Virtual Bank Usage Intention in Hong Kong: Exploring System Quality, UI Design, Security, Service Quality, Utilitarian Expectations, Word-of-Mouth, Brand Image and Rewards

The rapid development of fintech has led to the emergence of virtual banks (also known as neobanks, challenger banks, digital banks, or online-only banks). These new banks, by eliminating physical branches, reducing costs, and offering convenient and innovative user experiences, are transforming traditional banking. In Hong Kong, nearly four years since its launch, virtual banks have amassed 2.2 million users but face intense competition and multiple challenges while also presenting significant growth opportunities. Analyzing target customers’ usage intentions and related factors is crucial for enhancing service quality, market strategies, customer relationships, and innovation. This study explores the impact of system quality, user interface, security, service quality, utilitarian expectations, word of mouth, brand image, and reward systems on user intentions. Using a triangulation method that combines literature review, focus groups, pilot surveys, and large-scale surveys, 259 valid questionnaires were collected via snowball sampling and analyzed using SPSS. The results show significant correlations for all eight hypothesized factors. The research aims to provide insights and recommendations for the industry to improve overall service quality and increase market share, benefiting all stakeholders.

Introduction

The development of fintech has been a series of gradual transformations (Shantiet al., 2024), and it has now entered a new stage (Sardar & Anjaria, 2023). As society increasingly shifts towards cashless transactions, digital banking continues to expand (Lindström & Nilsson, 2023). According to a survey by the American Bankers Association, most consumers (71%) prefer managing their accounts via mobile apps or computers (ABA, 2023). This trend accompanies the decline of traditional banking; in 2023, over 2,500 branches in the United States closed (Bankrate, 2024).

Virtual banks (also known as neobanks, challenger banks, digital banks, or online-only banks) epitomize modern fintech innovation, representing a crucial step towards the full digitalization of traditional banking. They offer comprehensive online banking services without the need for physical branches, thereby enhancing convenience and efficiency (HKMA, 2018; Statista, 2024a). Terminologies for virtual banks vary globally: in mainland China, they are often called “direct banks,” in Hong Kong, “virtual banks,” in Taiwan, “Internet-only banks,” while in South Korea and Singapore, they are commonly referred to as “digital banks.” In India and Western countries, terms like “neobank,” “challenger bank,” “direct bank,” “digital-only bank,” or “Internet-only bank” are prevalent (forbes.com, 2024).

Globally, the total value of virtual banks reached US$18.6 billion in 2018, with 26 million users and 50 million accounts (Business Insider, 2019). By 2022, the number of neobanks/virtual banks worldwide had grown to 397 (Simon-Kucher, 2022). It is projected that by 2024, these banks will handle transactions worth US$6.37 trillion, with a user penetration rate of 3.89%, rising to 4.82% by 2028 (Statista, 2024b). In terms of revenue, virtual banks generated approximately US$96 billion in 2023, with expectations to exceed US$2 trillion by 2030 (cnbc.com, 2023).

The Hong Kong Monetary Authority (HKMA) has been promoting financial inclusion since 2017 and initiated the development of virtual banks. In 2018, it received 33 applications and issued eight licenses in 2019 (HKMA, 2019). All eight virtual banks launched operations by 2020. By the end of 2023, the number of virtual bank customers in Hong Kong reached 2.2 million (HK01, 2024). However, the market is highly competitive, with significant user overlap and over 50% dormant accounts. The average customer deposit is 30 times less than that of traditional banks and continues to decline. Total deposits only account for 0.2% (HK$32.2 billion) of Hong Kong’s HK$15.4 trillion in customer deposits. All eight virtual banks are still facing losses, with a combined mid-2023 loss of approximately HK$1.43 billion, individual losses ranging from HK$62 million to HK$318 million, and some investors hesitant to provide further funding (fintechnews.hk, 2024a, 2024b). Nonetheless, the 0.2% market share also indicates significant growth potential. Analyzing target customers’ usage intentions and related factors is crucial for improving virtual banks’ service quality, market strategies, customer relationships, and innovation.

Literature Review

Virtual banks are a significant fintech innovation, marking a key step in the digital transformation of traditional banking. Customers and the public are key stakeholders. This study explores factors influencing Hong Kong citizens’ intentions to use virtual banks, such as system quality, user interface design, security assurance, service quality, utilitarian expectations, word-of-mouth, brand image, and reward systems. The study integrates existing literature to propose and validate hypotheses.

System Quality

System quality is crucial for evaluating information systems’ performance, directly impacting users’ acceptance and continued use (Ivanova & Noh, 2022). For virtual banks, system quality affects customers’ transaction security, operational convenience, and information processing effectiveness (Nagyet al., 2024).

User Interface Design

User interface design is pivotal for virtual banks, evolving from visual design to comprehensive user interaction and experience design (Ivanova & Noh, 2022). Good interface design enhances usage intentions and builds user trust. It is especially critical in fintech’s rapidly developing virtual banking sector (Ravelino & Susetyo, 2023).

Security Assurance

Security is fundamental in virtual banking due to increasing user concerns about financial transaction safety (Ivanova & Noh, 2022). Security assurance significantly impacts users’ willingness to use virtual banks, encompassing technical aspects and psychological trust (Ashfaqet al., 2020).

Service Quality

Service quality is a key indicator of a company’s competitiveness, directly affecting user satisfaction and loyalty in virtual banks. It influences users’ intention to use and continued usage behavior (Ivanova & Noh, 2022; Dharmawanet al., 2023).

Utilitarian Expectation

Utilitarian expectation is crucial in technology acceptance and evaluating expected practical benefits. Understanding users’ utilitarian expectations is essential for grasping their usage intentions in virtual banks (Nagyet al., 2024; Kamdjouget al., 2021).

Word of Mouth

Word of mouth is a powerful influence on consumer decision-making, especially for virtual banks lacking physical touchpoints. With social media’s rise, word of mouth has expanded to digital platforms, impacting user growth and brand image (Yip & Mo, 2020; Ansary & Nik Hashim, 2018).

Brand Image

Brand image, reflecting consumers’ perceptions and emotions toward a brand, significantly influences purchase decisions (Islamiet al., 2023). In virtual banking, a strong brand image is essential for attracting and retaining customers and enhancing loyalty and usage intentions (Ansary & Nik Hashim, 2018; Hsu, 2023).

Rewards

Reward systems promote virtual banking services and build user loyalty. Effective strategies like cashback points rewards, and interest rate discounts encourage adoption and continued use, significantly enhancing customer satisfaction (Matousek & Xiang, 2021).

Intention to Use

Intention to use indicates the likelihood of future consumer use of a product or service. In virtual banking, it reflects potential customers’ acceptance and willingness to adopt services, influenced by system quality, interface design, security, and other factors (Ansary & Nik Hashim, 2018; Xuet al., 2019; Rakibet al., 2022).

Hypothesized Relationships of Factors

System Quality and Service Quality

In financial technology, system quality critically influences virtual banking service quality, encompassing dimensions like reliability, response time, availability, adaptability, and ease of use. User evaluations of their virtual banking experience are closely tied to system quality (Ivanova & Noh, 2022). Studies indicate that system quality affects perceptions of service quality and indirectly influences user satisfaction and loyalty (Norman, 2013).

H1: System quality positively impacts the service quality of virtual banks in Hong Kong.

User Interface Design and Service Quality

User interface design is crucial in fintech, promoting service use and enhancing user experience. Intuitiveness, aesthetics, and personalization of interface design directly affect perceptions of service quality (Ivanova & Noh, 2022). Usability and aesthetics are core elements, ensuring task completion and enriching emotional experiences, respectively, thereby influencing satisfaction and loyalty (Kimet al., 2009).

H2: User interface design positively impacts the service quality of virtual banks in Hong Kong.

Security Assurance and Service Quality

Security is a top concern in virtual banking, referring to the protection of personal information and funds from unauthorized access, leaks, or fraud. Security significantly influences trust and satisfaction in virtual banking (Venkateshet al., 2003) and is a prerequisite for continued service use (Ashfaqet al., 2020).

H3: Security assurance positively impacts the service quality of virtual banks in Hong Kong.

Service Quality and Usage Intention

Service quality significantly impacts consumers’ usage intentions in virtual banking, affecting satisfaction and loyalty. Dimensions like reliability, responsiveness, assurance, empathy, and tangibility directly influence whether consumers choose to use virtual banking services (Ivanova & Noh, 2022; Norman, 2013).

H4: Service quality positively impacts the usage intention of virtual banks in Hong Kong.

Utilitarian Expectations and Usage Intention

Utilitarian expectations refer to anticipated benefits like time savings, cost reduction, or increased efficiency. In virtual banking, these might include transaction convenience and fee transparency. Utilitarian expectations significantly influence usage intentions in fintech applications (Kamdjouget al., 2021; Alenizi, 2023).

H5: Utilitarian expectations positively impact the usage intention of virtual banks in Hong Kong.

Word-of-Mouth and Usage Intention

Word-of-mouth, involving informal consumer communication, can increase trust in virtual bank services. Electronic word-of-mouth (eWOM) spreads quickly through social media and review sites, significantly influencing consumer behavior (Yip & Mo, 2020; Ansary & Nik Hashim, 2018).

H6: Word-of-mouth positively impacts the usage intention of virtual banks in Hong Kong.

Brand Image and Usage Intention

Brand image encompasses consumers’ overall perception and cognition of a brand. In virtual banking, a strong brand image is crucial for attracting and retaining customers (Islamiet al., 2023). Positive brand image influences trust and usage intentions (Ansary & Nik Hashim, 2018).

H7: Brand image positively impacts the usage intention of virtual banks in Hong Kong.

Reward and Usage Intention

Reward systems, like high deposit interest rates and attractive account opening offers, significantly enhance usage intentions by increasing perceived value and satisfaction (Hsu, 2023; HKAB, 2023).

H8: Reward systems positively impact the usage intention of virtual banks in Hong Kong.

After reviewing all the literature, the proposed conceptual research model has been set as Fig. 1.

Fig. 1. Proposed conceptual research model.

Research Methods

This study employs a triangulation (Alenizi, 2023) approach, combining quantitative and qualitative data analysis. It integrates literature review, focus groups, individual interviews, a small pre-test survey, and a large-scale survey to gather both primary and secondary qualitative and quantitative data. Statistical Packages for Social Science (SPSS) version 29 software was used for the analysis of data.

Data Collection

This study’s data collection was conducted in two phases. Initially, primary data was gathered through qualitative research, which included a focus group of six participants and individual interviews with 8 participants. Insights from these discussions were used to design a preliminary questionnaire, which was administered through face-to-face interviews. Following adjustments based on feedback, a large-scale online survey was conducted using the snowball sampling method (Parkeret al., 2019), marking the quantitative research phase. The second phase involved secondary data collection through a comprehensive review of academic literature, journals, and other relevant materials to support the research.

Questionnaire Design

The questionnaire utilized a five-point Likert scale (Likert, 1932) ranging from “strongly disagree” to “strongly agree,” with demographic questions positioned at the end. A pilot test involving 43 participants led to revisions, reducing the number of questions from 44 to 37 to enhance the quality of responses (Table I).

Table I. Questionnaire Design

Variable	Factor	Keywords	References
Independent	System Quality (SysQua)	(1) Compatible, stable, smooth	Ivanova and Noh (2022)
		(2) Fast app response
		(3) Timely, effective support
Independent	User Interface (UseInt)	(1) Easy app use	Ivanova and Noh (2022)
		(2) Beautiful app design
		(3) Easy menu navigation
Independent	Security Assurance (SecAss)	(1) High security	Ivanova and Noh (2022)
		(2) Transparent transactions
		(3) Safer cybersecurity	Ashfaq et al . (2020)
Independent	Service Quality (SerQua)	(1) Meets daily needs	Ivanova and Noh (2022)
		(2) Satisfied with inquiries
		(3) Satisfied with services
		(4) Overall satisfaction
Independent	Utilitarian Expectation (UtiExp)	(1) Financial control	Kamdjoug et al . (2021)
		(2) Effective financial mgmt.
		(3) Good account management
		(4) Easier than others
		(5) No geographic limits
Independent	Word of Mouth (WorMou)	(1) Seek advice first	Ansary and Nik Hashim (2018)
		(2) Likely accept advice
		(3) Likely choose recommendations
Independent	Brand Image (BraIma)	(1) Better features	Ansary and Nik Hashim (2018)
		(2) Top industry brand
		(3) Stable, reputable
		(4) Logo as factor	Reyes et al . (2018)
Independent	Rewards (Rew)	(1) New account rewards	Hsu (2023)
		(2) Tier rewards
		(3) Encourages tier rewards
Dependent	Intention to Use (Int)	(1) Willing to use	Xu et al . (2019)
		(2) Continue using
		(3) Recommend to others
		(4) Plan to use	Rakib et al . (2022)

Sampling Objects

The study targeted Hong Kong residents aged 18 and above. According to the Hong Kong Census and Statistics Department, the target population at the end of 2023 was N = 6,565,000 (HKCSD, 2024). To achieve a 95% confidence level with a 5% margin of error, considering a confidence interval of 1.96, a success probability of 0.79, and a failure probability of 0.21, the minimum required sample size was calculated to be 255. The formula for the minimum required sample size:

$$\begin{array}{r}n=\frac{N{\left(\frac{z}{e}\right)}^{2}pq}{N-1+{\left(\frac{z}{e}\right)}^{2}pq}\end{array}$$

Meaning of each parameter in the equation:

N – 6,565,000 target population

z – 1.96 at 95% confidence level

e – 5% margin of tolerable error level

p – successful rate from pilot test-Int Q36 = 34/43

q – failure rate from pilot test-Int Q36 = 9/43

Results

The analysis of both qualitative and quantitative data provides a comprehensive understanding of the research topic. Qualitative data sheds light on the underlying reasons and development of new phenomena, offering a deeper insight into the subject matter (Hausken-Sutteret al., 2023). In contrast, quantitative data empirically supports the research hypotheses, quantifying findings with statistical evidence. This combined methodology enables a multi-faceted examination of the data, enhancing the interpretation of the survey results (Schoonenboom, 2023).

Descriptive Analysis

The descriptive analysis in this study is divided into two parts: an analysis of qualitative data from the focus group and individual interviews and an analysis of quantitative data from the questionnaire.

Focus Group and Individual Interviews

To explore the factors influencing the intention to use virtual banks, a focus group with 6 participants and individual interviews with 8 participants were conducted. The responses revealed that system quality and interface design are crucial factors. Participants emphasized the importance of a user-friendly interface and an efficient, secure transaction environment. They particularly valued the swift resolution of any system issues.

Security assurance concerns emerged prominently among respondents. Many expressed anxieties over the lack of physical branches in virtual banks, perceiving this as a potential risk to their funds’ safety. Some even feared the possibility of bank insolvency. This misunderstanding and unease about the term “virtual” indicates a significant need to enhance public trust in virtual banks.

Regarding service quality, users expect easy access to customer support and faster services compared to traditional banks. Utilitarian expectations also played a significant role; many users were attracted to virtual banks by account opening rewards, high-interest deposits, and cashback offers. Brand image and word of mouth also impacted usage intentions. Some participants felt that the absence of physical branches made virtual banks seem less stable than traditional banks, contributing to unease and misunderstandings about the “virtual” aspect.

Overall, while virtual banks have advantages in system quality and interface design, issues of security assurance and stability remain major barriers to their widespread acceptance.

Questionnaire Survey

A total of 43 interview questionnaires and 216 online questionnaires were collected, resulting in a final valid sample of 259 individuals, comprising 146 males and 113 females. To ensure the credibility and diversity of the questionnaire results, there were no restrictions on gender, age, income, education level, or prior experience with virtual banks.

Average Analysis of Quantitative Data

The questionnaire included nine factors, each comprising 3 to 5 questions, totaling 32 questions. Responses were rated on a 5-point Likert scale (Likert, 1932), with 1 indicating “strongly disagree” and 5 indicating “strongly agree.” The item with the highest average score best represented respondents’ intentions, while the item with the lowest score least represented them. Among the nine factors, utilitarian expectation had the highest average score (4.01), and intention to use had the lowest average score (3.58) (Table II).

Table II. Likert 5-Point Scale Min, Max, Mean Distribution for each Factor

Factor	Question	Minimum	Maximum	Mean	Standard deviation
SysQua	(2) Fast app response	1	5	3.92	0.847
	(3) Timely, effective support	1	5	3.59	0.920
UseInt	(1) Easy app use	1	5	3.91	0.883
	(2) Beautiful app design	1	5	3.82	0.868
	(3) Easy menu navigation	1	5	3.82	0.876
SecAss	(3) Safer cybersecurity	1	5	3.86	0.917
	(1) High security	1	5	3.72	0.944
SerQua	(1) Meets daily needs	1	5	3.85	0.915
	(2) Satisfied with inquiries	1	5	3.59	0.887
UtiExp	(5) No geographic limits	1	5	4.01	0.824
	(3) Good account management	1	5	3.70	0.932
WorMou	(2) Likely accept advice	1	5	3.71	0.939
	(1) Seek advice first	1	5	3.63	0.957
BraIma	(4) Logo as factor	1	5	3.90	0.852
	(1) Better features	1	5	3.73	0.820
Rew	(1) New account rewards	1	5	3.80	0.944
	(3) Encourages tier rewards	1	5	3.67	0.922
Int	(1) Willing to use	1	5	3.94	0.836
	(3) Recommend to others	1	5	3.58	0.951

Reliability Analysis

Cronbach’s alpha value of 0.7 is considered acceptable reliability (Tavakol & Dennick, 2011). In this study, the alpha coefficients were 0.974 for face-to-face surveys and 0.972 for online surveys, both exceeding 0.7. All variables had values between 0.804 and 0.921, indicating very high internal reliability (Table III).

Table III. Reliability Detection Coefficient Values

Variables		Cronbach’s alpha
System Quality (SysQua)		0.856
User Interface Design (UseInt)		0.804
Security Assurance (SecAss)		0.873
Service Quality (SerQua)		0.893
Utilitarian Expectation (UtiExp)		0.921
Word of Mouth (WorMou)		0.86
Brand Image (BraIma)		0.902
Rewards (Rew)		0.919
Intention to Use (Int)		0.872
Data collection method	Collection quantity	Cronbach’s alpha
Face-to-face interview method	43	0.974
Online collection	216	0.972
Total	259	0.972

Factor Analysis

This study will use factor analysis to test the relationship between variables by KMO test, Bartlett sphericity test, eigenvalue, and rotation component matrix.

KMO (Kaiser-Meyer-Olkin) and Bartlett Spherical Test

Through analysis, the KMO value of this study is 0.957, indicating that the data is highly suitable for factor analysis. Additionally, the statistical value for Bartlett’s test (Bartlett, 1937). of sphericity was 7688.462 with 496 degrees of freedom and a p-value less than 0.001, well below the common significance threshold (p < 0.05). This indicates that the correlations between items are significant, making the data very appropriate for factor analysis (Table IV).

Table IV. KMO and Bartlett Test

Kaiser-Meyer-Olkin measure of sampling adequacy		0.957
Bartlett’s test of sphericity	Approximate chi-square test	7688.462
	Degrees of freedom	496
	Significance	<0.001

Eigenvalue or Intrinsic Value (Eigenvalue)

Through eigenvalue analysis, it was determined that the nine research factors explained 72.649% of the variance. These nine factors were divided into four components. Component 1 included intention to use, service quality, and utilitarian expectation, with a total eigenvalue of 17.349 and a variance percentage of 54.216%. Component 2 comprised interface design, system quality, and security assurance, with a total eigenvalue of 2.758 and a variance percentage of 8.618%. Component 3 included word of mouth and a reward system, with a total eigenvalue of 1.226 and a variance percentage of 3.832%. Finally, Component 4 was brand image, with a total eigenvalue of 1.093 and a variance percentage of 3.416%, bringing the cumulative percentage to 70.082% (Table V and Fig. 2).

Table V. Analysis Results of Eigenvalues

Component	Factor	Total	Initial eigenvalues % of variance	Cumulative % of variance
1	(1) Willing to use	17.349	54.216	54.216
2	(2) Continue using	2.758	8.618	62.834
3	(3) Recommend to others	1.226	3.832	66.666
4	(4) Plan to use	1.093	3.416	70.082

Fig. 2. Analysis results of eigenvalues, scree plot.

The factor loadings for intention to use were 0.836, 0.835, and 0.793, reaching a good level, consistent with the studies by Rakibet al. (2022) and Xuet al. (2019). Service quality had acceptable loadings of 0.614 and 0.559, aligning with Ivanova and Noh (2022). Utilitarian expectation showed acceptable loadings of 0.686, 0.621, 0.577, 0.558, and 0.552, similar to Kamdjouget al. (2021). Security assurance had acceptable loadings of 0.585, 0.575, and 0.569, also consistent with Ivanova and Noh (2022). System quality had acceptable loadings of 0.700, 0.667, and 0.643, again aligning with Ivanova and Noh (2022). Interface design showed acceptable loadings of 0.738, 0.630, and 0.581, consistent with Ivanova and Noh (2022). The reward system had good loadings of 0.845, 0.834, and 0.818, in line with Hsu (2023). Word of mouth showed acceptable loadings of 0.663, 0.661, and 0.599, consistent with Ansary and Nik Hashim (2018). Finally, brand image had acceptable loadings of 0.661, 0.660, 0.637, and 0.564, aligning with the studies by Ansary and Nik Hashim (2018) and Reyeset al. (2018) (Table VI).

Table VI. Factor Analysis–Factor Rotation Matrix$\text{a}$

C	Factor	Keywords	FL		FL (Ref)
1	Intention to Use (Int)	(4) Plan to use	0.836	0.769	Rakib et al . (2022)
		(2) Continue using	0.835	0.951
		(1) Willing to use	0.793	0.925	Xu et al . (2019)
	Service Quality (SerQua)	(4) Overall satisfaction	0.614	0.867	Ivanova and Noh (2022)
		(1) Meets daily needs	0.559	0.81
	Utilitarian Expectation (UtiExp)	(4) Easier than others	0.686	0.673	Kamdjoug et al . (2021)
		(5) No geographic limits	0.621	0.788
		(2) Effective financial mgmt.	0.577	0.792
		(3) Good account mgmt.	0.558	0.834
		(1) Financial control	0.552	0.832
2	Security Assurance (SecAss)	(2) Transparent transactions	0.585	0.926	Nagy et al . (2024)
		(1) High security	0.575	0.916
		(3) Safer cybersecurity	0.569	<0.7
	System Quality (SysQua)	(2) Fast app response	0.700	0.83	Nagy et al . (2024)
		(1) Compatible, stable, smooth	0.667	0.85
		(3) Timely, effective support	0.643	0.814
	User Interface Design (UseInt)	(1) Easy app use	0.738	0.84	Nagy et al . (2024)
		(3) Easy menu navigation	0.630	0.896
		(2) Beautiful app design	0.581	0.879
3	Rewards (Rew)	(2) Tier rewards	0.845	0.82	Hsu (2023)
		(1) New account rewards	0.834	0.85
		(3) Encourages tier rewards	0.818	0.74
	Word of Mouth (WorMou)	(2) Likely accept advice	0.663	0.72	Ansary and Nik Hashim (2018)
		(1) Seek advice first	0.661	0.74
		(3) Likely choose recommendations	0.599	0.71
4	Brand Image (BraIma)	(3) Stable, reputable	0.661	0.74	Ansary and Nik Hashim (2018)
		(1) Better features	0.660	0.69
		(2) Top industry brand	0.637	0.73
		(4) Logo as factor	0.564	0.7	Reyes et al . (2018)

The formula for the varimax criterion (Kaiser, 1958):

$$\begin{array}{r}v=\sum _s\{n\sum _j\left(\frac{a_{js}^2}{h_j^2}\right)2^2-{\left[\sum _i\left(\frac{a_{is}^2}{h_i^2}\right)\right]}^2\}/n^2\end{array}$$

Meaning of each parameter in the equation:

v – the varimax criterion value is maximized to achieve simpler, more interpretable factors in factor analysis.

∑s – summation over all factors. This indicates that the calculation is done for each factor.

$n$ – the number of variables in the dataset.

$a_{js}^2$ – the loading of the $j$-th variable on the $s$-th factor. It represents how much a variable contributes to a factor.

$h_i^2$ – the communality of the i^th variable is the proportion of variance accounted for by the common factors.

∑ j – summation over all variables for a specific factor calculates each variable’s contribution to that factor.

∑I – summation over all variables. This part is used to adjust the factor loadings to improve interpretability.

Linear Regression Analysis

Service Quality of Linear Regression Analysis

According to the data in Table VII, the R-squared value of model 3 is 0.785, which means that user interface design explains 78.5% of service quality. The R-value is 0.785, which is a moderate positive correlation. The larger the gap between the F-value and the P-value, the greater the confidence between the two factors. The F value is 136.683, and the p-value is < 0.001 (when p < 0.05, it means that the confidence reaches 95%), which proves that the perceived ease is 95% confidence in service quality. The significance of each item in the model is less than 0.05 (p < α = 0.05), indicating that user interface design has a significant impact on service quality.

Table VII. Linear Regression Analysis for Service Quality

Model summarye
Model	R	R square	Adjusted R square	Standard error of the estimate
3	0.785c	0.617**	0.612	0.570

Regression equation:

$$\begin{array}{rl}& ServiceQuality=0.181+0.296\\ & \times UserInterfaceDesign3\left(easymenunavigation\right)\\ & +0.331\times SecurityAssurance3\left(Safercybersecurity\right)\\ & +0.332\\ & \times SystemQuality1(Compatible,stable,smooth)\end{array}$$

From the data on the equation, each of the service quality (a total of three items) is important to usage intention. The score of ``compatible, stable, smooth'' is the highest and the most important (Table IX, Figs. 3 and 4).

Fig. 3. Histogram of residuals.

Fig. 4. Residuals of scatter plot.

Linear Regression Analysis of Usage Intention

According to the data in Table VIII, the R-squared value of model 4 is 0.532, which means that word of mouth explains 53.2% of usage intention. The R-value is 0.730, which is a moderate positive correlation. The larger the gap between the F-value and the P-value, the greater the confidence between the two factors. The F-value is 72.276, and the p-value is <0.001 (when p < 0.05, it means that the confidence reaches 95%), which proves that the perceived ease is 95% confidence in usage intention. The significance of each item in the model is less than 0.05 (p < α = 0.05), indicating that word of mouth has a significant impact on usage intention. Regression equation:

Table VIII. Anova$\text{a}$ Table for Service Quality

Model		SS	DF	MS	F	Sig.
3	Regression	133.069	3	44.356	136.683	<0.001d
	Residual	82.753	255	0.325
	Total	215.822	258

Table IX. Coefficients$\text{a}$ Table for Service Quality

Model		Unstandardized coefficients		Beta	t	Sig.	Collinearity statistics
		B	Standard error				To	VIF
3	(Constant)	0.181	0.185		0.979	0.329
	UseInt3	0.296	0.058	0.283**	5.092	<0.001	0.487	2.055
	SecAss3	0.331	0.051	0.332**	6.443	<0.001	0.565	1.769
	SysQua1	0.332	0.058	0.298**	5.698	<0.001	0.548	1.824

$$\begin{array}{rl}& UsageIntention=0.139+0.234\\ & \times WordofMouth2\left(likelyacceptadvice\right)\\ & +0.200\times UtilitarianExpectation4\left(easierthanothers\right)\\ & +0.284\times Rewards2\left(tierrewards\right)\\ & +0.213\times ServiceQuality2\left(satisfiedwithinquiries\right)\end{array}$$

From the data on the equation, each of the word of mouth (a total of four items) is important to usage intention. The scores for “tier rewards” and “likely accept advice” are the highest and the most important (Tables X-XII, Figs. 5 and 6).

Table X. Linear Regression Analysis for Usage Intention

Model summarye
Model	R	R square	Adjusted R square	Standard error of the estimate
4	0.730d	0.532**	0.525	0.655
d: Predictors: (Constant), WorMou2-21. Likely to accept advice, UtiExp4-17. Good account management, Rew2-31. Tier rewards, SerQua2-11. Satisfied with inquiries

Table XI. Anova$\text{a}$ Table for Usage Intention

Model		SS	DF	MS	F	Sig.
4	Regression	124.180	4	31.045	72.276	<0.001e
	Residual	109.102	254	0.430
	Total	233.282	258

Table XII. Coefficients$\text{a}$ Table for Usage Intention

Model		Unstandardized coefficients		Beta	t	Sig.	Collinearity statistics
		B	Standard error				To	VIF
4	(Constant)	0.139	0.208		0.667	0.506
	WorMou2	0.234	0.056	0.241**	4.158	<0.001	0.550	1.818
	UtiExp4	0.200	0.061	0.196**	3.279	0.001	0.513	1.950
	Rew2	0.284	0.058	0.273**	4.934	<0.001	0.599	1.668
	SerQua2	0.213	0.062	0.198**	3.407	<0.001	0.544	1.838

Fig. 5. Histogram of residuals.

Fig. 6. Residuals of scatter plot.

Analysis Results

After conducting statistical analysis using SPSS software, the study obtained the R square with value motivated explanation at a 0.01 significance level. The hypotheses testing results from H1 to H6 and H8 indicate a positive impact, accepting the alternative hypotheses (Table XIII and Fig. 7).

Table XIII. Hypothesis Testing Result

Code	Research hypothesis	Beta	Result
H1	System quality has a positive impact on service quality	0.298**	Supported
H2	User interface design has a positive impact on service quality	0.283**	Supported
H3	Security assurance has a positive impact on service quality	0.332**	Supported
H4	Service quality has a positive impact on usage intention	0.198**	Supported
H5	Utilitarian expectations have a positive impact on usage intention	0.196**	Supported
H6	Word-of-mouth has a positive impact on usage intention	0.241**	Supported
H7	Brand image has a positive impact on usage intention	–	Not Supported
H8	Reward systems have a positive impact on usage intention	0.273**	Supported

Fig. 7. Theoretical research model with analysis results. Note: ^**p < 0.01 correlation analysis/linear regression analysis.

Conclusions and Recommendations

Conclusions

This study employed a triangulation approach, blending qualitative and quantitative research methods, to explore factors influencing the intention to use virtual banks in Hong Kong. From literature reviews, focus groups, interviews, and surveys, 259 valid questionnaires were analyzed.

The findings indicate that users value system quality, interface design, security assurance, service quality, utilitarian expectations, word-of-mouth, and rewards when considering virtual banks. Reward systems are crucial for encouraging account openings and continued usage (Reyeset al., 2018).

Utilitarian factors, such as convenience and efficiency, are key determinants. Users seek seamless operation and fast data processing, reflecting the positive impact of system quality on usage intention (Ivanova & Noh, 2022). A user-friendly interface with features like fingerprint and facial recognition enhances user satisfaction.

Moreover, trust in virtual banks and the intention to use them heavily depend on reliable security assurance measures and the quality of services provided. Security assurance research encompasses both technical aspects and the psychological construction of user trust (Dharmawanet al., 2023). Qualitative findings indicate that some target customers are concerned about the term “virtual,” questioning the security assurance and stability of virtual banks.

Recommendations

Academic Recommendations

Future studies should explore the integration of emergent technologies such as AI, big data analytics, and blockchain within the virtual banking sector (Kalyani & Gupta, 2023). This could potentially elevate the efficiency of financial services, enhance customer experiences, and refine user engagement strategies, thereby enriching the fintech discourse and offering actionable insights for industry application.

Managerial and Industry Recommendations

1. Security Assurance and Trust: Security assurance and stability are critical for the success of virtual banks. Building customer trust is essential for broader market acceptance. Research shows that many people are wary of the term “virtual” due to concerns over security and stability. It is suggested that the term “virtual bank” be replaced with “Licensed Digital Bank,” “Licensed Online Bank,” or “Digital-only Bank.” Additionally, a comprehensive communication strategy should be employed to enhance users’ understanding and confidence in the bank’s security measures. This includes regular promotional campaigns, security education, and awareness activities. Initiatives such as easy-to-understand videos across various platforms, social media safety tips, expert-hosted TV and radio segments targeting older demographics (Bansal & Choudhary, 2023), and regular security assurance events and safety update reports are recommended.

2. User Interface (UI) Design: An excellent UI is crucial for virtual banks, as it is the primary platform for customer interaction. A well-designed interface enhances user experience and triggers positive “feel-good” effects (Choiet al., 2021). Companies like Microsoft, Apple and social media platforms like TikTok and Instagram have shown that competitive advantage often boils down to superior UI/UX design (Refokus, 2020). From a cost-effectiveness perspective, investing in UI/UX is more efficient than traditional branch setups and advertising. Virtual banks should allocate more budget to UI/UX design, collaborate with renowned design firms, and establish multidisciplinary teams to support design efforts. Continuous monitoring and improvement of user interaction and satisfaction are essential to maintain a competitive edge.

3. System Quality: Virtual banks should continuously enhance their system quality to ensure platform stability and efficiency, improving transaction speed and data processing capabilities. This will increase user trust, satisfaction, and loyalty. Systematic IT outsourcing (ITO) services can provide ongoing enhancements in professional technology, cost efficiency, risk management, scalability, and flexibility (Mo & Chiang, 2024).

4. Reward Systems: Incentive programs can effectively increase user activity and loyalty (HKAB, 2023). Account opening rewards and high-interest deposits attract new users, while high-interest term deposits boost deposit volumes. Loyalty programs offering points, cash or spending rebates, discounts, and personalized financial products can further enhance customer retention and satisfaction. These programs should be regularly updated to maintain their appeal and competitiveness.

5. Brand Building and Word-of-Mouth Marketing: Strengthening brand image can increase user recognition and trust in virtual banks. Effective WOM and eWOM marketing can elevate brand visibility and attract new users (Yip & Mo, 2020; Dharmawanet al., 2023). Virtual banks should encourage satisfied customers to share positive experiences and leverage social media and other online platforms to amplify their reach. Clear and appealing brand positioning, combined with effective online marketing strategies, can enhance brand visibility and attract potential users. Collaborating with other fintech innovators to co-develop new financial products and services can also expand business scope and enhance innovation capacity and market competitiveness.

Research Limitations and Future Directions

Despite the rigor of this study, limitations due to time and resource constraints were inevitable, providing avenues for future research. Future research should consider employing probability sampling techniques and expanding the sample size to enhance the generalizability of the findings. The theoretical framework in this study may not cover all potential factors influencing the intention to use virtual banks. Future research could explore additional factors that might impact this intention (Bangaet al., 2023) and examine how emerging technologies like blockchain and AI affect system quality and user experience in virtual banking.