The role of user satisfaction in advancing sustainable mobility: evidence from emerging market

Environmental degradation, rising fuel prices and advancements in battery technology have all contributed to the global trend toward sustainable mobility, as electric vehicles (EVs) are emerging as a vital player (IEA, 2021). The vehicles present various advantages likened to fossil-fueled vehicles [provided that electricity for charging is generated from renewable sources (Muratori et al., 2021)] such as zero tailpipe emissions, no reliance on petroleum, improved fuel economy, lower maintenance and convenient at home recharging (Collett et al., 2021; Collett et al., 2021; Sadeghian et al., 2022). The number of EVs globally reached about 57 million at the end of 2024, of which more than 90% are in China, Europe and America (IEA, 2024). However, in emerging markets, such as those in Africa, the adoption of EVs is still low. Despite this barrier, rising interest in sustainable solutions offers an opportunity to understand user satisfaction, helping bridge the gap between advanced and emerging markets (Dijk and Orsato, 2013).

Bearing in mind the potential of EVs in reducing air and noise pollution, it is essential to explore the cause influencing user satisfaction and their intent to continue using them (Cruz-jesus et al., 2023; Rehman and Jajja, 2025). However, research on satisfaction or dissatisfaction of EVs owners or users; especially from emerging countries’ perspective is very limited. Satisfaction with EVs is often linked to several factors, including vehicle performance, cost efficiency and the availability of charging infrastructure (Rezvani et al., 2015a; Collett et al., 2021; Rehman and Jajja, 2025). In emerging countries, however, the situation is different. A study by Hardman et al. (2018) recognized that these satisfaction determinate factors are layered with additional complexities, such as inconsistent electricity supply, regulatory uncertainties and socioeconomic factors. Respondents in previous researches never used EV users and are psychologically removed from the vehicles in this way, which reduces the reliability of conclusions on adoption derived from their answers (Rezvani et al., 2015b and Song et al., 2022). The contextual and methodological shortcomings in existing research underscore the need to explore the factors that influence EV users’ satisfaction within the specific context of emerging markets. Notably, prior studies have predominantly focused on the general public or potential adopters, rather than actual EV users or owners (Song et al., 2022). As a result, a comprehensive investigation into the satisfaction levels of current EV users is essential to gain deeper insights into the present realities and future prospects of EVs adoption.

The findings by Hasan (2021) highlights the relationship between battery electric vehicle (BEV) owners’ satisfaction and their intention to repurchase and recommend the vehicle to others. He further describes that when consumers plan purchasing new products, mostly costly and innovative items with significant uncertainty, such as BEVs, they tend to seek guidance from existing users, and are highly influenced by their evaluations. Another study by Jebril et al. (2021) found that, the installation of charging infrastructures and their types have an effect on the satisfaction of EV users in Jordan. A key component of providing services that comprehend and meet consumers’ requirements and wants is ensuring their satisfaction.

EVs, characterized by their low energy consumption, reduced emissions and other benefits, have got increasing attention (Cruz-jesus et al., 2023; Rahi et al., 2025). Interest in adopting EVs has been growing progressively across emerging markets, making the topic an increasingly relevant. Recent studies such as Moeletsi (2021) and Phuthong et al. (2024) have explored environmental, economic and infrastructural factors influencing users’ satisfaction. Pang et al. (2023) referred in their research that satisfaction is either material or spiritual. Material satisfaction relates to product use value, including quality and design. Spiritual satisfaction refers to the emotional and identity-related benefits customers derive from the product. In countries such as India, China and Brazil, financial incentives and supportive policies play a substantial role in shaping potential adopters’ decisions (Abdul et al., 2024). However, the limited availability of charging infrastructure remains a major obstacle in many emerging nations. A research by Moeletsi (2021) in South Africa found that addressing the challenges associated with charging infrastructure networks leads to higher levels of satisfaction among users. Infrastructure shortages in Sub-Saharan Africa hinder EV transition; limited charging and grids demand comprehensive public strategy including solar, network impacts, residential solutions, consumer needs and private engagement (Bartlett et al., 2001). Tanzania consumers give high value to driving range and low price as the most desirable attributes (Malima, 2025).

Key constructs include perceived value, expectation, and performance. Perceived value, the customer’s assessment of benefits relative to costs (of EVs in this research) (Zeithaml, 1988), mediates the relationship between expectations and satisfaction by increasing or diminishing perceived benefits (Chen and Chen, 2010). Satisfaction happens when performance meets or surpasses expectations, while unmet expectations lead to dissatisfaction (Oliver, 1997). Satisfaction, viewed as transactional or cumulative (Anderson and Fornell, 2010), results from the interplay of expectation, performance and perceived value.

A surge in interest in consumer satisfaction research has led to a number of careful analyses of the origins and consequences of satisfaction cognitions (Ollver, 1980). The latest developments in ECT continue to build on this original model developed by Oliver, which has been widely adopted and adapted, particularly in consumer satisfaction, marketing and information system research (Ryzin, 2004 and Chen et al., 2022). ECT has been extended to accommodate new variables and contexts, such as digital services, technology adoption and environmental psychology (Halstead et al., 2007; Ryzin, 2013) and in recent years, (Cruz-jesus et al., 2023; Rahi et al., 2025) used them in their EVs adoption researches. For instance, in information systems, researchers often integrate ECT with models like the TAM to study user satisfaction and adoption intention to continue using technologies (Bhattacherjee, 2001). However, a key limitation in the application is that only a few researchers like Cruz-jesus et al. (2023) have studied in the context of EVs.

The model provides a structured framework for considering factors affecting the adoption of technologies like EVs in emerging markets. TAM emphasizes the role of perceived usefulness (performance), where individuals assess how technologies provide practical benefits such as cost savings, reduced travel time, or environmental impact (Davis et al., 1989). In resource-constrained emerging markets, these perceived benefits significantly affect satisfaction and engagement with sustainable mobility options (Afroz et al., 2015). Individuals may also consider how easily they can access charging infrastructures, navigate apps for shared mobility, or maintain an EVs (Krishnan and Koshy, 2021; Tuan et al., 2022; Chandra et al., 2024). Challenges such as accessing charging infrastructures or navigating apps highlight its importance in emerging markets with technical or infrastructural barriers (Venkatesh and Bala, 2008; Krishnan and Koshy, 2021 and Chandra et al., 2024). This framework helps identify key determinants for improving satisfaction and promoting sustainable mobility (Afroz et al., 2015 and Heuveln et al., 2021).

EV users’ satisfaction is impacted by a complex interaction of economic, psychological and technological aspects (Lampo et al., 2022 and Rehman and Jajja, 2025). Economic incentives are critical to enhancing EV users’ satisfaction developing markets. Financial support systems, such as subsidies, tax reductions and reduced registration fees have been recognized as key drivers in increasing EV users’ satisfaction (Goel et al., 2021). Governments and interested parties could start lowering the cost of EVs to boost their market share. According to Electric Mobility Mission Plan 2020, financial incentives are key drivers of EVs adoption among customers (Ansab and Kumar, 2022). Important economic factors that adopters take into account include the initial purchase price, the availability of government subsidies, operational expenses (such as gas vs electricity) and maintenance costs, all of which have an impact on users’ pleasure. Therefore, EV users’ satisfaction is influenced by economic, psychological and technological factors, with financial incentives (such as subsidies, tax reductions and lower cost) playing a vital role in improving adoption and users’ satisfaction. As presented in Figure 1, EV performance, price, charging infrastructure and battery range are hypothesized to influence users’ satisfaction through perceived value and moderated by government incentive.

From this we hypothesize that:

The development of charging infrastructure is essential for providing convenient and accessible charging options, thereby enhancing the attractiveness of EVs (Abdul et al., 2024). Charging infrastructure plays a vital role in shaping EVs users’ satisfaction, as remarked by (Jaiswal et al., 2022; Cruz-jesus et al., 2023). Quicker charging time increases customer satisfaction by bringing the refueling time of EVs closer to that of conventional internal combustion engine vehicles (Muratori et al., 2021). Equally, investments in public charging infrastructure have boosted consumer confidence, improving user satisfaction (Patel et al., 2022). The distance can travel per full charge is one of the most essential aspects impacting EVs users’ satisfaction (Jaiswal et al., 2022; Cruz-jesus et al., 2023). Longer battery range EVs are typically preferred by users because they lessen range anxiety of battery power when traveling a considerable distance from a charging infrastructure. Concerns about sustainability also surface, including how to efficiently and sustainably deliver electricity from renewable energy sources, control the grid’s electrical demand and build new charging infrastructures (Muna and Kuo, 2022). Generally, charging infrastructure influences EV users’ satisfaction by ensuring accessibility, faster charging and extended battery range, thus reducing range anxiety, enhancing confidence and promoting sustainable energy mobility systems. Hence:

An EV’s ability to accelerate and handle can have a big impact on how satisfied a consumer is. Because electric motors produce rapid torque, many EVs can accelerate quickly, improving the driving experience (Galvin, 2017). Globally, the rate of yearly energy consumption per km slowed during the 2010s, but starting in 2019, there have been noticeable gains as sales shares of EVs have started to rise significantly (Serohi, 2022; ITS, 2023; IEA, 2024). The adoption of EVs, which include both BEVs and plug-in hybrid (PHEVs), is primarily responsible for the recent acceleration of energy efficiency improvements. This is because electric powertrains use three to six times less energy to cover a unit of distance than powertrains that rely on fuel engines (ITS, 2023). Therefore, EV performance, principally rapid acceleration and efficient powertrains, considerably improve users’ satisfaction whereas improving global energy efficiency through widespread adoption of both EVs. Hence:

A lack of effective urban sustainability planning and limited public participation in policymaking are key obstacles to achieving sustainable development (Hossen, 2025). Incentives are provided to facilitate the adoption of specific initiatives. While some countries offer significant subsidies and incentives to promote EVs adoption; others are still developing their EVs policies or may not provide similar benefits, potentially discouraging prospective buyers (Abdul et al., 2024). For instance, similar to Norway, the Danish government has supported BEVs by exempting them from the registration tax, which is relatively high in both countries compared to their neighboring nations (Haustein et al., 2021). Furthermore, technological advances in battery life, charging time and vehicle performance are key factors contributing to increased satisfaction levels in global markets (Krishna, 2021). Improvements in battery technology (Hasan, 2021 and Jaiswal et al., 2022) have helped solve concerns over the efficiency and reliability of EVs, leading to a more positive user experience. A major contributing reason to the satisfaction of many users is the environmental impact of driving an EVs as opposed to fuel cars (Manoj and Jayaprakash, 2023). Increased awareness of environmental benefits is positively correlated with higher satisfaction levels among EVs owners (Lee and Cho, 2021). The concept of perceived value is well-established in consumer and marketing literature. According to Chandra et al. (2024), “Perceived value refers to the consumer’s assessment of the benefits they receive from a product in relation to the price they pay”. It denotes a consumer’s evaluation of whether the quality and features received are worth for the price paid (Hamzah and Tanwir, 2021).

Overall, limited urban sustainability planning, policy gaps and unequal EV incentives hinder adoption, while improvements in battery technology, performance and environmental awareness significantly increase perceived value and overall EV users’ satisfaction globally. The authors hypothesize as follows:

This study uses quantitative research methods, which emphasize data that primarily rely on numerical analysis. The study model consists of elements adapted from previous articles; most of them were done in developed countries.

Ethiopia reportedly has about 30,000 EVs out of 1.2 million cars plying its roads in 2024 (Kassu, 2024). Based on the formula of Cochran (1977) for determining the sample size for a proportion under simple random sampling, when the population is finite, is adjusted by the finite population correction (FPC), the sample size became 379. The sample size required for a 5% margin of error at a 95% confidence level with maximum variability (assuming p = q = 0.5), was again adjusted using Bartlett et al. (2001) formula to consider Anticipated Return Rate (90%) and the final sample size is 422.

The authors used a systematic random sampling technique, wherein every Kth unit was selected from the sampling frame randomly, which consisted of the license plate numbers of EVs. According to Chandra et al. (2024), “In systematic sampling, every Kth individual is chosen in a continuous manner from a population list, starting from a random point.” A random starting point 58 was determined through a lottery draw from the range of 1–71, representing the sampling interval. Subsequent selections followed a fixed interval of 71 units, yielding the sequence 58, 129, 200 […] continuing systematically through the population of 30,000 EVs plate numbers.

Quantitative data obtained from systematic randomly selected EV users in Ethiopia through paper-based surveys administered from September to November, 2024 (there is no significant seasonal as well as fuel price variation effect in this duration). A seven-point Likert scale was used for questions related to variables. This type of scale is suitable for analytical tools such as SEM (Jaiswal et al., 2022). The scale is structured as follows: 1 represents strong disagreement, 2 disagreement, 3 somewhat disagree, 4 neutral, 5 somewhat agree, 6 agree and strongly agree was denoted by 7. The questionnaires were distributed in Addis Ababa City, where nearly all EVs are located. Questionnaires were delivered and subsequently collected in person. Out of the 422 targeted EV users, 326 completed the questionnaires in full and returned, resulting in a response rate of 77.25%. Incomplete and unreturned questionnaires, totaling 96, were excluded from further analysis. The instrument, comprising 37 questions, was developed based on insights from previous studies.

SmartPLS4.1.0.9 (Ringle et al., 2024) and SPSS 20 were used to analyze the collected data. Partial least squares (PLS)-SEM is a causal modeling technique designed to maximize the enlightened variance of dependent latent variables (Hair et al., 2014 and Hair et al., 2017). PLS-SEM was chosen due to its diverse advantages. First, PLS-SEM has minimal requirements for measurement scales. Second, it can accommodate both reflective and formative indicators for latent variables. Third, it is compatible for analyzing small sample sizes and data that do not follow a normal distribution. Hair et al. (2018) stated that “greater statistical power means that PLS-SEM is more likely to identify relationships as significant when they are indeed present in the population”. A PLS path model is composed of two components. The first is a structural model (the inner model of PLS-SEM) that connects the constructs. The second component consists of the measurement models (PLS-SEM outer model) of the constructs, which depict the relationships between the indicator variables and the constructs (Hair et al., 2022).

Valid data were obtained from 326 respondents and their demographic characteristics are presented in  Appendix Table AI. The majority of respondents are male 288(88.3%), with females comprising 38(11.7%). Most respondents are 36–45 years, 181(55.5%), followed by 26–35 years 129(39.6%). Younger and older age groups constitute a small proportion 6(1.8%) and 10(3.1%), respectively. About education, above half of the respondents hold a master’s degree 186(57.1%), while bachelor’s degree holders make up 93(28.5%). Those with a PhD or above account for 40(12.3%) and the rest 7(2.1%) are Grade 12 complete. Occupationally, most respondents work in the public sector 181(55.5%), followed by the private sector 109(33.4%). Self-employed individuals denote 22(6.7%), while students and others form smaller segments 2(0.6%) and 12(3.7%). Monthly family income is concentrated in the 10,000–30,000 Birr range (49.4%), with 16.9% earning above 90,000 Birr. Smaller groups fall within other income brackets. Concerning EVs, hybrid models are the most popular choice 176(54%) and plug-in vehicles 35(10.7%) are least preferred. Apart from demographic questions, the variable scales were developed from studies on factors like price value (Venkatesh et al., 2012), perceived value (Li and Shang, 2020; Yang and Peterson, 2004), vehicle performance (Tuan et al., 2022; Uzir et al., 2021), charging infrastructure (Tuan et al., 2022), government incentives (Tuan et al., 2022) and customer satisfaction Cruz-jesus et al., 2023; Uzir et al., 2021) and modified to fit the study purpose.

This study used Cronbach’s alpha coefficient and composite reliability (CR) values to evaluate the inter-item consistency of the measurement items. For both Cronbach’s alpha and CR, a value of 0.708 or higher is considered acceptable (Hair et al., 2014). Hair et al. (2022) clearly put that, reliability is classified as excellent if it surpasses 0.9, good if above 0.8, adequate if above 0.7, questionable if greater than 0.6, and poor if below 0.5. As Table 1 depicts Cronbach’s alpha (α) values across all constructs exceed the minimum threshold of 0.708, confirming internal consistency. To ensure the survey’s effectiveness, a pilot test was conducted to identify issues before full deployment. Shukla (2008) suggests a sample of 15–30; thus, the authors surveyed 26 EVs users in Addis Ababa. While not fully representative, this approach helped assess question clarity and potential ambiguities. The CR values (both rho_a and _c) additionally validate the reliability. All constructs have CR values greater or equal to 0.708, with most having values greater than 0.900. Thus, construct reliability was successfully established.

Validity refers to the accuracy of an evaluation, determining whether the theoretical and practical meanings truly reflect the fundamental concept being assessed (Hair et al., 2022). Convergent validity is assessed through factor loadings and average variance extracted (AVE). All items across constructs have loadings > 0.5, with most above 0.8, demonstrating strong indicator reliability. Government incentive (0.960) and charging infrastructure (0.942) have higher loadings and, signifying that these indicators effectively represent their respective constructs. Even constructs with slightly lower AVE, such as perceived value (AVE = 0.644) and Price (AVE = 0.734), exceed the minimum requirement, confirming their acceptability.

Discriminant validity was established using the Fornell and Larcker criteria (Fornell and Larcker, 1981) and the HTMT criteria. According to the Fornell and Larcker ratio (Table 2), the square root of AVE for each research construct was confirmed to be greater than the corresponding correlation values between each pair of constructs (Uzir et al., 2021; Hair et al., 2022). Hence, the finding in the table shows that square root of AVE is above the matching row and column value confirming discriminant validity requirements are proven.

Similarly, Heterotrait–monotrait ratio (HTMT) test (Table 3) was done and validity was established. It is another enhanced method, as an alternative of Fornell–Lacker criterion, to check discriminant validity. Hair et al. (2022) and Zhang et al. (2021) defined HTMT as “the mean value of the indicator correlations across constructs”. Hair et al. (2022) suggest the HTMT of correlations to assess discriminant validity stating that “While the Fornell–Larcker criterion has long been the primary criterion for discriminant validity assessment, more recent research highlights that the HTMT criterion should be the preferred choice”.

The model fit summary in Table 4 approves a robust and well-fitting structural model. The SRMR values for both the estimated model (0.031) and the saturated model (0.030) are well below the recommended threshold of 0.08, indicating a minimal difference between observed and predicted correlations. Additionally, d_ULS (0.452) and d_G (0.675) remain within acceptable ranges, further supporting the model’s validity. The Chi-square values (1078.731 for the saturated model and 1088.016 for the estimated model) suggest an adequate model fit, though chi-square is sensitive to sample size. The Normed Fit Index (NFI) values (0.905 and 0.904) exceed the 0.90 threshold, demonstrating a satisfactory incremental fit.

Once satisfactory results were obtained for the measurement model, the structural model was also evaluated using numerous fit indices and the coefficient of determination (R²), followed by hypotheses testing. The R² values for Perceived Value (0.434) and Satisfaction (0.781) (as shown on Figure 2 and in Table 5) indicate that the independent variables explain a substantial proportion of variance, particularly in Satisfaction, affirming the model’s predictive strength and its relevance in understanding EVs users’ perceptions. “As a general guideline, R² values of 0.75, 0.50 and 0.25 can be considered substantial, moderate, and weak, respectively” (Hair et al., 2022).

The Q² values (0.419 and 0.773) indicate the predictive power of the model in the analysis.

These values assess how well the model predicts unseen data, ensuring that the relationships identified are generalizable beyond the sample data set. “As a rule of thumb, Q² values higher than 0, 0.25 and 0.50 depict small, medium and large predictive relevance of the PLS-path model” (Hair et al., 2018). In this study, the Q²predict values of 0.419 and 0.773 indicate that the model exhibits strong predictive accuracy, suggesting that it is well-suited for predicting future or out-of-sample data. The findings reinforce the robustness and validity of the proposed model, particularly in the context of EVs user satisfaction, adoption, or perceived value.

Variance inflation factor (VIF) is used to measure multicollinearity among predictors and all values are below five. Multicollinearity is a significant issue for formative constructs, as the first-order constructs collectively contribute to the variance of the related formative second-order constructs (Li and Shang, 2020). A VIF value below 3 is generally acceptable, indicating that multicollinearity is not a concern. Collinearity issues are commonly uncritical if VIF = 3–5 (Hair et al., 2022).

Hypothesis testing outcomes convey significant insights into the factors influencing EVs users’ satisfaction. The remarks of Hair et al. (2022), “The examination of total effects between constructs, including all their indirect effects, provides a more comprehensive picture of the structural model relationships” is fully taken for the test. In H1 (as shown by Figure 2 and Table 6) price influences satisfaction negatively and statistically significant (β = −0.136, T = 2.202, 95% BC CI [−0.240, −0.038]), suggesting that higher EVs prices reduce user satisfaction. This finding aligns with prior studies such as (Tuan et al., 2022) emphasizing price sensitivity among consumers in emerging markets, where affordability remains a key barrier to EVs adoption. Likewise, charging infrastructure (H2) proves a positive and significant impact on satisfaction (β = 0.219, T = 2.824, 95% BC CI [0.083, 0.337]), indicating that well-developed charging networks enhance user experience. Krishna (2021) and Hasan (2021) supported this finding that accessibility of public charging and workplace charging influence EVs uptake considerably. The distance traveled on each trip by an EVs is crucial, as the range remains limited, particularly given the frequent lack of charging infrastructures (Cruz-jesus et al., 2023). This result highlights the key role of charging accessibility in boosting EV adoption and reducing range anxiety in areas with limited infrastructure.

On the other hand, H4 was supported in the direct relationship (no mediation and moderation as shown in Figure A2) with score of β = 0.246 and p -value < 0.05. However, the hypothesis is not or closer to be supported with the presence of mediator and moderator (β = 0.009 and p -value = 0.063). H3 is not supported as its path coefficient is β = 0.010 and p -value > 0.05 at 95% level of confidence. These findings suggest that, in the context of this study, battery range and performance (almost nearly significant) do not significantly influence user satisfaction. One possible explanation is that current EVs models in the studied market offer adequate range and performance, leading users to prioritize cost and infrastructure over these factors (Muratori et al., 2021). Alternatively, perceptions of battery range and performance might be shaped by expectations rather than actual experience, warranting further investigation into the role of expectation-confirmation in EVs users’ satisfaction. Besides, most respondents (55.2%) have owned their EVs for less than six months, suggesting recent adoption. Additional reason may be the difference in culture with developed economy that leads to variation in avoiding uncertainty (Hofstede, 2011).

H5. To examine the mediating effects within the relationships, the bootstrapping method recommended by Hair et al., (2022) and Hair et al., (2018) was used to ensure accurate outcomes when evaluating the confidence intervals (CIs) of the indirect relationships. The analysis (as shown in Table 7) provides critical insights into how perceived value influences the relationship between the attributes and user satisfaction. For battery range (H5a), perceived value does not mediate the relationship with satisfaction, as the direct effect (β = 0.028, p > 0.05) remains insignificant, while the indirect effect through perceived value (β = 0.017, p < 0.05) is also weak. This proposes that users may not perceive battery range as a determinant value, possibly because current EVs’ models meet their daily mobility needs. Similarly, vehicle performance (H5c) reveals a full mediation role, meaning perceived value fully describes the link between performance and satisfaction. The direct path (β = 0.062, p > 0.05) is insignificant, while the indirect effect via perceived value (β = 0.032, p < 0.05) is significant, indicating that users do not derive satisfaction directly from performance but rather from how they perceive its contribution to overall value.

On the other hand, charging infrastructure (H5b) and price (H5d) demonstrate partial mediation effects, suggesting that both factors influence satisfaction directly and indirectly through perceived value. The direct effect of charging infrastructure on satisfaction (β = 0.214, p  < 0.05) remains significant, but its indirect impact through perceived value (β = 0.023, p  < 0.05) is also noteworthy. This implies that while charging accessibility directly enhances satisfaction, users also appreciate its contribution to perceived value. In contrast, price negatively impacts satisfaction (β = −0.151, p  < 0.05), with an additional negative indirect effect through perceived value (β = −0.029, p  < 0.05). Researches (Li and Shang, 2020) support the mediation role of perceived value in measuring technology adoption. Studies (Uzir et al., 2021) indicate that perceived value enhances customer satisfaction by shaping their expectations and experiences, ultimately leading to loyalty. For example, Liang et al. (2024) found that perceived value significantly influences customer-brand relationships and loyalty in sustainable mobility contexts.

H6. Government incentive is moderating the negative relationship of price and satisfaction. The study assessed the effect of government incentive on price’s influence on satisfaction. As Table 8 reveals, with the absence of government incentive as moderators, R-square was 0.732 or 73.2%, whereas it was changed to 0.781 or 78.10% when the moderator was included. The difference in government incentive included is 0.781–0.732 = 0.049 or 4.9% with a significance moderation role at 5% confidence level. The negative direct effect (Figure A1) is (−0.234*), and it indicates that higher EVs prices lower user satisfaction. The interaction effect (−0.151*) suggests significant moderation, meaning government incentives help mitigate the negative influence of price on satisfaction. When incentives are provided, the adverse impact of high prices on satisfaction is reduced, making EVs more attractive to consumers (Abdul et al., 2024). This highlights the crucial role of supportive policies in improving customer perception and market adoption.

Importance performance map (Figure 3) describes six variables: battery range, charging infrastructure, government incentive, perceived value, performance and price. They are mapped based on their relative significance (x-axis: total effects) and performance (y-axis: satisfaction levels). From the map, government incentives (purple) and perceived value (orange) show high importance and relatively strong performance, suggesting that they are key determinants of user satisfaction. Charging infrastructure (cyan) also has a high importance score but moderate performance, indicating a potential area for improvement. Battery range (red) and vehicle performance (green), despite having moderate performance, show low importance, which aligns with the hypothesis testing results where these factors were not significant predictors of satisfaction. Remarkably, price (blue) has negative importance and lower performance, reinforcing its negative impact on satisfaction, as users may perceive EVs as expensive relative to their value.

Grounded in CST, this study examines the direct and indirect effects of EVs charging infrastructure, price, performance and battery range on user satisfaction in Ethiopia, an underexplored emerging market. Using PLS-SEM, the analysis highlights perceived value as a mediator and government incentives as a moderator variable. Data were collected from 326 EVs users in Addis Ababa. The authors confirmed reliability and validity through Cronbach’s alpha, AVE, composite reliability, HTMT, Fornell–Larcker and cross-loadings, ensuring model robustness. Hypothesis testing shows that price negatively affects satisfaction (β = −0.136, p  < 0.05), supporting with affordability concerns in emerging markets (Cruz-jesus et al., 2023). Conversely, charging infrastructure significantly enhances satisfaction (β = 0.219, p  < 0.05), reinforcing its role in user experience (Hasan, 2021). However, performance (β = 0.009, p  > 0.05) and battery range (β = 0.010, p  > 0.05) do not significantly influence satisfaction, contrasting with findings from developed markets (Cruz-Jesus et al., 2023). These results suggest that Ethiopian EVs users prioritize cost and infrastructure over technical attributes, possibly due to short-distance travel patterns and adequate standard performance in available models.

Mediation analysis denotes perceived value’s role in determining satisfaction. Charging infrastructure positively influences satisfaction both directly (β = 0.214, p  < 0.05) and indirectly through perceived value (β = 0.023, p  < 0.05), reinforcing its double impact. Price exerts a negative direct effect (β = −0.151, p  < 0.05) and indirect effect (β = −0.029, p  < 0.05), indicating that high EVs’ costs lessen perceived value and, subsequently, satisfaction. Remarkably, performance is fully mediated by perceived value (β = 0.032, p  < 0.05), meaning users evaluate it based on its contribution to overall value rather than as an independent factor. In contrast, battery range reveals no mediation, suggesting its low relevance in this context. Moderation analysis emphasizes the role of government incentives in offsetting price concerns. Incentives significantly moderate the price-satisfaction link (β = 0.151, p  < 0.05), mitigating affordability constraints. The strong direct impact of price (β = −0.234, p  < 0.05) reinforces the need for policy interventions to enhance accessibility and user satisfaction.

By applying Expectation-Confirmation Theory (ECT) to sustainable mobility, the study highlights affordability, infrastructure and perceived value as key post-adoption factors in satisfaction. It extends ECT by showing that user expectations evolve with external conditions, not just initial perceptions. This is consistent with the researches of Abdul et al. (2024), Cruz-jesus et al. (2023) and Li and Shang (2020). Moreover, this study provides a theoretical contribution by positioning perceived value as a full mediator in the EVs adoption process. While prior research has examined perceived value as a direct determinant of satisfaction, this study illustrates how it mediates the effects of core EVs’ attributes: charging infrastructure, price, performance and battery range on user satisfaction. This insight enriches value-based theories, such as the Technology Acceptance Model (TAM), by demonstrating that consumers do not evaluate EVs’ attributes in isolation but rather in terms of their collective contribution to overall value perception.

Beyond theoretical advancements, this study offers valued practical insights for policymakers, automakers and infrastructure designers seeking to improve EVs acceptance in emerging market. Given that affordability and charging infrastructure significantly impact satisfaction, policymakers should prioritize financial incentives (e.g. subsidies, tax exemptions) and investments in widespread charging networks. This approach aligns with consumer priorities and can accelerate EVs adoption in resource-constrained settings which are in line with the findings of Phuthong, et al. (2024) and Moeletsi (2021). EVs manufacturers should focus on optimizing perceived value rather than solely improving technical specifications. This means designing cost-effective models with a balance between affordability and performance, ensuring that price-sensitive consumers perceive sufficient benefits to justify their investment.

Findings indicate that charging accessibility is a stronger determinant of satisfaction than battery range in emerging markets. Developers should enhance charging infrastructure density, especially in urban and suburban areas, to ease range anxiety and improve convenience Abdul et al. (2024) and Cruz-jesus et al. (2023). As post-adoption experiences shape satisfaction, automakers and service providers should strengthen support systems, including user education, after-sales service and usage incentives. Although infrastructure and subsidy costs are high, the long-term benefits such as lower emissions and health costs, reduced fuel imports, improved energy security and boosted green innovation generally outweigh them.

Ethiopia, with its big renewable energy from hydropower, photovoltaic, wind and geothermal sources (IEA, 2024, 2023; ITS, 2023; Collett et al., 2021), can benefit a lot by a significant reduction of billions of dollars for importing fuel per year if it gives high priority to establishing charging infrastructures, by minimizing blackouts (Hardman et al., 2018) and improving government incentive policies (though tax exemption inconsistencies are being observed on EVs). This research bridges key gaps in EVs user satisfaction literature by integrating established theories with empirical insights from emerging markets. By demonstrating the dominant influence of affordability and infrastructure over traditional technical attributes, the study offers a refined understanding of EVs adoption dynamics in developing economies. Practically, it offers actionable insights for stakeholders aiming to accelerate sustainable mobility transitions through strategic policymaking, tailored business models and infrastructure investments (enhancing private charging infrastructures as a business entity with reasonable recharging cost to EVs users).

This research, while offering valuable insights, has certain limitations that should be acknowledged. The study focuses on EVs users in an emerging market context (Ethiopia) that has distinctive culture, economic status and rules and policies, which may limit the generalizability of the findings to developed economies where infrastructure, incentives and consumer behaviors differ significantly. Methodologically, it may have sampling bias caused by human error and outdated frame.

Future studies could:

• compare markets to test if satisfaction drivers vary across economic and regulatory settings;

• address the limited generalizability due to the small sample by using larger, more representative samples to boost statistical power; and

• complement the current quantitative approach with qualitative methods for deeper insights.

The cross-sectional design restricts causal inference, and reliance on self-reported data may introduce bias. Longitudinal and mixed-method designs are recommended to enhance validity and depth.

External funding was not received for this research.

The authors acknowledge the use of OpenAI's ChatGPT for language refinement in the preparation of this manuscript.