One of the most mundane activities for grown-ups is grocery shopping (Singh & Söderlund 2020). To make grocery shopping more convenient and fun for shoppers, mobile grocery shopping apps have become a pivotal strategy for retailers to compete in the multi-trillion-dollar grocery industry. Mobile grocery shopping apps offer various advantages to shoppers, such as portability, mobility, personalisation and ubiquity (Kumar, Chakraborty & Bala 2023).

Younger consumers dominate mobile grocery shopping services, while fewer older consumers use them (Hoh et al. 2023). This divide in mobile grocery shopping application usage is notable in South Africa (Mastercard 2023). However, across the world, the elderly segment presents good business opportunities in terms of size, wealth and consumption patterns, and businesses should proactively rethink strategies to fit this market segment (Chaouali & Souiden 2019). Thus, even in an emerging market country such as South Africa, attracting more elderly consumers to mobile grocery shopping services could assist retailers in recouping the costs of developing and maintaining mobile grocery shopping services, and tapping into the wealth of this niche segment.

Prior research on mobile grocery shopping in emerging markets made valuable contributions by mainly conducting research among the general population of a country (e.g. Chakraborty 2023; Hoh et al. 2023; Kim 2021). However, few respondents in these samples represented the elderly, defined as individuals 65 years and older (Berg & Liljedal 2022). The objective of this study is to investigate the determinants of elderly South African consumers’ intention to adopt mobile grocery shopping services, using one of the most cited technology adoption theories for research in a variety of settings and populations – the unified theory of acceptance and use of technology 2 (UTAUT2) framework (Blut et al. 2021). Such a study will contribute to the theoretical understanding of the elderly’s perceptions of mobile grocery shopping services, influencing their intention to use them.

The value of the study is as follows. The findings of this study add vital insights, filling gaps in knowledge and serving as a basis for future-related research in emerging markets. A fundamental task in developing a marketing strategy is the identification of meaningful sources of heterogeneity within a market, which enables the design of differentiated and segment-specific approaches to effectively engage diverse consumer groups. It is believed that the elderly as a market segment may not be homogenous in behaviour when using mobile services (Arenas-Gaitán, Villarejo Ramos & Peral-Peral 2020). Therefore, from a theory-building perspective and developing meaningful insights for retailers, it is essential to explore moderating variables that may highlight differences among elderly consumers regarding their intentions to use mobile grocery shopping services.

The manuscript is organised as follows. It begins with a review of mobile grocery shopping research in emerging markets to position the study in the existing literature. Next, it discusses the UTAUT2 model and formulates hypotheses based on its tenets and elderly consumer literature. Following this, the research methodology and data analysis plan are outlined. The main results of the hypothesis testing are presented, followed by a discussion of these findings. Finally, the manuscript concludes with the study’s theoretical contributions, managerial implications, limitations and suggestions for future research.

Table 1 presents an overview of research published in academic journals on adopting mobile grocery shopping services in emerging markets.

The key theories used in prior research are the technology acceptance model (Al Amin et al. 2021; Bruwer, Madinga & Bundwini 2022; Pasupuleti, Jeyavelu & Seshadri 2021; Yuuiarty et al. 2020; Shukla & Sharma 2018), theory of planned behaviour (Kim & Moon 2023; Shukla & Sharma 2018) and the stimulus-organism-response model (Chakraborty 2023; Ligaraba et al. 2023). Other theories used in prior research include the valence theory (Hoh et al. 2023) and the uses and gratifications theory (Al Amin et al. 2021).

In Table 1, prior research primarily focused on collecting data from the general population in Malaysia, South Africa, India and South Korea, resulting in only a few respondents being elderly shoppers (65 years old and older). Also, some of the studies in Table 1 (Al Amin et al. 2021; Kim 2021; Pasupuleti et al. 2021) do not include any respondents who met the criteria for being elderly.

In conclusion, previous research on mobile grocery shopping in emerging markets has primarily focused on the general population, neglecting to investigate elderly consumers’ adoption of such a service. Notably, the summary in Table 1 does not include a UTAUT2 study. According to Tamilmani et al. (2021b), the UTAUT2 has two unique characteristics that make it more suitable than other frameworks for the study. The UTAUT2 has high predictive validity, around 74% of behavioural intention. Also, compared to the TAM and TPB, the UTAUT2 focuses on adopting new technologies by incorporating intrinsic motivation. Thus, using UTAUT2 to investigate the intention of the elderly to adopt mobile grocery shop services in the emerging market context would add valuable insights to close the gap in the literature.

The theoretical model used to achieve the objective of the study is shown in Figure 1. The model is based on the UTAUT2 framework, which was proven valid for studying consumer technology adoption in emerging markets (e.g. Malekpour et al. 2025). Tan and Ooi (2018) asserted that facilitation conditions become irrelevant in mobile shopping intention because of prior mobile device experience. In the context of this study, elderly mobile grocery shopping users must be smartphone users to use mobile grocery shopping services. Being smartphone users, the elderly would also have experience in using mobile apps. Therefore, facilitating conditions were excluded as a variable from the conceptual model of the study. Lastly, age and gender were included as moderators based on the recommendation by Blut et al. (2021) that UTAUT studies should, at a minimum, include the two moderators.

To address the study’s objective, a non-probability sampling procedure was used to collect data from 269 elderly consumers who own a smartphone and live in five retirement complexes in two South African cities. All questionnaires were screened for completeness and consistency. No cases were removed because of missing data. The retirement complex manager distributed the questionnaires to all residents who were willing to participate in the study. Sixty-five is the most used threshold age in marketing research on the elderly (Berg & Liljedal 2022). Therefore, all respondents had to be 65 years or older. Forty-three per cent of the respondents were males. Most male respondents had a monthly spendable income of R10 001–R15 000 (24.8%), made two shopping trips monthly (38.8%) and spent R501–R1500 per trip (73.6% = 36.8% + 36.8%). Most female respondents also had a monthly spendable income of R10 001–R15 000 (33.6%), made four or more shopping trips per month (53.0%) and spent R501–R1000 per trip (42.1%).

Data were collected using hard-copy questionnaires distributed by the management offices in the retirement complexes. This data-collecting process is similar to the study of Kelfve et al. (2020). The multi-item scale used to measure each variable in the conceptual model was adapted from previous studies. The measurement model included 29 items across seven constructs. Performance expectancy was measured using five items developed by Pappas et al. (2014). Effort expectancy was assessed with four items, also from Pappas et al. (2014). Social influence was measured using four items sourced from Yang (2010) and San Martín and Herrero (2012). Price-value and hedonic motivation were each measured using three items developed by Venkatesh et al. (2012). Habit was captured using four items also from Venkatesh et al. (2012), and behavioural intention was measured using three items from Chen, Hsu and Lu (2018). More information on the scales is available in Appendix 1. The respondents could indicate their agreement with each statement using a 7-point Likert scale (1 = strongly disagree and 7 = strongly agree).

SmartPLS version 4 was used to test the hypotheses. The guidelines outlined by Hair et al. (2019) were considered for assessing the reflective measurement model and reporting the results of the structural model assessment.

The outer loadings were inspected for indicator reliability. Evidence of indicator reliability is an outer loading of 0.708 and higher. The internal consistency reliability of the multi-item scales was assessed using Cronbach’s alpha and composite reliability (CR). Evidence for internal consistency reliability is an index value of at least 0.7. Convergent validity is established by confirming that each construct’s average variance extracted (AVE) is above 0.5. This study evaluated the multi-item scales used in the survey for discriminant validity using heterotrait-monotrait ratios of correlations (HTMT). The benchmark for discriminant validity in a measurement model is the correlation ratio that does not exceed 0.85.

The structural model was assessed using a three-step process, which involved testing for collinearity, estimating path coefficients and interpreting the model’s predictive accuracy (Hair et al. 2019). R² was used to assess the model’s in-sample predictive accuracy. The PLSpredict algorithm was used to compare the prediction errors of the Partial Least Squares (PLS) path model against simple mean predictions. A positive Qpredict2 value indicates that the prediction error of the Partial Least Squares Structural Equation Modeling (PLS-SEM) results is smaller than the prediction error of simply using mean values (SmartPLS 2024b). The cross-validated predictive ability test (CVPAT) is an alternative to PLSpredict. Cross-validated predictive ability test applies an out-of-sample prediction approach to calculate the model’s prediction error, which determines the average loss value. The average loss value is compared to the average loss value of a prediction using indicator averages (IAs), which is considered a naïve benchmark, and the average loss value of a linear model (LM), which is considered a more conservative benchmark. The difference of the average loss values should be below zero and statistically significant to claim that path model has better predictive capabilities compared to the prediction benchmarks (SmartPLS 2024a). Age and gender were tested separately as moderators, because the moderation hypotheses did not state, for example, that the moderation effect of gender depends on the age of the elderly. As age is a continuous variable in the study, the moderation hypotheses were tested by creating an interaction effect of the independent variable and the moderator. The default data metric of standardisation was used in the moderation analysis to create the interaction effect.

The study set out to identify the factors influencing South African elderly grocery shoppers’ intention to use mobile grocery shopping services. Our study provides evidence of the robustness of the UTAUT2 in explaining behavioural intentions, as the variance in the elderly’s behavioural intention was 76%. Effort expectancy, subjective influence, price-value, hedonic motivation and mobile shopping habit positively influenced the elderly’s intention to use mobile grocery shopping services. Only performance expectancy did not statistically significantly influence behavioural intention.

Grocery shopping is primarily a utilitarian task; therefore, it is reasonable to expect that performance expectancy would influence behavioural intention in the study. This reasoning also aligns with the systematic review by Yap et al. (2022) that included the influence of performance expectancy on the behavioural intention of the elderly technology users. However, in this study, performance expectancy did not influence intention. A possible explanation for this contradictory finding is the smaller screens on mobile phones, coupled with the elderly’s psychological and physical challenges, which could make the elderly doubt the usefulness of mobile grocery shopping applications. It might also be that the elderly have little interest in using mobile grocery shopping applications. Braun and Osman (2024) state that the elderly refuse to engage in new trends and prefer to shop at physical stores.

The strong positive influence of effort expectancy on behavioural intention aligns with Naatu, Selormey and Naatu (2025). The smaller screens of mobile phones, combined with the psychological and physical challenges of the elderly, may have caused the elderly grocery shoppers in the study to doubt the usefulness of grocery shopping apps.

Hedonic motivation had the second strongest influence on behavioural intention. The strong influence may arise from older adults using the entertainment value of mobile shopping to justify their choice in light of the challenges they might perceive in using mobile grocery shopping services (San-Martín, Prodanova & Jiménez 2015).

The positive, statistically significant influence of subjective norms, price-value and habits aligns with prior research using the UTAUT2 to investigate technology adoption in an emerging market. Hoque and Sorwar (2017) reported that subjective influence, a variable similar to subjective norm, had a weak influence on the intention to use mHealth services in Bangladesh. In their study conducted in Iran, Malekpour et al. (2025) reported that perceived price-value and habit positively influence the use of electronic channels for grocery shopping. The focus of the study can explain the weak influence of habit – the elderly’s intention to use mobile grocery shopping services – justifying including (general) mobile shopping habits and not specifically the habit of using mobile grocery shopping services.

The statistically significant moderation effects of age and gender mirror established views regarding age and gender differences in technology adoption. Crucially, the statistically significant moderation effects indicate that such differences can exist in the elderly segment in an emerging market country.

In conclusion, the findings of this study challenge the core assumptions of the UTAUT2 framework by demonstrating that effort expectancy, rather than performance expectancy, is the primary driver of intention among elderly users in emerging markets. This deviation highlights the necessity of testing existing technology adoption frameworks to identify age-related and contextual factors. Additionally, the significant moderation effects of age and gender emphasise that elderly consumers represent a heterogeneous group, reinforcing the importance of tailoring technology adoption models to specific demographic and market contexts.

The study’s contributions to theory building are as follows. Previous research on mobile grocery shopping in emerging markets largely overlooked the elderly, a key demographic for retailers in these countries. In response to the call by Braun and Osman (2024) for more research on the elderly’s use of online shopping channels, this study presents new insights into elderly shoppers’ mobile grocery shopping behaviour in emerging markets.

The UTAUT2 is considered the most comprehensive theory in understanding individual technology adoption and use (Tamilmani, Rana & Dwivedi 2021a). Our study provides more evidence of the robustness of the UTAUT2 in explaining behavioural intentions, as the explained variance in the elderly’s behavioural intention was 76%. The study’s results also support the high predictive validity of the UTAUT2 (Tamilmani et al. 2021a).

Specific theoretical contributions can be presented by comparing the study results to the UTAUT meta-analysis by Blut et al. (2021). The meta-analysis results of Blut et al. (2021) show that performance expectancy should influence behavioural intention more strongly than effort expectancy. In contrast, effort expectancy had a statistically significant influence on intention behaviour in this study and performance expectancy did not influence behavioural intention statistically significantly. The non-statistically significant influence of performance expectancy challenges the key premise that the technology should be useful for it to be considered to be used by an individual (Venkatesh et al. 2003). The influence of hedonic motivation on the behavioural intention of the elderly was stronger than the influence reported by Blut et al. (2021), while the influence of habit was weaker than that reported by Blut et al. (2021). These results present evidence that the behavioural intention of elderly grocery shoppers to use mobile shopping services in an emerging market is influenced by a unique combination of variables. Thus, researchers cannot generalise from previous research which UTAUT variables would influence the behavioural intention of the elderly to use mobile grocery shopping services.

The findings of age and gender differences in the UTUAT2 relationship further elevate the theoretical contributions of the study. In the broader body of knowledge on mobile grocery shopping services, scant evidence of such differences is reported on the elderly’s use of innovative services. The findings on age and gender differences significantly bolster the argument that the elderly market segment in emerging markets is heterogeneous rather than homogeneous.

As effort expectancy strongly influences behavioural intention among elderly grocery shoppers, South African grocery retailers should focus on minimising the elderly’s perceived effort required for using their mobile grocery shopping app. It is imperative to implement clear and intuitive navigation mechanisms, underscored by descriptive icons, to facilitate ease of use by the elderly. A minimalistic design approach, prioritising essential features, is advocated to streamline user interaction and focus attention on core functionalities. Adopting consistent layout and design patterns is crucial for fostering user familiarity and reducing the cognitive load of learning new interface structures. Accessibility considerations can be improved by ensuring the app supports system-wide font size adjustments. The user interface should also have a high contrast between text and background to improve readability.

Developers and marketers must recognise the importance of hedonic motivation in shaping the user experience for this demographic. South African grocery retailers should integrate features that augment the intrinsic enjoyment of the shopping process, such as personalised recommendations to make the shopping experience more relevant and engaging, and social interaction features can provide a sense of community and belonging.

The price-value ratio can be improved by offering bulk online-only deals to make mobile grocery shopping more attractive for the elderly. Delivery charges can be lowered for elderly shoppers. In addition, South African grocery retailers can designate a day of the week for free delivery for elderly shoppers.

Given the age and gender differences observed, it is vital for grocery retailers in South Africa to recognise in their marketing efforts that the elderly market segment is heterogeneous. Grocery retailers should recognise that older respondents show a stronger positive relationship between effort expectancy and behavioural intention. Tailored displays can be positioned near the entrances of stores and at check-out counters to educate ‘older’ elderly shoppers about the ease of use of mobile grocery shopping services. The individuals shown in these displays should represent different elderly people of different ages. Testimonials of ‘younger’ elderly shoppers can be displayed in-store to market the hedonic aspects of mobile grocery shopping. The habit of using mobile grocery shopping services strongly influenced the intention of elderly male grocery shoppers only. Grocery retailers can partner with other mobile services to market mobile grocery shopping services to elderly male shoppers. The spill-over effect of the positive perceptions of the partner’s mobile service to the grocery retailer’s mobile shopping service would enhance the success of integrated strategies such as advertisements of the grocery retailer displayed within the partner’s mobile service or special promotions attracting elderly male shoppers to trial the mobile grocery shopping service.

A limitation of the study is that the research was conducted among the elderly in a single country. In addition, a cross-sectional research design was used. Replication of the study in different emerging market countries is vital to present evidence of the robustness of the findings of this study. Replication will also lead to researchers conducting a meta-analytic structural equation modelling (MASEM) study in the future to confirm the robustness of the results in this study. In such a MASEM, the moderating role of gender and age can be retested on a more significant sample of elderly grocery shoppers. The MASEM study can also explore national culture as a moderator, providing valuable insights for grocery retailers in emerging markets.

The sample represents elderly individuals who live in retirement complexes in two main cities. Therefore, it is vital that researchers in the future replicate the study in low-income groups that have access to this service and elderly consumers living in other types of establishments to determine the robustness of the findings.

The influence of facilitating conditions on the elderly’s intention to use mobile grocery shopping services was not investigated in this study based on the findings by Tan and Ooi (2018). Future research can include facilitating conditions to confirm whether it might be a factor influencing elderly consumers’ intention to use mobile grocery shopping services.

Future research can also extend the UTAUT2 psychological or emotional factors. For example, Huang (2023) recommends that anxiety, utilitarian value and trust are also important predictors of the elderly’s mobile shopping intention. Like grocery retailers in developed markets, grocery retailers in emerging markets are also concerned about retaining elderly grocery shoppers who are already using their mobile grocery shopping service. Thus, future research could examine the continuous use of mobile grocery shopping by the elderly by extending the expectation-confirmation model proposed by Bhattacherjee (2001) with technology readiness variables or by researching cross-channel synergies and dissynergies.

The findings of this research hold significant value for academics and retailers alike. By shedding light on the factors influencing the elderly’s intention to use mobile grocery shopping services in an emerging market country, this study provides valuable insights that can inform future academic research. The study’s results present vital information for academics interested in researching the elderly’s use of mobile services and new technological innovations. Furthermore, the results contribute to filling the gap in the elderly’s use of mobile grocery shopping services. At the same time, the study provides much-needed insights for grocery retailers to enhance the adoption of mobile shops by the elderly, a growing segment in emerging markets and worldwide. For retailers, the managerial recommendations should prompt them to re-evaluate their marketing of mobile grocery shopping services to align with the perceptions of the elderly. In conclusion, grocery retailers should show through their marketing of mobile grocery shopping services that the elderly are just as important a market segment as younger shoppers.