Flying to the customer: a meta-analysis of consumer adoption of drone delivery

Drone delivery has been increasingly prominent on industrial and political agendas across the globe, driven by economic incentives to make processes faster, more flexible and dependable, as well as being a “green” transport option (Benarbia and Kyamakya, 2021). Drones, or unmanned aerial vehicles (UAVs), are autonomous flying machines of different sizes and configurations based on the use case. Given their low cost, time-saving nature, environmental friendliness and minimum human involvement, drones are often considered the best option for parcel delivery (Kaplan and Heaslip, 2024). Companies like Amazon, Swiggy, Zomato and Rakuten are currently leading the drone package delivery market (The Business Research Company, 2024). Amazon Prime Air was one of the first companies to offer drone delivery services (DDS), taking less than 30 min to deliver small packages to clients (Wilke, 2019). The DDS market is projected to grow at a compound annual growth rate of 20.33% between 2024 and 2029 and is forecasted to reach US$1.75bn by 2029 (Mordor Intelligence, 2024). Osakwe et al. (2022) predicted that by 2026, more than a million drones would be used for deliveries related to retail. Hence, there is a growing interest among researchers in this field.

Many studies have investigated the broader commercial applications of drones, such as their use in cargo transportation and business logistics (Fang et al., 2024; Li et al., 2024). On the other hand, some have precisely focused on their role in food delivery services (Leong and Koay, 2023). While most customers are in favour of DDS and would even pay extra for it, some are cautious about executing this innovation in a retail situation. Despite the increasing interest in DDS for business applications, its utilization still remains relatively low (Leon et al., 2024). People’s concerns about security and privacy regarding DDS usage are high, as this technology is still relatively novel and is being adopted gradually (Leon et al., 2023; Osakwe et al., 2022).

Scholars have postulated multiple variables influencing DDS adoption. These variables range from complexity to environmental concerns and innovation resistance to risk perception. Empirical studies on the intention to use DDS have yielded mixed results. For example, Schmidt and Saraceni (2024) discovered a negative relationship between perceived risk and behavioural intention, while Fang et al. (2024) reported no significant relationship between these two variables. Additionally, Edwards et al. (2024) suggested that perceived usefulness has no significant effect on intention, contrasting with Koh et al. (2023), who found a strong positive relationship.

These varied results have led to confusion among scholars, professionals and researchers, not helped by differences in cultural context. However, mixed results are typical in the social and behavioural sciences; individual research studies are frequently limited by their specific contexts. Meta-analysis serves to uncover a common truth by statistically aggregating conceptually similar studies (King and He, 2005). Even though several studies have examined the connection between DDS adoption and a variety of factors, there is a dearth of available meta-analyses on this topic (Kaplan and Heaslip, 2024). Additionally, numerous prior studies have examined DDS solely within a single country, limiting the ability to draw broad conclusions; according to Chen et al. (2024), significant cultural differences exist in the acceptance of DDS. Managers should develop distinct ways to enhance the intention to adopt DDS across dissimilar cultures, since different components function differently in different cultures.

Existing studies suggest that uncertainty-avoidance and masculinity are important dimensions of culture (Hofstede, 2001) in the context of technology adoption. Srite and Karahanna (2006) performed a multiple-country study and discovered that these dimensions significantly affect the likelihood of adopting technology. This finding has been replicated in many studies investigating the influence of cultural differences on technology adoption (Kumar et al., 2023). Therefore, the present study considers these dimensions to assess the relevance of culture to the DDS adoption (Chen et al., 2024).

Based on the above discussion, this article has two objectives. First, we will conduct a comprehensive meta-analysis of studies on DDS adoption intention published up to April 2025. The meta-analysis will focus on articles that identify significant variables affecting DDS adoption intention. Second, we will investigate the moderating influence of culture on these variables. By analysing prior studies, the present meta-analysis contributes to existing knowledge by identifying previously studied antecedents of DDS adoption intention. Additionally, analysing the moderating effect of culture deepens our understanding of which factors hold greater significance within a particular cultural context.

Prior research on DDS adoption intention has drawn from multiple theoretical frameworks. These include the Technology Acceptance Model (TAM; Davis, 1989); the Theory of Planned Behaviour (TPB; Ajzen, 1991); the Unified Theory of Acceptance and Use of Technology (UTAUT; Venkatesh et al., 2003); and perceived risk theory (Featherman and Pavlou, 2003). Our meta-analysis organizes the antecedents identified in previous studies into four theoretically grounded categories: motivational factors, individual characteristics, socio-environmental concerns and credibility factors, each reflecting a coherent stream of theoretical reasoning. A single theoretical framework, such as TPB, might provide constructs from different categories, but they are categorised here according to their functional role. For example, perceived behavioural control and subjective norm (socio-environmental concerns) both originate in TPB, but the former is characterised here under individual characteristics, while the latter is categorised under socio-economic concerns.

Furthermore, the selection of antecedents is based on their consistent identification in prior studies as key determinants of DDS adoption. This approach helps to provide a more comprehensive understanding of the adoption of DDS as it captures the multidimensional nature of consumer decision-making. Thus, this study is intentionally integrative instead of theory-specific.

Motivational factors capture TAM and its extension-based beliefs related to perceived usefulness and hedonic value (Van der Heijden, 2004). Individual characteristics reflect TPB constructs, such as attitude and perceived behavioural control, and are extended with personal innovativeness from the diffusion of innovations theory (Agarwal and Prasad, 1998). Socio-environmental concerns incorporate subjective norms from TPB alongside environmental motivations from the green behaviour literature. Finally, credibility factors encompass trust and risk-related constructs derived from perceived risk theory (Featherman and Pavlou, 2003). Our framework offers a structured explanation of DDS adoption research that has used multiple theoretical frameworks. The external drivers (motivational factors and socio-environmental concerns) and internal drivers (individual characteristics and credibility factors) together shape the behavioural intention. Thus, the framework is based on established adoption mechanisms rather than just an aggregation of constructs.

Fang et al. (2024) and Schmidt and Saraceni (2024) noted inconsistencies in prior studies’ findings on perceived risk. Similarly, Edwards et al. (2024) and Koh et al. (2023) noted disagreements on perceived usefulness. These inconsistencies stem from multiple sources beyond cultural differences, including variations in technology maturity, deployment contexts (e.g. urban versus rural settings) and regulatory environments. The post-pandemic normalization of contactless delivery further complicates the situation (Kim et al., 2021; Yaprak et al., 2021). Additionally, drone-specific concerns, such as privacy intrusion and surveillance risks, introduce unique dynamics that are not present in conventional delivery technologies (Alhothali et al., 2024; Koh et al., 2023).

Nevertheless, these inconsistencies can be the outcome of a broader user perception regarding emerging technologies that we are not including in the present meta-analysis. For instance, concerns related to the visibility of technology, perceived intrusiveness and automation anxiety may affect users’ evaluation of DDS (Ali et al., 2021; Öztürk et al., 2024). Further, factors such as trust can play an important role in shaping user acceptance of artificial intelligence and autonomous systems (Lukyanenko et al., 2022).

Perceived usefulness indicates how much a person thinks DDS would help them (Davis, 1989; Liu et al., 2020). Consumers will be more receptive to such delivery methods if delivery drones are able to meet their last-mile delivery needs (Edwards et al., 2024). Furthermore, not all consumers are conversant with DDS, because these services are a relatively recent development. However, prior studies have found a positive relationship between perceived usefulness and intention to adopt DDS (Fang et al., 2024; Schmidt and Saraceni, 2024). Hence, the following hypothesis is proposed:

Perceived ease of use refers to how easy consumers perceive a system to be to use. Thus, it is the degree to which a person believes that using a particular system requires effort (Van der Heijden, 2004). Perceived ease of use is vital during the pre-implementation phase (Venkatesh et al., 2003), and research shows that it affects the intention to adopt DDS (Jasim et al., 2022; Liu et al., 2020). Therefore, we propose:

People may choose to use a technology that is fun and enjoyable (Hollebeek et al., 2019); consumers seek instant gratification, delight and imagination (Arnold and Reynolds, 2003). Consumers with a strong hedonic motivation might participate in contactless delivery by DDS (Jasim et al., 2022). Prior researchers have also found enjoyment as a driver for the adoption of DDS (Koay and Leong, 2023; Mathew et al., 2021):

Attitude refers to how people feel about carrying out a specific behaviour, whether that feeling is positive or negative (Fishbein and Ajzen, 1977). Based on TPB, consumers’ attitude toward technology directly impacts their adoption intention (Ajzen, 1991; Li et al., 2024). For example, a consumer might use DDS for picking up a food parcel when their attitude towards it is positive (Yaprak et al., 2021). Scholars have postulated that there is a positive relationship between attitude and DDS adoption (Edwards et al., 2024; Kumar et al., 2024). Hence, we posit that:

Personal innovativeness in the context of technology is the willingness to try new information systems (Agarwal and Prasad, 1998). Innovativeness affects the individual’s intention to adopt innovations, especially in the context of goods or services that have never been used before, such as DDS (Chen et al., 2022). Previous research studies have found a positive relationship between personal innovativeness and DDS adoption intention (Fang et al., 2024; Mittendorf et al., 2017). Therefore, we propose the following hypothesis:

Perceived behavioural control is defined as the ability of a person to perform a specific behaviour (Ajzen, 1991). It influences behavioural intention and can lead to changes in behaviour (Kumar et al., 2024). It depicts the ease or difficulty an individual perceives in executing the behaviour (Jasim et al., 2022). Perceived behavioural control demonstrates the respondents’ perception of their ability to use DDS. Studies have discovered that perceived behavioural control influences DDS adoption (Jasim et al., 2022; Li et al., 2024). Therefore, we hypothesize:

Previous researchers have found that friends or family members influence one’s actions (Fishbein and Ajzen, 1977; Schaarschmidt et al., 2021). Subjective norms have been found to reduce individuals’ resistance to using technology and enhance its adoption (Kim and Hwang, 2020). There is a high likelihood of using innovative and new logistics technologies such as DDS if there is social support for their adoption (Cai et al., 2021). Earlier studies have postulated a positive relationship between subjective norms and DDS adoption (Jasim et al., 2022; Liu et al., 2020). Hence, we propose the following hypothesis:

Environmental concerns reflect an individual’s interest in dealing with environmental problems (Han et al., 2022). Individuals who care about the environment are likely to take action and act sensibly towards ecological challenges (Albayrak et al., 2013). People who recognize environmental challenges are more likely to adopt DDS, as it provides a sustainable alternative to traditional delivery methods by reducing fossil fuel consumption and lowering carbon emissions. Existing studies have also found that environmental concerns lead to DDS adoption (Mathew et al., 2021; Mittendorf et al., 2017):

Trust refers to beliefs in common ideas, goals, morality and the trustee’s integrity (Leon et al., 2023). Individuals develop trust through personal interactions with service personnel, such as truck drivers, in the case of traditional delivery methods. However, the same is not true with drones (Ramadan et al., 2017). The lack of face-to-face interaction makes trust in the DDS and the service provider even more important. Previous researchers have also found trust to be an important antecedent for DDS adoption (Jasim et al., 2022; Osakwe et al., 2022). Hence, we hypothesize:

Performance risk is the possibility that a product will not work as expected and will not provide the necessary benefits (Featherman and Pavlou, 2003). Drones are expected to improve delivery speed and efficiency (Schmidt and Saraceni, 2024). On the other hand, Ravich (2015) argued that drones can collide and fall, which can pose a danger to pedestrians. Such likelihoods increase doubts about the performance of drones. Individuals worry about such risks, which in turn hinders DDS adoption. Studies have also shown a negative relationship between performance risks and DDS adoption intention (Liu et al., 2020; Yoo et al., 2018). Hence, we propose the following hypothesis:

Privacy risk refers to the leakage of confidential information without the permission of the consumers (Featherman and Pavlou, 2003). Zhou (2011) found that privacy risk reduces the adoption of technological innovation. Unauthorized collection of data and images during delivery by drone is the main concern apart from the issues related to safety and security (Koh et al., 2023). Individuals also fear that drones might encroach unlawfully and damage their property (Alhothali et al., 2024). Khan et al. (2019) articulated that customers who worry less about privacy develop positive views about DDS. Previous studies have found that if DDS poses less privacy risk, individuals will be more likely to adopt it (Fang et al., 2024; Leong and Koay, 2023). As a result, the following hypothesis is presented:

Most prior studies on DDS have been conducted in a single-country context, ignoring the cultural nuances. Only a few studies have explored cross-cultural DDS variations and suggested further cross-cultural investigations (Chen et al., 2024). Among Hofstede’s cultural dimensions, uncertainty-avoidance and masculinity–femininity are particularly pertinent (Hofstede, 2001). Furthermore, in the past, scholars have also used these two dimensions as moderators in technology adoption and have derived valuable implications (Goyal et al., 2025; Leonidou et al., 2013). In high-uncertainty-avoidance cultures, individuals may exhibit heightened sensitivity to perceived risks and technological ambiguities, which may diminish adoption intentions (Featherman and Pavlou, 2003; Hofstede, 2001). Similarly, masculine cultures prioritize achievement, performance and efficiency, thereby amplifying the influence of instrumental beliefs such as usefulness and convenience on adoption intentions (Hofstede, 2001). Other cultural dimensions like power distance or individualism-collectivism are relevant in broader technology adoption contexts. However, they exhibit less direct interaction with the core antecedents of DDS adoption and offer limited explanatory value in this specific context.

In high-uncertainty-avoidance cultures, individuals give more importance to the quality of technological services, such as reliability, responsiveness and assurance (Ladhari et al., 2011). Researchers have also underscored that people in such cultures strongly emphasize product quality (usefulness and ease of use) to deal with uncertain situations (Hung and Chou, 2014). Additionally, individuals from high-uncertainty-avoidance cultures are more cautious (Hofstede, 2001) and seek a sense of control during the adoption process to minimize uncertainty. Hence, perceived behavioural control plays a key role in determining their intentions to adopt technologies such as DDS.

Individuals often feel reassured by listening to the experiences of other people (Hwang and Lee, 2012). Thus, subjective norms are important for the acceptance of new technologies. Lai et al. (2016) also found that the impact of social influence or subjective norms on behavioural intentions to use new technologies was stronger in high-uncertainty-avoidance cultures. Similarly, security and privacy concerns are more important in high-uncertainty-avoidance cultures (Cyr, 2013). In high-uncertainty-avoidance cultures, there is a high openness to innovation and readiness to adopt new technologies (Steenkamp et al., 1999). Hence, personal innovativeness becomes a key factor in DDS adoption within low-uncertainty-avoidance cultures. Thus, we propose:

The masculinity dimension captures the degree to which a culture emphasizes traditionally masculine values, such as achievement, assertiveness and material success. It contrasts with feminine values, such as cooperation, modesty, care for the weak and quality of life (Hofstede, 2001). This cultural orientation shapes individuals’ evaluation and response to novel technologies such as drones. For instance, individuals in feminine cultures may place greater value on emotional gratification, ease of use, environmental benefits and societal consensus, aligning with constructs like perceived enjoyment, perceived ease of use and subjective norms. Conversely, individuals in masculine cultures may prioritize instrumental benefits, such as perceived usefulness and personal achievement and may exhibit greater scepticism toward risks that may threaten performance outcomes. Prior meta-analyses have found that the masculinity dimension significantly moderates consumer behaviour and technology adoption intentions (Goyal et al., 2025; Kumar et al., 2023). Thus, we propose:

Figure 1 represents the conceptual model for the present study.

The study followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines to ensure a transparent and systematic selection of studies. The identification, screening, eligibility and inclusion stages were conducted in accordance with the PRISMA framework (Page et al., 2021). A PRISMA flow diagram is provided to illustrate the study selection process in a structured manner (Figure 2). In the first step, the research articles were thoroughly searched from reputable electronic databases, including Scopus, EBSCO and ProQuest. Apart from databases, conference proceedings from other well-known conferences were also taken into consideration, e.g. the Association for Information Systems conferences. The following search string was used to extract relevant publications: (Drone OR UAV OR “Unmanned Aerial Vehicle”) AND (Delivery OR Logistic* OR Parcel) AND (Adoption OR Acceptance OR Consumer OR Customer OR Intention). We specifically selected publications that contained these search terms within the title, keywords or abstract. The present meta-analysis considered studies published till April 2025. The first database search found 1,571 publications.

In the second step, the following criteria were used for shortlisting the publications for the present meta-analysis:

• The publication should have evaluated at least one relationship of our conceptual model quantitatively;

• The publications should contain the correlation value or another measure (such as F ratio or Student’s t or beta value) which could be converted to the correlation value;

• The publications must contain the sample size; and

• The publications should be in English.

In the third step, the title, keywords and abstract of the extracted publications were read by two researchers independently to remove seemingly irrelevant and duplicate publications. At the end of this step, 201 publications were deemed suitable for further analysis. The inter-rater reliability during this selection phase was outstanding, with a 95% agreement rate and a Cohen’s kappa of 0.86. Any disagreements were resolved through discussions between the two researchers.

Next, both researchers read the full text of the 201 extracted publications to remove publications that did not meet the criteria mentioned in the second step. By the conclusion of this step, 49 publications were considered appropriate for meta-analysis.

In the fifth step, five more publications met the inclusion criteria after doing forward and backward searches. Finally, 54 publications and 58 studies were selected for the analysis (four publications had two sub-studies with different samples).

We recorded each publication’s title, author names, journal, sample size, country name and correlation values in a spreadsheet. In our meta-analysis, we used correlation coefficients as effect sizes. When studies did not provide correlation coefficients, we converted other statistical measures, such as regression coefficients, into correlations based on the method outlined by Peterson and Brown (2005). Specifically, we applied their formula: r = β + 0.05λ, where λ = 1 if β is positive and λ = 0 if β is negative, to estimate the correlations.

Following prior meta-analytic practices (Baptista and Oliveira, 2016), conceptually equivalent constructs were merged to ensure theoretical consistency across studies. Specifically, perceived usefulness (TAM) and performance expectancy (UTAUT) were treated as a single construct, as they capture the instrumental belief that a technology enhances task performance. Similarly, perceived ease of use (TAM) and effort expectancy (UTAUT) were combined, as both reflect the degree of effort associated with using a system (Venkatesh et al., 2003). These aggregations result from conceptual similarities across theoretical models rather than only from statistical considerations.

Only constructs with at least 3 prior-study correlation values available were included as antecedents, ensuring statistical robustness and reliability in line with meta-analytic guidelines (Borenstein et al., 2011). Furthermore, we recorded the uncertainty-avoidance and masculinity scores for each country as indicated by Hofstede (2001) and classified the countries into high vs low cultural groups by using the median split method to facilitate subgroup comparisons across cultural contexts, as done in prior studies (Gupta et al., 2025; Kirca et al., 2005; Kumar et al., 2023). Although this approach may reduce variance, it has been widely adopted in cross-cultural meta-analytic studies using country-level cultural indices. Furthermore, a limited number of effect sizes per relationship restricts the feasibility of continuous moderation analysis.

The sample of this meta-analysis consists of 54 publications and 58 studies. These studies were performed in several countries, including China, Malaysia, the USA, India and South Korea. It thus reflects a wide geographical distribution. There is considerable variation in sample size among the studies reported, ranging from 83 to 1108; this indicates heterogeneity across contexts and populations. Along with key constructs examined in prior literature, our data set also captures variation in cultural dimensions (uncertainty-avoidance and masculinity). The sample thus supported the examination of cultural moderation effects in the present meta-analysis. A detailed overview of the included studies and their characteristics is provided in Table S1 in the supplementary material. The references of these publications are also provided in the supplementary material.

Statistical analysis was conducted using the correlation value as the effect size measure. According to Cohen et al. (1983), a higher effect size value indicated more substantial presence of the phenomenon. The present meta-analysis followed the procedures outlined by Lipsey and Wilson (2001) and Hedges and Vevea (1998).

Table 1 shows the meta-analysis results for the direct relationships. The results show that all motivational factors – namely perceived usefulness, perceived ease of use and perceived enjoyment – have a significant and positive relationship with the DDS adoption intention. These results confirm support for H1, H2 and H3. Among these, perceived usefulness has the strongest impact on adoption intention, followed by perceived ease of use and perceived enjoyment. Additionally, all individual characteristics, attitude, personal innovativeness and perceived behavioural control were significantly and positively associated with DDS adoption intention, with attitude having the strongest effect, thereby supporting H4–H6. Socio-environmental concerns – namely subjective norm and environmental concerns – also demonstrate significant positive relationships, with subjective norm exerting a stronger influence, supporting H7 and H8. Finally, among credibility factors, trust has a positive relationship with DDS adoption intention. In contrast, performance risk does not have a significant effect, and privacy risk negatively impacts adoption intention, providing support for H9 and H11, while H10 is not supported.

The heterogeneity observed among the sample studies suggested the presence of a variable moderating the relationships. This meta-analysis investigated culture as a potential moderator, as presented in Tables 2 and 3. Relationships of personal innovativeness and trust with adoption intention could not be examined because of the limited number of studies on these antecedents.

The Q-between statistics revealed that uncertainty-avoidance significantly moderates the majority of the relationships (Table 2). In countries with a high-uncertainty-avoidance culture, adoption intention was shown to have a significantly stronger relationship with all of the following: perceived ease of use, perceived enjoyment, attitude, perceived behavioural control, subjective norms, environmental concerns and privacy risk. Conversely, uncertainty-avoidance did not significantly moderate the relationships between adoption intention and perceived usefulness or performance risk. For perceived usefulness, despite a stronger effect in high-uncertainty-avoidance contexts, the difference was not statistically significant. Similarly, performance risk was negatively related to adoption intention across both groups, but the moderation effect was not significant. Overall, the results partially support H12.

As presented in Table 3, the Q-between statistics indicate that culture type significantly moderates the influence of perceived usefulness, perceived ease of use, perceived enjoyment, attitude, perceived behavioural control, subjective norms and environmental concerns on DDS adoption. Conversely, no significant moderating effects were found for performance risk or privacy risk, as indicated by non-significant Q-between values. While privacy risk showed slightly stronger negative effects in feminine cultures, the differences were not statistically significant. Overall, most relationships were significantly moderated, partially supporting H13.

We also performed a publication bias test. Publishing bias is the phenomenon wherein a desire for statistically significant outcomes in study results influences a journal’s publishing decisions. For this goal, Rosenthal’s fail-safe N test was employed in this study (Rosenthal, 1979). The current meta-analysis shows no publication bias, according to the fail-safe N values (see Table 1).

The present meta-analysis identified four key categories of factors: motivational, individual characteristics, socio-environmental concerns and credibility factors. Motivational factors, namely perceived usefulness, perceived ease of use and perceived enjoyment, have a positive influence on DDS adoption intention. This demonstrates that individuals are more likely to adopt DDS when they perceive the service as useful, easy to use and enjoyable (Koh et al., 2023). Furthermore, individual traits positively influence DDS adoption. Users with a positive attitude toward innovation and an inclination to try new technologies will have a higher tendency to adopt DDS (Kim et al., 2021; Leong and Koay, 2023). This might be because, rather than a specific feeling, attitude indicates an overall evaluation or feeling with respect to a technology. When a positive perception regarding DDS is generated among people, they develop a stronger intention to use it, as the existing literature on technology adoption suggests (Ajzen, 1991; Li et al., 2024).

Also, the significance of perceived behavioural control suggests that consumers’ belief in their ability to use the technology affects their DDS adoption intentions (Cai et al., 2021). With respect to socio-environmental concerns, the significant relationship with subjective norm highlights the role of social influences in driving DDS adoption intention. Such influences include recommendations from peers, family or social media influencers (Jasim et al., 2022). Also, the relationship between environmental concerns and DDS adoption intention suggests that sustainability is a meaningful driver for consumers. This implies the growing importance of eco-friendly practices in the decision-making process (Mathew et al., 2021). Also, as anticipated, trust and privacy risks serve as essential credibility factors influencing the adoption of DDS (Jasim et al., 2022). This implies that individuals are inclined to evaluate the security and reliability of such systems before committing to DDS use. Trust is pivotal in building user confidence, and privacy concerns affect their willingness to share personal information (Leon et al., 2023). Privacy concerns make people feel that they will lose control of their sensitive information. Such concerns are personal, immediate and difficult to ignore, which makes people uncomfortable about sharing their data or being monitored (Featherman and Pavlou, 2003).

Interestingly, findings reveal that performance risk does not impact the intention to adopt DDS (Fang et al., 2024), indicating that concerns about DDS’s efficiency and effectiveness do not notably deter adoption intentions. This implies that users are generally satisfied with DDS’s basic performance and are instead more focused on risks associated with trust, security and personal data (Osakwe et al., 2022). One possible reason for this might be that users expect the service provider to handle performance-related issues like delays or failures. Thus, performance-related concerns may not strongly influence users’ adoption of DDS (Featherman and Pavlou, 2003).

The present study also investigated the moderating influence of culture on the adoption of DDS. The findings revealed a stronger influence in high-uncertainty-avoidance cultures for the following constructs: perceived ease of use, perceived enjoyment, attitude, perceived behavioural control, subjective norms, environmental concerns and privacy risk (Alhothali et al., 2024; Hwang et al., 2021; Li et al., 2024). This indicates that individuals in such cultures emphasize ease, enjoyment, attitudinal beliefs and contextual reinforcements to mitigate the inherent uncertainty associated with adopting emerging technologies like DDS. The stronger negative effect of privacy risk in high-uncertainty-avoidance cultures indicates that people in such cultures are more sensitive to this risk, and hence they place greater emphasis on privacy-related concerns. On the other hand, the relationship becomes negative and insignificant in low-uncertainty-avoidance cultures, indicating that privacy concerns play a less influential role in shaping adoption decisions.

In addition, the study also examined the moderating influence of the masculinity–femininity cultural dimension. The findings suggest that, compared with masculine cultures, individuals in feminine cultures are more strongly influenced by factors such as perceived usefulness, perceived ease of use, perceived enjoyment, attitude, perceived behavioural control, subjective norms and environmental concerns when forming intentions to adopt DDS. Furthermore, the consistent negative effect of privacy risk across masculine as well as feminine cultures suggests that privacy concerns are universal and less dependent on cultural differences related to masculinity or femininity.

Interestingly, performance risk did not show statistically significant differences across either dimension of culture. This indicates that performance risk perceptions are more universally shared, regardless of cultural orientation. The consistent insignificance of performance risk across both cultural contexts suggests that these concerns are not strongly shaped by the differences in culture. These concerns reflect an overall tendency of people to rely on service providers for DDS performance.

The present meta-analysis brings out several theoretical contributions. First, it is one of the first comprehensive studies that investigates the factors influencing the intention to adopt DDS by synthesizing the outcomes of prior quantitative studies and employing meta-analysis. These were further grouped into four broad categories: motivational factors, individual characteristics, socio-environmental concerns and credibility factors. Second, a review of existing DDS studies revealed that the empirical findings were fragmented, with varying effect sizes and significance levels. This meta-analysis advances the DDS literature by consolidating the strengths and significance of the relationships between key antecedents and DDS adoption intention.

Third, most prior DDS research has primarily focused on a limited set of antecedents based on one or two established theories, such as TAM, TPB and UTAUT (Ajzen, 1991; Venkatesh et al., 2003). This study, however, broadens the scope by incorporating multiple constructs from different theoretical frameworks. Finally, there is a scarcity of evidence on the moderating role of culture in DDS adoption studies. This study makes a significant contribution by offering valuable insights into the moderating impact of Hofstede’s dimensions on DDS adoption intention.

The outcomes of the present meta-analysis provide the following key implications for organizations. First, organizations should emphasize the functional benefits of DDS, such as improved efficiency, faster delivery times and convenience. The industry should focus more on the design and user experience of DDS platforms; the platforms should be user-friendly and designed to minimize complexity. Organizations should enhance the consumer experience by incorporating elements of enjoyment and excitement into the DDS experience through features such as real-time tracking. They could also devise unique delivery presentations that appeal to consumers’ hedonic motivations.

Second, the relationships of attitude, personal innovativeness and perceived behavioural control with DDS adoption intention also provide useful implications. Organizations can recognize and segment early adopters who are tech-savvy and naturally drawn to novel technologies to fast-track the adoption of DDS in this group. Organizations can build positive attitudes and boost perceived control through educational content, clear instructions, pilot projects and demonstrations.

Third, the awareness of green delivery options may resonate particularly well with consumers’ values and encourage them to adopt the technology. Hence, companies should inform consumers about the sustainability benefits of DDS, such as reduced carbon footprint and minimized traffic congestion. Furthermore, subjective norms can be built by success stories of adoption in similar communities or industries, which can develop social pressure and peer influence. This can be enhanced through influencer marketing, social campaigns and community-based initiatives.

Fourth, the privacy risk and trust relationships with DDS adoption intention also provide valuable implications for practitioners. Clear communication about the reliability of drones, safety measures and privacy protection policies may address lingering doubts by building confidence in the safety of personal information. Moreover, if an organization is involved in community outreach, shares the success stories of customers and partners with trustworthy firms, it can further enhance trust.

Fifth, the moderator analysis suggests that managers should adapt strategies based on cultural contexts rather than adopting a uniform approach. In high-uncertainty-avoidance cultures, most determinants, such as ease of use, enjoyment, attitude, perceived behavioural control, subjective norms, environmental concern and privacy risk, have strong effects on adoption intention. Thus, firms should focus on reducing uncertainty by providing clear information, detailed guidance and reliable service communication. At the same time, they should highlight the steps taken to ensure data privacy. In contrast, in low-uncertainty-avoidance cultures, firms may adopt a relatively balanced communication strategy through which they focus on general value propositions for DDS.

Finally, the effects of factors such as perceived usefulness, perceived ease of use, perceived enjoyment, attitude, perceived behavioural control, subjective norm and environmental concerns are stronger in feminine cultures. Hence, where a feminine culture is predominant, firms should emphasize functional benefits, ease of use, user experience and environmental value. Firms can benefit by getting customer testimonials, peer recommendations and influencer endorsements. Additionally, communications regarding the sustainability benefits of DDS can be effective in these contexts. However, in masculine cultures, firms should highlight core functional benefits rather than multiple persuasive aspects of DDS. Overall, the findings highlight the need for culturally sensitive DDS strategies across cultural contexts. Table 4 summarizes the conclusions and their implications.

The present study has several limitations that future researchers should address. First, this study identified 11 constructs as key determinants for DDS adoption intention. Future studies should consider additional factors, e.g. other dimensions of risk such as safety concerns (related to accidents and drone malfunction) and legal risks (e.g. liability in case of mishaps). Second, while the study analysed uncertainty-avoidance and masculinity as cultural dimensions, future researchers should incorporate Hofstede’s other cultural dimensions (individualism, power distance and long-term orientation) for a deeper examination of cultural moderating effects. Third, integrating additional country-specific and firm-specific moderators, such as the Human Development Index, Global Innovation Index, industry sector and company size, could provide deeper insights into the factors influencing DDS adoption intention.

Fourth, future researchers can include additional factors related to emerging technologies in their models to study their effect on DDS adoption. These factors can include technological intrusiveness, automation-related concerns and trust in artificial intelligence systems. Fifth, with the growing interest in DDS adoption and the challenges associated with it, future researchers should develop empirical studies or comprehensive literature reviews to better understand the dynamics at play. The use of median split to operationalize cultural dimensions may reduce variability and assume homogeneity within groups, potentially obscuring nuanced cross-country differences. Future research may employ continuous moderation techniques or multi-level approaches to capture more fine-grained cultural effects.

Sixth, future research should move beyond simply adding new antecedents and instead focus on advancing stronger theoretical perspectives. For instance, future studies can extend current adoption-centric models by incorporating post-adoption frameworks, such as expectation-confirmation theory, to examine continued use and loyalty toward DDS. In addition, researchers may draw on institutional theory to better understand how regulatory environments, infrastructure readiness and ecosystem maturity shape consumer acceptance, thereby shifting attention from individual-level factors to broader contextual influences.

No funding was received for this research.

The supplementary material for this article can be found online.