Perceived ease of use is a factor of technology acceptance model (TAM) theory. Davis (1989:320) defined PEOU as ‘the degree to which a person believes that using a particular system would be free of effort’. The results of empirical studies have supported that PEOU plays an important role in technology adoption (Arora, Singh & Gupta 2022; Arumugam, Ismail & Joeharee 2020). This tells that metered taxi operators in South Africa are more likely to use technologies that they find easy to use.

Davis (1989:320) defined PU as ‘the degree to which a person believes that using a particular system would enhance his or her job performance’. From the perspective of Rogers (2010), the usefulness of a technology is affected by the shared view of potential users in a network or association. As metered taxi operators in South Africa are aligned to networks or associations (Ngubane et al. 2020), the usefulness of a transport management system is determined by the network or association to which the operators belong. This situation was explained by Venkatesh et al. (2003), who said that social influence affects how individuals perceive the usefulness of a technology.

The Unified Theory of Acceptance and Use of Technology 2 (UTAUT2) postulated that the price value of a technology determines the BI to use the technology by consumers (Venkatesh et al. 2003). Consumers will adopt a technology if they can afford the hardware, services and operational costs. Utami et al. (2021) found that transport entrepreneurs require capital to buy smartphones and broadband to operate e-hailing services. In the same vein, Boateng, Appau and Baako (2022) highlighted that it costs money for transport entrepreneurs to join an e-hailing application such as Uber. In South Africa, to operate an Uber car, the car model should be in good condition and fully insured (Ngubane et al. 2020). As most metered taxi operators fund their businesses, the costs of adopting e-hailing applications have a direct impact on BI to use the application.

The second level of Maslow’s hierarchy of needs is safety, which refers to security, stability and protection (Maslow & Lewis 1987). In business, safety is equated to minimising the probability of risk and uncertainty (Holton 2004). Risks associated with operating a transport business include damage to vehicles, inability to receive clients, and injuries to drivers and customers. Concerning customer and driver safety, studies carried out in Ghana revealed that customers of e-hailing perceived e-hailing taxis as safe (Boateng et al. 2022). This is because e-hailing cars are tracked in real time, the identification details of drivers and the customers are shared, cars are insured, and drivers undergo criminality screening (Arumugam et al. 2020). However, in the South African transport environment, e-hailing operators are prone to targeted violence by transport networks or associations fighting over the domination of profitable routes (Henama & Sifiso 2017). Because of such violence, e-hailing operators risk their vehicles being destroyed.

Trust is the perceived risk of willingness to accept that one part will do good for another part (eds. Marková & Gillespie 2007). A part can be a person, an institution or a system. Trust is an important factor in the adoption of online services (Arora et al. 2022). Trusting online services is a challenge because many people operate from the point of distrust. After all, when online, people are afraid of cybercrime, for example, credit card fraud, and account hacking (eds. Marková & Gillespie 2007). Mechanisms to improve online customer trust are known to include product advertisement, product quality, product testimonials (Li 2019), information accuracy, security and product usability (Fong et al. 2023). As an example, e-hailing customers trust that the service provider will provide them with a car, and transport operators trust that the quality of e-hailing apps is of good quality for customers to book a car (Fong et al. 2023).

The convenience of a service is also perceived as the experience that the consumer benefits from using a service, which can be positive or negative (Kumar, Sachan & Dutta 2020). Empirical studies on e-hailing revealed that convenience is a benefit of e-hailing (Arumugam et al. 2020; Boateng et al. 2022; Utami et al. 2021). Utami et al. (2021) found that e-hailing services are convenient for secure movement of money because all transactions are carried online.

Attitude is an evaluative response, which is an affect for or against an object (Ajzen 2005). Arul and Misra (2002) argued that attitudes are tendencies with the ability to promote or impede an action at individual or group level. There are factors that affect attitude towards an object, for example, PEOU and PU in the TAM (Davis 1989), and social pressure expectations and self-efficacy (Ajzen 2005). Furthermore, the past literature has reported on many other factors that determine attitude in technology adoption, for example, trust and convenience (Liu & Tai 2016). In the context of the adoption of e-hailing, there are empirical studies that found significant relationships between attitude and PEOU, and attitude and PU, for example, as by Shah et al. (2020).

Davis, Bagozzi and Warshaw (1989:984) referenced Fishbein and Ajzen (1975) to define BI as ‘a measure of the strength of one’s intention to perform a specified behaviour’. Behavioural intention has been measured through various determinants depending on the theoretical model; for example, in TAM, the determinants are PU and ATU (Davis et al. 1989), and in UTAUT, the determinants are performance expectancy, effort expectancy, social influence and facilitating condition (Venkatesh et al. 2003). There are empirical studies that adopted and extended adoption theories to test BI, for example, a study by Kumar et al. (2020).

The study participants were metered taxi entrepreneurs who were randomly sampled from Sandton, in the Johannesburg metropolitan area. Participant selection was based on availability and desire to participate in the study, and no compensation was paid. The questionnaire was self-administered from April to September 2020. In total, 250 questionnaires were completed; however, 227 had usable data and were analysed, which met the recommended threshold sample size of between 200 and 500 for quantitative analysis (Hair et al. 2019).

The construction of the questionnaire was based on the constructs of the research model (Figure 1). The questionnaire covered both the dependent and independent constructs of the model. Attitude towards use was measured by six independent variables, while BI was measured by seven independent variables. The construct TWTMS was measured with five items. The questionnaire measured the responses on an unlabelled five-point Likert scale, from 1 (strongly disagree) to 5 (strongly agree).

The data were cleaned; a statistical analysis was performed to measure central tendency and normality, and multilinear regression analysis was used for hypotheses testing. The SPSS statistical package (version 24) was used to analyse the data.

Ethical clearance to conduct this study was obtained from the University of South Africa College of Science, Engineering and Technology’s (CSET) Research and Ethics Committee (No. 062/NM/2018/CSET_SOC).

The internal consistency of the model was measured using the Cronbach’s alpha, with alpha coefficients ranging between 0.907 and 0.989 (see Table 2). Aligned with Pallant (2016), the Cronbach’s alpha values were above 0.70, which suggest that the measurement scales were internally consistent, and the collected data were reliable.

The provided analyses examine regression models across various dependent variables and predictor combinations (see Table 3–Table 11). For ATU, PC, PP and PT, they collectively contribute to explaining variance, with Model 3 being the most effective. Conversely, in predicting BI, models incorporating PC, PS, PT and PU demonstrate superior explanatory power, particularly Model 4. Meanwhile, the impact of TWTMS on BI is highlighted, with Model 1 showcasing the strongest association.

Six hypotheses (H1, H2, H3, H4, H5 and H6) were tested to assess the ATU of transport management systems by metered taxi entrepreneurs. Three of the hypotheses were significant (H3, H5 and H6) and are discussed.

The first significant hypothesis was H3, which found that the PP positively influences the ATU (β = 0.159; p < 0.05). These results suggest that metered taxi entrepreneurs will have a positive attitude towards transport management systems if they find them to be of economic value. There is a scarcity of studies that have tested this factor on the adoption of transport management systems by metered taxi entrepreneurs. However, a study that investigated customer loyalty to e-deals (Cheah, Phau & Liang 2015) found that perceived value positively predicted attitudes towards e-deals.

The second significant hypothesis was H5, which found that PC positively influences the ATU (β = 0.409; p < 0.05). These results suggest that the subjective feeling that transport management systems provide the convenience of managing metered taxis will positively affect the attitude of entrepreneurs towards using them. There is a lack of studies that have tested this factor on the adoption of transport management systems by metered taxi entrepreneurs; however, there are other studies that found a significant relationship between these factors, for example, in mobile learning (Chang, Yan & Tseng 2012) and mobile payment (Sari, Habibi & Hayuningputri 2022).

The third significant hypothesis was H6, which found that PT positively influences ATU (H6) (β = 0.481, p < 0.05). These results suggest that the trustworthiness of transport management systems will positively influence metered taxi entrepreneurs to believe that they are useful in managing their businesses. These results corroborate the findings of other studies that investigated the adoption of e-hailing applications and found that trust influences ATU (Arora et al. 2022; Razi, Tamrin & Nor 2019).

Seven hypotheses (H7, H8, H9, H10, H11, H12 and H13) were tested to evaluate BI to adopt transport management systems by metered taxi entrepreneurs. Four of the hypotheses were significant (H7, H10, H11 and H12).

The first significant hypothesis was H7, which found that PU positively influences BI (β = 0.209; p < 0.05). The results suggest that taxi entrepreneurs will be inclined to use transport management systems if they perceive the system as valuable or beneficial to their businesses. The results corroborate the findings of Arora et al. (2022) and those of Arumugam et al. (2020).

The second significant hypothesis was H10, which found that PS positively influences BI (β = 0.244; p < 0.05). The results suggest that metered taxi entrepreneurs will use transport management systems if they are guaranteed security, integrity and confidentiality. These results align with the findings of Arumugam et al. (2020), who found that drivers and passengers felt that they were secure when using e-hailing because cars are tracked, and drivers’ and passengers’ identification details are captured. The issue of physical security is of concern in South Africa because the metered taxi transport environment has been reported to be volatile (Henama & Sifolo 2017).

The third significant hypothesis was H11, which found that PC positively influences BI (β = 0.302; p < 0.05). These results suggest that BI to adopt will increase if transport management systems are perceived to improve business convenience. These results are consistent with the findings of prior studies on e-hailing (Arumugam et al. 2020; Utami et al. 2021). E-hailing was found to be convenient because it provides geolocation to pick up and drop clients (Utami et al. 2021) and provides an estimation of travel time to pick up and drop clients (Boateng et al. 2022).

The fourth significant hypothesis was H12, which found that PT positively influences BI (β = 0.238; p < 0.05). The results suggest that if metered taxi entrepreneurs trust that the transport management system is reliable, secure or efficient, adoption will increase. These findings are aligned with the finding of Mohamad et al. (2016), who found that taxi operators have a high likelihood of adopting Uber if they find it trustworthy.

H14 was significant and found that BI positively influences TWTMS (β = 0.313; p < 0.05). These results can be interpreted to mean that metered taxi entrepreneurs are willing to adopt and use transport management systems to reap the benefits. The results of this study satisfy the TAM model as proposed by Davis (1989), who postulated that a positive BI results in the AU of the technology. Furthermore, the results of this study align with those of Haba and Dastane (2018), who found a significant relationship between BI and AU of taxi hailing mobile applications in Malaysia.