a positive perception of the security of online data and trust in e-commerce applications are less prone to hacking attacks. They are very critical regarding privacy policies and security; thus, they pay closer attention to privacy policies and are less likely to be a victim of hacking attacks. Furthermore, we developed a model to further improve the trust and security of customers engaging in e-commerce in Pakistan.

Limitations

Our study was based on a smaller sample size which lacked gender and geographical diversity, as our sample was male-dominated, and geographical diversity was not taken into account across Pakistan. Furthermore, most respondents were advanced computer users; hence, they may have more insights into technical aspects and may be more critical of security aspects. Due to these weaknesses, the results may not be reflective of the total population; accordingly, in future studies, we intend to test the effectiveness of our model in a more comprehensive population. Since our study relied upon quantitative data, there were no detailed insights into understanding why the respondents have a certain viewpoint.