UTAUT2 and other extensions

The original UTAUT framework was developed to explain and predict the acceptance of technology in an organisational context (Venkatesh et al., 2003), although, later it was tested in non-organisational settings too (Venkatesh, Thong & Xu, 2012; Venkatesh, Thong & Xu, 2016). Over the years, UTAUT showed wide application, which enhanced the generalisability of the theory (Venkatesh, Thong & Xu, 2012; Neufeld, Dong & Higgins, 2007). Given the variance of information communication technologies and the advances in the sector, a number of scholars extended UTAUT to adapt it to the context or improve its predictive power (Venkatesh, Thong & Xu, 2012).

The adaptations of the model were underpinned by four main approaches, reflecting a) the modification of the model to different contexts, b) the alterations of the endogenous variables, c) the addition of attitudinal antecedents, and d) the examination of various moderating variables. The first stream of research extended the model to apply it to new technologies (e.g. enterprise systems, e-health systems), focus on new user segments (e.g. healthcare professionals), and examine it in new geographical and cultural settings (e.g. India, China) (Chang et al., 2007; Yi et al., 2006; Gupta, Dasgupta & Gupta, 2008). For instance, the model was extended by a set of web-specific constructs, including trust and personal web innovativeness to explore how well it predicts the use of web tools (Casey & Wilson-Evered, 2012). Another stream of research extended UTAUT by incorporating additional endogenous variables (e.g. (Sun, Bhattacherjee & Ma, 2009)), such as satisfaction and continuous intention to use (Maillet, Mathieu & Sicotte, 2015). The third stream of research scrutinised additional determinants of use and behavioural intention, such as task-technology fit and personality traits (Zhou, Lu & Wang, 2010; Wang, 2005). Finally, some studies extended UTAUT by introducing new contextual and moderating variables, such as culture, ethnicity, religion, employment, language, income, education and geographical location, among others (Im, Hong & Kang, 2011; Al-Gahtani, Hubona & Wang, 2007; Riffai, Grant & Edgar, 2012).

Although the adaptations of the model enriched the understanding of the theory applications, the research was mainly limited to organisational settings (Chang et al., 2007; Yi et al., 2006; Gupta, Dasgupta & Gupta, 2008; Im, Hong & Kang, 2011; Al-Gahtani, Hubona & Wang, 2007). The literature lacked evidence about a user behavioural model, which could explain the utilisation of technology by consumers rather than employees. However, such evidence was important, given arguments in prior studies suggesting that the determinants of acceptance in organisational and non-organisational (i.e. consumer) settings are not the same. It was found that the importance of the factors reflecting the costs and benefits of behaviour varied based on the context (e.g. (Brown & Venkatesh, 2005; van der Heijden, 2004; Brown, Venkatesh & Bala, 2006; Brown & Venkatesh, 2005; Kim, Malhotra & Narasimhan, 2005)).

Given the above limitations, Venkatesh et al. proposed an extension of UTAUT, named UTAUT2 (Venkatesh, Thong & Xu, 2012). The model (Figure 2) set out to address two main objectives. First, compared to all prior attempts to extend the model, UTAUT2 was not designed to have a specific focus (e.g. new technology, geographical location). Instead, the goal of the theory was to represent an overarching framework for examining technology acceptance. The extension was designed to give a higher precision in explaining user behaviour (Venkatesh, Thong & Xu, 2012; Alvesson & Kärreman, 2007). The second objective was to propose a behavioural model of consumer technology acceptance, in contrast to UTAUT, which was developed to examine technology in organisational settings. To fulfil the objective, Venkatesh et al. planned to extend the UTAUT model with new constructs, tackling behavioural and attitudinal determinants of the utilisation of technology in the non-organisational context (Venkatesh, Thong & Xu, 2012). The authors developed UTAUT2 by introducing three new constructs and altering some relationships (e.g. removing the voluntariness) in the original model to adapt it to the consumer technology use context. Such an approach offered a new theoretically justified mechanism for predicting technology acceptance, which was encouraged and endorsed by prior research (Bagozzi, 2007; Venkatesh, Davis & Morris, 2007). In addition to advancing the technology acceptance literature (Venkatesh, Thong & Xu, 2012), UTAUT2 aimed to achieve wider generalisability by addressing the private user segment .

UTAUT 2 postulates that the use of technology by individuals is underpinned by the effect of the three additional constructs, namely, hedonic motive, cost/perceived value and habit, moderated by age, gender and experience. Hedonic motivation is defined "as the fun or pleasure derived from using technology, and it has been shown to play an important role in determining technology acceptance and use" (Venkatesh, Thong & Xu, 2012). The inclusion of this construct was justified by the findings of prior studies in the IS and marketing domains which found that the perceived hedonic nature of the outcome (e.g. perceived enjoyment) was a significant predictor of consumer technology use (Brown & Venkatesh, 2005; van der Heijden, 2004). The rationale for integrating cost in the new model was based on the relative importance of the factor in the context of consumer product use compared to the usage of technology in workplace settings. For example, when technology is used by employees in organisations, users do not feel responsible for the cost that is associated with the use of technology, due to the lack of direct financial implications for them (Venkatesh, Thong & Xu, 2012). In contrast, the use of consumer technology implies a higher perception of the responsibility, due to direct costs borne by the use of technology. The lower the costs, the more intensive is the use of technology (Venkatesh, Thong & Xu, 2012; Brown, Venkatesh & Bala, 2006; Brown & Venkatesh, 2005). Since UTAUT and UTAUT2 utilised subjective measures, the cost factor was represented by price value. Price value is defined as "consumers’ trade-off between the perceived benefits of the applications and the monetary cost for using them" (Venkatesh, Thong & Xu, 2012). A positive relationship between perceived value and intention to use indicates that a user perceives the benefits of technology use as higher and more important than the associated monetary costs. The third variable included in UTAUT2 is habit, which is defined as "the extent to which people tend to perform behaviours automatically" (Venkatesh, Thong & Xu, 2012). The construct was operationalised based on prior studies which had brought the automaticity perspective into the research. In contrast to a reason-oriented framework (e.g. TRA and TPB), which states that behavioural intention results from deliberate evaluations, the automaticity perspective considers technology use to be an automatic and unconscious behaviour (e.g. (Limayem, Hirt & Cheung, 2007; Kim, Malhotra & Narasimhan, 2005)). Habit was hypothesised to have a direct and indirect effect on actual use through behavioural intention (Venkatesh, Thong & Xu, 2012). However, the effect of either of the paths is dependent on the degree to which people rely on routinised behaviour in accepting/using technology (Venkatesh, Thong & Xu, 2012; Ajzen, 2011). The extended version of UTAUT resulted in a number of theoretical contributions. The model explains 74 %of the variance in behavioural intention and 52 % of the variance in technology use, which suggests that the model has high predictive validity when applied to the consumer segment. The supported effects of price value, hedonic motivation and habit indicate three significant drivers of consumers’ intention to use or actual use of technology (Venkatesh, Thong & Xu, 2012). Specifically, the introduction of the habit factor demonstrated the alternative theoretical mechanism in examining technology use (Bagozzi, 2007). Such an approach challenged the role of intention (Venkatesh, Davis & Morris, 2007), which was commonly used as a proxy for behaviour (e.g. (Venkatesh et al., 2003; Ajzen, 2011)). The inclusion of hedonic motivation in the model was found to be more important than performance expectancy and was significant across a wide range of studies (Alalwan, Dwivedi & Rana, 2017; Megadewandanu, Suyoto & Pranowo, 2016). In addition, the integration of price value in UTAUT2 addressed the need to measure the costs of IS use in the consumer context. Finally, extended UTAUT determines the role of personal factors (gender, age, and experience) in moderating the effect that hedonic motivation, price value and habit have on behavioural intention and/or use (Venkatesh, Thong & Xu, 2012).

Applications

UTAUT and UTAUT2 have been tested in different geographical contexts to understand the role of culture in technology adoption and solidify the generalisability of the theory tenets (Gupta, Dasgupta & Gupta, 2008; Im, Hong & Kang, 2011; Venkatesh, Thong & Xu, 2012). The majority of findings showed that the role of UTAUT constructs was significant irrespective of the difference in cultures. For example, the employment of the model in a comparative study on technology acceptance in the USA and China demonstrated the high explanatory power of the model across the two geographical settings. However, the model accounts for a greater variance in the behavioural intention when fewer moderators are tested (Venkatesh, Thong & Xu, 2012). When UTAUT was examined in Korea and the USA, the strength of the relationships slightly varied, although the significance was invariant across the two samples (Im, Hong & Kang, 2011). Similar results were observed when the UTAUT model was tested cross-culturally in individualistic vs. collectivistic nations. The model was shown to be viable in both types of cultures, but the strength of the relationships was different, suggesting a strong moderating role of culture on the model paths (Udo, Bagchi & Maity, 2016). UTAUT2 was also validated in different countries with contrasting cultures, economies and level of technology penetration. In Jordan, mobile banking adoption was not affected by social influence (Alalwan, Dwivedi & Rana, 2017). When comparing the adoption of education technology in Korea, Japan and the US, both the strength of the relationships and the significance of the effects were different across samples. For Korean users, the intention to use e-learning correlated with habit and perceived efficacy. For Japanese users, the behavioural intention was underpinned by habit, price value and social influence, while US users stressed only the habit and price value factors. Surprisingly, effort expectancy was not significant for any country, which might indicate that the technology being tested did not demand any effort to operate it (Jung & Lee, 2020). UTAUT2 applications demonstrate that insight into the conditions associated with culture is required, such as nations’ socio-economic status or norms.

The original and extended UTAUT models have been used to examine technology acceptance in a number of different sectors, such as healthcare (Chang et al., 2007), e-government (Gupta, Dasgupta & Gupta, 2008; Chan et al., 2010), mobile internet (Venkatesh, Thong & Xu, 2012; Thong et al., 2011), enterprise systems (Chauhan & Jaiswal, 2016; Ling Keong et al., 2012) and mobile banking and apps (Zhou, Lu & Wang, 2010; Mütterlein, Kunz & Baier, 2019). The applications of UTAUT demonstrated a strong dependence of behavioural intention on the two perception factors, namely perceived performance and perceived ease of use. For example, the technology acceptance framework was used to understand the acceptance of a pharmacokinetics-based clinical decision support systems. All constructs had significant effects on intention, except for facilitating conditions, which influenced only the actual utilisation of the technology (Chang et al., 2007). The investigation of the factors driving the adoption of e-government by employees in a state organisation in a developing country demonstrated the significant influence of all the UTAUT variables moderated by gender, while performance and effort expectancy showed the strongest effects (Gupta, Dasgupta & Gupta, 2008). When the model was used to explore the acceptance of ERP software training, three out of four predictors of use intention were found to be significant. While effort expectancy, performance expectancy and facilitating conditions influenced employees' intention to adopt training tools, the effect of social influence was not supported. Such findings were probably due to the instrumental nature of ERP software and the high contingency of its use on utility factors that overshadow the role of social influence on users’ decisions (Chauhan & Jaiswal, 2016). The applications of UTAUT2 showed that the significance and the strength of behavioural determinants differed across cases. The utilisation of UTAUT2 to investigate the antecedents of mobile app adoption confirmed the role of performance expectancy, social influence, hedonic motivation and habit (Mütterlein, Kunz & Baier, 2019). However, in two other studies investigating mobile banking adoption, the role of social influence was not confirmed (Ajzen, 2011; Baptista & Oliveira, 2015). The strongest observed effects were demonstrated by performance expectancy, hedonic motivation and habit (Baptista & Oliveira, 2015).

Practical Implications

UTAUT and UTAUT2 could have a number of applications in practice. UTAUT can be used to examine the anticipated acceptance rate of a product and ensure sufficient stock to satisfy the consumers’ demand. Evidence that the model provides can be used by practitioners to design more user-oriented products. UTAUT underscores the role of social influence and facilitating conditions, thus highlighting the importance of contextual analysis in strategies for technology implementation and promotion (Venkatesh, Thong & Xu, 2012). The application of UTAUT2 enables technology producers and vendors to measure how the trade-off between monetary price and the value of the product influences the utilisation of their technology. Companies have the opportunity to reconsider cost-structures to adjust the pricing policy to the relative value attached to the product, because the benefits that users get from the purchase of technology may not justify the price that they pay. By investigating the effect of habit on users’ intention, technology producers and distributors are able to define the marketing communication strategies that may address the beliefs that fuel automatic behaviour (e.g. advertising the utility of the product in various scenarios). By measuring the effect of hedonic motivation, product developers and managers can adjust the offering in such a way as either to enhance the hedonic value of technology or augment hedonic cues for marketing the product. Finally, the moderation effects in UTAUT2 enable practitioners to identify which user segment demands more marketing effort to address habits, deliver hedonic value and demonstrate better value for money (Venkatesh, Thong & Xu, 2012).

Limitations

The Unified Theory of Acceptance and Use of Technology provides a holistic tool to measure technology acceptance and technology use (Venkatesh et al., 2003; Venkatesh, Davis & Morris, 2007). However, despite the rigorousness of the model, UTAUT has some theoretical and methodological limitations that were not addressed in further studies (Venkatesh et al., 2003; Venkatesh, Davis & Morris, 2007). UTAUT faced critique with regards to its inability to explain behavioural intention in different settings. Limited external validity of the model motivated further studies to extend the model by adding additional determinants of behaviour, such as trust, self-efficacy, computer self-efficacy, innovativeness, perceived threats, perceived risk (Martins, Oliveira & Popovič, 2014; Slade et al., 2015). Also, the model was extended by introducing new moderating effects, such as income, location, culture, technology readiness (Im, Hong & Kang, 2011; Borrero et al., 2014) (for a more comprehensive insight see the review by (Venkatesh, Thong & Xu, 2016)). Still, some key factors, like computer self-efficacy, remained under-researched. Although it was confirmed that this factor plays a role in behavioural intention (Bandura & Locke, 2003), only an indirect effect of self-efficacy on intention was tested while developing UTAUT (Venkatesh et al., 2003).

The concern over the wide application of UTAUT was noted by Dwivedi et al. (Dwivedi et al., 2019), who stated that the majority of studies in the IS context cite the original UTAUT paper without using the model. Those surprising findings lead to the conclusion that UTAUT might not be as robust as it claimed to be, given overrated citations compared to the actual implication of the theory. Thus, based on the analysis of MASEM (Combined meta-analysis and structural equation modelling), a revised version of UTAUT was proposed, which included attitude construct as a partial mediator of the effects of exogenous constructs on behavioural intentions (Dwivedi et al., 2019).

The major methodological limitation of UTAUT concerns the development of the scales that were used to measure the core constructs. For the final measurement development, the study used the highest loading items for each scale. While this approach was supported by the literature (Hevner et al., 2004), there was debate as to whether it may be useful to validate the measurements or even develop new ones to eliminate potential content validity issues (Venkatesh et al., 2003). In addition, the intention to use and use behaviour scales were adopted from prior studies (e.g. (Davis, 1989)), but alternative measurements should be developed and validated in future studies (Venkatesh et al., 2003).

UTAUT2 also has some limitations inherent in the methodology. The model utilises a self-reported scale to measure intention to use (Venkatesh, Thong & Xu, 2012), which jeopardises the accuracy and validity of the research conclusions. UTAUT2 shares this limitation with many other technology acceptance models (e.g. TAM, original UTAUT) (Venkatesh et al., 2003; Davis, 1989)). In addition, similar to other technology acceptance models, UTAUT 2 can face a threat of common method variance (Straub & Burton-Jones, 2007; Sharma, Yetton & Crawford, 2009). To reduce the potential of common method bias, different methodological approaches need to be used (e.g. using experimental settings that can make manipulation checks possible).