Investigating Antecedents of Intention to Use Green Agri-Food Delivery Apps: Merging TPB with Trust and Electronic Word of Mouth

Abstract

The rapid expansion of digital platforms has significantly influenced consumer purchasing behaviors, particularly in the agri-food sector. Therefore, this paper investigates the key factors driving customers’ intention to use green agri-food delivery apps (GAFDAs) by integrating trust and electronic word of mouth (eWOM) into the Theory of Planned Behavior (TPB) framework. Additionally, this study examines gender as a moderating variable, assessing whether its influence alters the relationships between key determinants and behavioral intention. Data were collected from 252 Algerian consumers, and the proposed model was tested using SmartPLS 4 and SPSS 26.0. The results confirm that attitude, subjective norms, perceived behavioral control (PBC), trust, and eWOM positively and significantly influence the intention to use GAFDAs, with PBC emerging as the strongest predictor. Moreover, gender moderates the effect of trust on behavioral intention, with trust significantly influencing men’s adoption decisions but not those of females. In contrast, subjective norms and PBC are stronger predictors for female consumers. These findings highlight the importance of gender-specific marketing strategies to enhance GAFDA adoption. This study contributes to the literature by extending TPB with trust, eWOM, and gender moderation, offering valuable insights for marketers, policymakers, and app developers promoting sustainable food consumption.

1. Introduction

The COVID-19 pandemic has led to the widespread adoption of food delivery apps among customers [1,2,3]. These apps are “mobile applications that allow users to order and receive food offline” ([4], p. 3). While online food delivery services offer numerous benefits, such as job creation and sales opportunities, they are not without criticism. Negative impacts include (1) societal effects, such as public health concerns and increased traffic; (2) environmental issues, such as the generation of more food and plastic waste; and (3) economic challenges, such as higher commission fees. These factors have led to boycotts by some customers and restaurants [1]. From a health perspective, consuming unhealthy food (e.g., French fries) and adopting poor eating habits (e.g., frequent fastfood consumption) not only contribute to weight gain and obesity but also increase the risk of chronic diseases like type 2 diabetes [5]. Okumus and Bilgihan [6] (p. 32) argued that food ordering apps could serve as “tools to promote healthy eating behaviors”. In practice, however, these apps are often used to promote both healthy and unhealthy food options. Therefore, these apps can foster healthy behaviors only when they focus on green agri-food or organic products. Encouraging the consumption of sustainable foods, such as green agri-foods, through green food ordering apps is an effective solution to mitigate the negative impacts of traditional food delivery systems.

Generally, any food item produced using eco-friendly agricultural practices is considered a green agri-food product. This includes items such as oils, dairy, meat, vegetables, and fruits. In the field of green agri-food production, Wang and Fan [7] (p. 3) emphasize that “environmental protection and product quality are prioritized, with a focus on pure naturalness and a ban on the use of harmful chemicals such as pesticides, fertilizers, and herbicides”. As a result, green agri-foods are perceived by customers as organic, environmentally safe, healthy, and pollution-free [7,8,9]. The adoption of digital technologies in agri-food supply chains is also expected to contribute to food waste reduction, improve food security, and enhance sustainability [10]. It has been shown to positively impact the performance of agri-food supply chains [11,12] and improve the performance of agri-businesses [13].

From an environmental perspective, Appiah et al. [14] found that “green supply chain practices” significantly influence environmental performance. In practice, the online delivery of green agri-food involves four key parties: (1) green agri-food suppliers (including distributors and grocery stores); (2) green agri-food ordering platforms that process orders and manages delivery; (3) delivery personnel responsible for transporting orders from suppliers to customers; and (4) customers, both existing and potential, who constitute the target market. Hence, mobile apps offer significant opportunities to promote the use of green agri-food. A recent report by Future Data Stat (2023) indicates that the global sustainable food delivery market is expected to expand significantly, rising from USD 112.67 billion in 2023 to USD 449.67 billion by 2030, with a “compound annual growth rate” of 19.7% [15]. The report highlights that North America—particularly the United States and Canada—represents the leading region in terms of sustainable food delivery services. Meanwhile, substantial growth is also being observed in European countries such as the United Kingdom and Germany, as well as in Asian markets including Japan and South Korea. In contrast, markets in South America, the Middle East, and Africa are experiencing more gradual, yet steady, growth.

In Algeria, the government is actively promoting green entrepreneurship, particularly within the agricultural sector. The Ministry of Agriculture has prioritized the development of organic and environmentally sustainable agri-food products, aiming both to satisfy domestic demand and to enhance the competitiveness of Algerian products in international markets. In terms of digital adoption, empirical studies indicate that Algerian consumers demonstrate a willingness to adopt food delivery applications, suggesting a favorable environment for the integration of mobile technologies in the agri-food sector [16]. Cui et al. [17] (p. 1) note that “consumers’ concerns about their personal health and food safety have stimulated the demand for green agri-food”. Furthermore, increasing awareness of green consumption has played a significant role in driving customers’ preference for green agri-foods [7]. However, due to the novelty of online green agri-food marketing in Algeria, challenges remain, including a lack of experience, trust issues, limited customer awareness of the benefits of these products, and problems with online payment systems. For instance, Ma et al. [18] highlighted that marketing green agri-food products online in China faces challenges such as seasonality, quality assurance, and storage costs.

Over the past five years, food delivery apps have attracted the attention of scholars in both developed and developing countries. Research has focused on investigating the intention to adopt these apps [1,19,20,21,22] as well as continuance intention [2,4,23] and examining user satisfaction and loyalty [24,25,26]. In addition to the above, Ma et al. [18] observed that male customers exhibited a more consistent willingness and behavior in purchasing green agri-food products than their female counterparts. Indeed, while women may express stronger pro-environmental attitudes, their purchasing behaviors may be influenced by additional factors such as price sensitivity, perceived convenience, or product availability. Francioni et al. [23] further investigated the moderating role of gender in behavioral intentions and found that the determinants of male consumers’ decision-making processes differed from those of female consumers. These variations highlight the importance of considering gender dynamics when examining the adoption of digital service. However, existing studies confirm that research on the intention to adopt food delivery apps is still in its early stages [27,28,29], and studies focusing on the intention to use green agri-food delivery apps (GAFDAs) are almost non-existent.

Theoretically, the “Theory of Planned Behavior” (TPB), developed by Ajzen [30], is one of the most widely used frameworks by scholars to predict human behavior in various fields and contexts [31,32,33,34]. TPB posits that behavior can be predicted by intentions, which, in turn, are influenced by attitudes, subjective norms, and “perceived behavioral control” (PBC) [30]. Behavioral intention is defined as a person’s willingness to voluntarily perform a specific behavior [35]. Several researchers have applied the TPB to investigate the intention to use food delivery apps [36,37]. The majority of these studies suggest that augmenting the original TPB model with additional constructs enhances its explanatory power regarding customer intentions. Researchers have integrated constructs such as the desire to use online food delivery and perceived risk [37], social isolation, perception of food safety, and food delivery hygiene [36], perceived security and word of mouth (WOM) [20], and perceived trust [38]. Additionally, some studies have combined “electronic word of mouth” (eWOM) with the TPB to predict tourists’ intentions [39,40]. Previous research has also highlighted that, beyond the core constructs of the TPB, several additional factors influence consumers’ intention to use food delivery apps. These include performance expectancy and congruity with self-image [27], price and social value [28], customer experience and ease of use [41], and effort expectancy [21,42]. Furthermore, Ma et al. [18] demonstrated that customer trust is a crucial factor in the successful online marketing of green agri-food products.

Practically speaking, with the widespread adoption of food ordering apps and the growing green agri-food market, stakeholders need to understand the key drivers of GAFDA adoption. Despite the increasing importance of healthy food, surprisingly few studies have investigated the intentions behind using green agri-food ordering apps. This paper seeks to fill this gap by applying an expanded TPB model, incorporating trust and eWOM, within the Algerian context. Therefore, this study aims to address three key research questions: (RQ1) What are the primary determinants of consumer intentions to use GAFDAs?; (RQ2) Do men and female differ in their predictors and intentions to use GAFDAs?; and (RQ3) Does gender moderate the influence of key determinants—such as attitudes, subjective norms, PBC, trust, and eWOM—on consumers’ intention to use GAFDAs? The results of this study are expected to provide valuable insights for stakeholders, particularly those involved in the green agri-food supply chain, green agri-food retailers, agri-food marketing managers, app developers, and policymakers focused on sustainability. These insights could help in the development of programs and strategies aligned with the “United Nations Sustainable Development Goals” (UNSDGs).

The remainder of this paper is structured as follows. Section 2 presents the theoretical background and hypotheses development, followed by Section 3, which details the methodology, including data collection and instrument development. Section 4 discusses the results, while Section 5 provides a detailed discussion of the findings in relation to prior research. Finally, Section 6 concludes with theoretical contributions, managerial implications, study limitations, and future research directions.

2. Theoretical Background and Hypotheses

2.1. Extended TPB in the Food Delivery App Context

The TPB has been widely recognized as a robust framework for predicting customer intentions and behaviors across various domains, including e-commerce, healthcare, and sustainable consumption. Ulker-Demirel and Ciftci [43] conducted a comprehensive literature review and confirmed the model’s broad applicability, emphasizing its utility in explaining consumer decision-making processes. However, scholars have also highlighted certain limitations of the TPB when applied to complex consumer behaviors. Mouloudj et al. [32] (p. 3) noted that “scientific evidence shows that the TPB model was successful in predicting intentions in several settings, but that evidence also demonstrated that incorporating additional constructs can be more useful in improving prediction accuracy”. This suggests that while TPB serves as a strong theoretical foundation, its explanatory power can be enhanced by integrating context-specific variables.

In the context of GAFDAs, consumer decision-making may go beyond the traditional constructs of the TPB. Since sustainable consumption decisions frequently involve elements such as trust and eWOM communication, it is essential to extend the TPB framework by incorporating these additional factors to more effectively explain behavioral intentions. Among these, trust has emerged as a critical determinant in the adoption of digital services [3,21,22]. Trust is particularly relevant in green food delivery services, where consumers may have concerns about the authenticity of sustainability claims, food quality, security (to avoid risks), and ethical sourcing [2,4,17,18,20,21,25]. While the existing literature has established that trust significantly influences online purchasing behavior [3], its role in the adoption of GAFDAs remains underexplored. Another essential construct shaping consumer behavior is eWOM. Prior research has demonstrated that online reviews and peer recommendations significantly affect consumer decision-making in digital marketplaces [6,20]. However, within the context of GAFDAs, the extent to which eWOM influences adoption intentions remains unclear.

On the other hand, some researchers have applied other conceptual frameworks to explain intentions to use food delivery apps, such as Protection Motivation Theory (PMT) [44], the “Unified Theory of Acceptance and Use of Technology” (UTAUT) [21,27,42], Gratification Theory [41], the Value–Belief–Norm (VBN) Theory [45], the Theory of Consumption Value [28,46], and the “Technology Acceptance Model” (TAM) [1,47,48,49]. Moreover, hybrid models have emerged as a promising approach to improving explanatory power. For instance, Troise et al. [22] combined the TPB with the TAM to investigate the intention to use food ordering apps. Similarly, Choe et al. [50] integrated the TPB and the TAM to explore the intention to use drone food delivery, highlighting the need for multi-theoretical frameworks in emerging technological contexts.

In the context of GAFDAs, these alternative models offer insights but also reveal certain limitations. PMT, for example, could help explain how perceived risks (e.g., food contamination or misleading sustainability claims) influence consumer intentions. While the UTAUT effectively captures digital service adoption, it does not inherently account for sustainability concerns, which are central to GAFDAs. Similarly, Gratification Theory emphasizes hedonic and utilitarian benefits such as convenience and enjoyment, but sustainability-related motivations may extend beyond mere gratification. The VBN theory provides a strong ethical foundation for sustainable consumption, yet it does not consider technological adoption barriers, making it less suitable as a standalone model for GAFDAs. Meanwhile, the TAM has been extensively used to explain technology adoption based on “perceived usefulness and perceived ease of use”. In the food delivery sector, it is frequently applied to examine consumer interactions with mobile apps and digital platforms. However, since the TAM primarily focuses on technological adoption without addressing behavioral intention determinants specific to sustainability, its predictive power may be enhanced by integrating TPB-based constructs. Additionally, the TCV highlights functional, emotional, and social values in shaping consumer choices. This perspective is relevant to GAFDAs, as consumers evaluate these services not only based on their environmental impact but also in terms of convenience, affordability, and personal satisfaction.

Based on these arguments, this paper extends the TPB model by incorporating trust and eWOM to improve its predictive ability regarding the willingness to use GAFDAs. This theoretical extension aligns with prior research advocating for the integration of external variables to enhance the predictive power of TPB, particularly in contexts involving digital platforms and sustainability considerations.

2.2. Hypothesis Development

2.2.1. Attitudes Toward GAFDAs

Attitude is defined as “the degree of a person’s positive or negative feelings about performing a target behavior” ([51], p. 984) or “the degree to which a person has a favorable or unfavorable evaluation or appraisal of the behavior in question” ([30], p. 188). In the context of green agri-food delivery, attitudes reflect a customer’s judgment of a particular app’s service as being preferred or non-preferred, beloved or unloved, favorable or unfavorable, and good or bad. Attitudes toward using food delivery apps are shaped by customers’ perceptions of their environmental impact, health benefits, and convenience. Studies have shown that individuals who perceive GAFDAs as sustainable and beneficial are more likely to develop positive attitudes toward their use [4,49]. Several empirical studies on digital apps have revealed the significant role of favorable attitudes in predicting the intention to use online food delivery apps [19,20,22,36,49,50], as well as the continuance to use food ordering apps [2,4,36] and mobile service apps [52] and the intention to use telemedicine apps [53]. Based on these findings, we propose the following hypothesis:

H1. 

Attitude positively influences the intention to use GAFDAs.

2.2.2. Subjective Norms

The second factor in the TPB model is subjective norms (or social influence), which are defined by Venkatesh et al. [54] (p. 450) as “the extent to which an individual perceives that important others believe he or she should apply the new system”. According to Ajzen [30] (p. 188), subjective norms are “the perceived social pressure to perform or not to perform the behavior”. In the context of green agri-food delivery, it is expected that people will seek input or advice from individuals in their reference groups (e.g., family, friends, colleagues, or professionals such as doctors and nutritionists) when considering ordering healthy food, especially if they lack sufficient information to make informed decisions. Çoker et al. [55] (p. 1) stated that “framing social norms around making healthy and eco-friendly food choices to refer to a close referent group may change their perceptions and ability to encourage sustainable and healthy food purchasing”. Many previous studies have reported that subjective norms significantly impact the intention to use food delivery apps [20,22,36,50], the desire to use online food delivery [37], and the intention to use health apps [53]. Social influence has been shown by Alalwan [56] and Zhao and Bacao [3] to have a significant association with the willingness to reuse food ordering apps. Aprile and Punzo [57] suggest that subjective norms have an indirect effect on intentions by influencing attitudes, PBC, and moral norms. Based on these findings, we propose the following hypothesis:

H2. 

Subjective norms positively influence the intention to use GAFDAs.

2.2.3. Perceived Behavioral Control (PBC)

PBC is a key construct introduced by Ajzen [30] in the TPB model, defined as “people’s perception of the ease or difficulty of performing the behavior of interest” (p. 183). In the context of green agri-food delivery, the presence of difficulties may lead to customers switching apps, postponing their orders, or even abandoning their orders if significant obstacles arise. Both ease of use and self-efficacy play important roles in shaping mobile app usage intentions [16,52]. Mobile apps are practical and easy-to-use resources that can help motivate individuals to adopt healthier food purchasing and eating habits [58]. Poon et al. [37] (p. 14) noted that issues such as “poor delivery service, food being stolen, wrong orders, leakage of personal information, financial disputes, potential physical harm, and the risk of contracting COVID-19 from the delivery person” can reduce customers’ willingness to use food ordering apps. Several empirical studies have found a positive association between PBC and the intention to use food ordering apps [36,37,50]. However, in a study by Yazdanpanah and Forouzani [59], the results did not support the effect of PBC on the intention to buy organic food. Nevertheless, in line with the TPB model, PBC is expected to influence the intention to use GAFDAs. Thus, we propose the following hypothesis:

H3. 

PBC positively influences the intention to use GAFDAs.

2.2.4. Trust in GAFDAs

In the era of digital marketing, issues related to confidentiality, privacy, and security have become major concerns for both marketers and customers. As a result, trust has emerged as a crucial factor in understanding customers’ adoption behavior toward modern mobile apps [19,60]. Trust is defined as “the willingness of a party to be vulnerable to the actions of another party based on the expectation that the other will perform a particular action important to the trustor, irrespective of the ability to monitor or control that other party” ([61], p. 712). Baş et al. [9] foundthat trust influences attitudes toward organic products. Indeed, customer trust may be influenced by factors such as perceived value and ease of use [47], as well as personalization and information quality [60]. In the context of GAFDAs, trust plays a crucial role in reducing customers’ concerns about food quality, online transactions, and data security. Customers are more likely to adopt these apps when they perceive that suppliers adhere to ethical and environmental standards, ensuring product authenticity and safety. Several studies have demonstrated that trust significantly influences attitudes [19], intentions toward using food delivery apps [21,47], and the intention to continue using food ordering apps [3,38]. Recently, Fu et al. [62] found that trust has a significant effect on the intention to purchase green agricultural food products online in China. While Troise et al. [22] found that trust did not significantly impact general food delivery app adoption, other studies suggest that trust plays a crucial role in consumer decisions related to sustainable and organic products. Given that GAFDAs emphasize environmental responsibility and health-conscious consumption, trust in product authenticity and service reliability is expected to positively influence customers’ intention to use these apps. Based on this, we propose the following hypothesis:

H4. 

Trust positively influences the intention to use GAFDAs.

2.2.5. Electronic Word of Mouth (eWOM)

eWOM is defined by Hennig-Thurau et al. [63] (p. 39) as “any positive or negative statement made by potential, actual, or former customers about a product or company, which is made available to a multitude of people and institutions via the Internet”. The Internet has opened up a vast space for customers to share their personal opinions and experiences with others [63], where individuals can post comments, photos, and videos during or after their experiences for a wide audience. These public reviews and experiences can significantly influence the perceptions of other people about GAFDAs, creating positive or negative impressions. Recommendations based on WOM in the context of food ordering apps can be influenced by factors such as attitudes, subjective norms, and security [20]. For instance, Jalilvand and Heidari [64] reported that both face-to-face WOM and eWOM positively influence behavioral intention. In tourism, for example, the intention to visit a destination can be influenced by the source credibility and argument quality of eWOM [40]. Similarly, Jalilvand and Samiei [39] found that eWOM positively impacts both travelers’ attitudes and intentions. Alalwan [56] also found that online reviews and ratings correlate positively with continued intentions to reuse food delivery apps. Based on these findings, we propose the following hypothesis:

H5. 

Positive eWOM positively influences the intention to use GAFDAs.

2.2.6. Gender as a Moderator

Gender differences have been extensively studied in consumer behavior, particularly in the adoption of digital platforms and online services [18,65]. Research suggests that women tend to be more concerned about environmental issues [66,67], making them more likely to choose organic food products. Several studies have also identified gender as a moderator in the relationship between psychological factors and behavioral intentions, influencing how individuals perceive and adopt technology-based services [68,69,70,71,72].

Previous research indicates that women are more influenced by perceived healthiness, attitude toward sustainability [23], service quality [70], perceived innovativeness [68], environmental concern [67], perceived risk [65], social norms, and trust-related concerns [73]. In contrast, men tend to prioritize perceived hygiene, ease of app usage [23], price value [70], brand image [69], system quality [72], and customer service [74] when adopting digital technologies.

Studies also highlight gender’s role in moderating key psychological relationships. For instance, Moon [66] found that gender moderates the link between subjective norms and the intention to choose green restaurants, with a stronger effect among women. Chan et al. [75] reported that subjective norms had a greater impact on men’s intentions to eat healthily, whereas attitude positively influenced women’s behavioral intentions but not men’s. Moreover, research suggests that women are generally more detail-oriented, whereas men focus on broader information processing, and women tend to be more susceptible to persuasion [65].

However, findings on gender’s moderating role remain inconsistent. Asaker [76] found no evidence that gender moderates the relationship between trust and intention to use. Similarly, in the context of “drone food delivery apps”, Hwang et al. [69] observed that while gender moderated the association between desire and intention, it had no significant effect on the link between image and intention. Hwang et al. [68] further found that gender moderated the relationship between product innovativeness and intention but did not significantly influence the connection between attitude and intention. Additionally, some studies have reported no significant moderating effect of gender [71,77], suggesting that its influence may depend on factors such as cultural background, technological familiarity, and the specific context of adoption.

Given these mixed findings, it is essential to further investigate whether gender significantly moderates the influence of key predictors—“attitude, subjective norms, PBC, trust, and eWOM”—on the intention to use GAFDAs.

H6. 

Gender significantly moderates the relationship between predictors (attitude = H6a, subjective norms = H6b, PBC = H6c, trust = H6d, eWOM = H6e) and the intention to use GAFDAs.

Figure 1 illustrates the theoretical framework.

3. Materials and Methods

3.1. Sample and Data Collection

The target population for this study consisted of customers aged 18 and above residing in three northern Algerian cities: Algiers, Blida, and Medea. Given the exploratory nature of this research and the challenges associated with accessing a complete sampling frame for digital consumers, we adopted a convenience sampling approach—a widely used and accepted method in consumer behavior and marketing research for efficiently collecting primary data (e.g., [21,32,56,66]). To enhance the representativeness of the sample, deliberate efforts were made to include participants with diverse backgrounds in terms of gender, age categories, and educational level. In total, 350 Algerian customers were recruited for the study. Participants were provided with detailed information regarding the research objectives and the estimated survey completion time, which was approximately 12 min. Since the questionnaire was designed to be anonymous and confidential, and participation was entirely voluntary, ethical approval and formal consent were not required [31,32,33].

Data collection was conducted between 5 May and 30 June 2023, with surveys distributed at various public venues such as shopping malls, grocery stores, and parks. No incentives were offered to participants. Of the 350 paper-based questionnaires disseminated, 252 were completed and deemed valid for analysis, resulting in a response rate of 72%. The demographic profile of the respondents is summarized in

Table 1. Demographics of the respondents (n = 252).

Variable	n	Percentage (%)
Gender
Male	137	54.34
Female	115	45.63
Age
18–35 years	113	44.84
36–50 years	98	38.89
51–65 years	34	13.49
Over 65 years	7	2.78
Marital Status
Single	88	34.92
Married	138	54.76
Others (Divorced or widowed)	16	6.35
Undefined	10	3.97
Education level
Secondary and lower	93	36.90
Undergraduate	127	50.40
Post-graduate	32	12.70
Monthly Income
<DZD 40,000	76	30.16
DZD 40,001–60,000	81	32.14
DZD 60,001–80,000	35	13.89
>DZD 800,000	31	12.30
Undefined	29	11.51

. The sample comprised 54.34% males, while 44.84% were between 18 and 35 years old. Additionally, 54.76% were married, 50.40% held an undergraduate degree, and 32.14% reported a monthly income ranging between 40,001 and 60,000 Algerian Dinar (DZD).

3.2. Instrument Development

This study employs a quantitative research approach, utilizing a structured questionnaire as the primary data collection tool. The questionnaire is divided into two sections: the first gathers demographic information, while the second focuses on measuring the study constructs, including both independent and dependent variables. To assess these constructs, validated measurement scales were adapted and contextualized for this study. The scale for attitudes toward GAFDAs was derived from Belanche et al. [20], Bouarar et al. [16], and Poon et al. [37]. Subjective norms were measured using items from Poon et al. [37] and Choe et al. [50], while PBC was adapted from the same sources. The trust scale was based on Troise et al. [22] and An et al. [78], whereas the eWOM scale was drawn from Jalilvand and Heidari [64]. Finally, customer intention was assessed using scales from Choe et al. [50] and An et al. [78].

The original questionnaire was developed in English and subsequently translated into Arabic. To ensure content validity, two subjectmatter experts reviewed the translated version, leading to minor revisions. Specifically, the fourth item under the attitude construct, “Using a green agri-food delivery app would be pleasant”, was removed. A 5-point Likert scale (1 = “strongly disagree”, 5 = “strongly agree”) was used to measure respondents’ level of agreement with various statements. To evaluate the clarity and comprehensibility of the items, a pilot study was conducted with 30 participants. The results indicated no significant issues, confirming the questionnaire’s suitability for the main survey.

Table 2. Constructs and items.

Constructs and Measurement Items	Source
Attitude toward green agri-food delivery apps (GAFDAs)	Belanche et al. [20]; Bouarar et al. [16]; Poon et al. [37]
AT1 Using GAFDA is a good idea
AT2 Using GAFDA is a wise idea
AT3 I like the idea of using GAFDAs
Subjective Norms	Poon et al. [37]; Choe et al. [50]
SN1 Most people who are important to me think I should use GAFDAs
SN2 People whose opinions I value would support me to use GAFDAs
SN3 Most people who influence my decisions agree with me about using GAFDAs
Perceived behavioral control (PBC)	Poon et al. [37]; Choe et al. [50]
PBC1 I have the resources, skills and opportunities to using GAFDAs
PBC2 I am confident that if I want, I can use GAFDAs
PBC3 I think there is nothing that will prevent me from using GAFDAs
Trustin green agri-food delivery apps (GAFDAs)	Troise et al. [22]; An et al. [68]
TR1 Transactions via GAFDAs are safe
TR2 The privacy of green agri-food delivery users is well protected
TR3 Green agri-food delivery services are reliable
Electronic word of mouth (eWOM)	Jalilvand and Heidari [64]
eWOM1I often read online comments to find out which GAFDAs leave “a good impression on others”
eWOM2 Online comments influence my decision to choose a GAFDA
eWOM3 I have confidence in online recommendations on GAFDAs
Intention to use green agri-food delivery apps (GAFDAs)	Choe et al. [50]; An et al. [78]
INT1 I intend to use GAFDAs in the future
INT2 I will plan to use GAFDAs in the future
INT3 I am likely to use GAFDAs in the future

Source(s): Authors’ own work.

presents a comprehensive overview of the measurement items used in this study.

4. Results

This section presents the findings from the statistical analysis. We used SPSS 26 to calculate demographic characteristics, means, and standard deviations. The quantitative analysis was carried out using Smart PLS 4, employing PLS-SEM and SPSS to test the model’s fit and the hypotheses.

4.1. Descriptive Statistics

Based on the analysis, the descriptive statistics presented in Appendix A.1 show several findings. Customers showed positive attitudes toward using GAFDAs, with a mean score of 3.69. Moreover, customers perceived a moderate level of social influence regarding GAFDAs, with a mean score of 3.60. Customers also had a high level of PBC, with a mean score of 4.17, suggesting that they feel confident about using GAFDAs. Customers expressed a moderate level of trust in GAFDAs, with a mean score of 3.48. Furthermore, the intention to use GAFDAs was strong among customers, with a mean score of 4.03. However, the perception of the role of eWOM in influencing GAFDA usage was relatively low, with a mean score of 3.08. These findings indicate that while customers show favorable attitudes and intentions to use GAFDAs, their trust and the influence of eWOM are moderate or lower, which may suggest areas for improvement.

Table 3. Analysis of gender-based differences in study variables.

Variables	Male (n = 137)		Female (n = 115)		t-Value
	Mean	Std	Mean	Std
Attitude	3.445	0.813	3.985	0.559	−6.024 **
Subjective norms	3.384	0.727	3.869	0.609	−5.676 **
PBC	3.929	0.651	4.478	0.553	−7.131 **
Trust	3.284	0.835	3.715	0.751	−4.271 **
eWOM	2.924	0.666	3.197	0.674	−3.276 *
Intention to use	3.688	0.482	4.446	0.333	−14.244 **

Note: ** p < 0.01, * p < 0.05. Source(s): Authors’ own work.

presents a comparative analysis of gender-based differences in the study variables, highlighting statistically significant distinctions between the male and female participants. The results indicate that females consistently reported higher mean scores across all the study constructs, including attitude (3.98 vs. 3.44), subjective norms (3.86 vs. 3.38), PBC (4.47 vs. 3.92), trust (3.71 vs. 3.28), eWOM (3.19 vs. 2.92), and intention to use GAFDAs (4.44 vs. 3.68).

The t-values confirm that these differences are statistically significant, particularly for attitude (−6.024 **), PBC (−7.131 **), and intention to use (−14.244 **), suggesting that women exhibit stronger positive perceptions and greater behavioral intentions toward GAFDAs compared to men.

4.2. Model Evaluation

Cronbach’s alpha coefficients were calculated to measure the internal consistency, where the Cronbach’s alpha values ranged between 0.760 and 0.944, which indicates that all the coefficients are higher than the threshold (i.e., 0.7) recommended by Hair et al. [79]. As shown in Figure 2 and

Table 4. Confirmatory factor analysis results.

Construct	Item	Loading	CA	CR	AVE
Attitude	AT1	0.887	0.870	0.879	0.793
	AT1	0.892
	AT1	0.893
Subjective norms	SN1	0.897	0.894	0.896	0.825
	SN2	0.926
	SN3	0.901
PBC	PBC1	0.871	0.835	0.836	0.752
	PBC2	0.872
	PBC3	0.858
Trust	TR1	0.970	0.944	0.946	0.899
	TR2	0.938
	TR3	0.936
eWOM	eWOM1	0.869	0.825	0.831	0.740
	eWOM2	0.872
	eWOM3	0.841
Intention to use	INT1	0.826	0.760	0.774	0.675
	INT2	0.863
	INT3	0.772

Note: Cronbach’s alpha (CA), composite reliability (CR), average variance extracted (AVE), perceived behavioral control (PBC), electronic word of mouth (eWOM), and all item loadings are significant at p < 0.001. Source(s): Authors’ own work.

, the indicator loadings ranged between 0.772 and 0.970, which indicates that all the loadings are greater than the threshold (i.e., 0.7; p < 0.001), as suggested by Hair et al. [79]. In addition, “composite reliability” (CR) values and “average variance extracted” (AVE) were calculated to test the internal reliability and convergent validity. The results shown in Table 4 revealed that the CR values and AVE values also exceeded the threshold of 0.7 and 0.5, respectively, as discussed by Hair et al. [79]. Hence, it is clear that the constructs of the proposed model have sufficient construct reliability and convergent validity. In addition, to assessment the discriminant validity we used both the Fornell–Larcker and “Heterotrait–Monotrait Ratio” (HTMT) criteria [79]. The results shown in Appendix A.1 show that the diagonal AVE scores (“the square root of AVE”) for each variable were greater than the variable’s correlations with any other variable, as suggested by Fornell and Larcker [80]. Likewise, the results shown in Appendix A.2 show that the HTMT values were below 0.90, as proposed by Henseler et al. [81], indicating satisfactory discriminant validity.

4.3. Hypothesis Testing

Our structural model showed a good model fit with the levels suggested by Hair et al. [79], as follows: χ2 = 527.102, p < 0.001, NFI = 0.834, RMR = 0.053, and RMSEA = 0.05. To test the significance of the path coefficients, the authors used a bootstrap procedure with 5000 samples.

Table 5. Test of hypotheses.

Relationship	β	t-Value	Sig.	Results
H1: Attitude → intention to use GAFDAs	0.252	3.491	0.000	S
H2: SNs → intention to use GAFDAs	0.169	2.421	0.016	S
H3: PBC → intention to use GAFDAs	0.284	3.866	0.000	S
H4: Trust → intention to use GAFDAs	0.148	2.380	0.017	S
H5: eWOM → intention to use GAFDAs	0.123	2.833	0.005	S

Note: Subjective norms (SNs), perceived behavioral control (PBC), electronic word of mouth (eWOM). R² = 0.559 (R² Adjusted= 0.550), S = Supported; NS = Not Supported. Source(s): Authors’ own work.

shows that attitudes (β = 0.252, t = 3.491, p < 0.001), subjective norms (β = 0.169, t = 2.421, p < 0.001), PBC (β = 0.284, t = 3.866, p < 0.001), trust (β = 0.148, t = 2.380, p < 0.001), and eWOM (β = 0.123, t = 2.833, p < 0.001) have positive influences on intention to use GAFDAs, therefore confirming hypotheses H1, H2, H3, H4, and H5. In addition, the R² adjusted value was 0.550; meaning the expanded TPB model explains 55% of the variance in intention to use GAFDAs.

Table 6. Summary of the gender-based multiple regression model.

Relationship	β	t-Value	VIF	Results
Males (R2 = 0.556)
H1m: Attitude → intention to use GAFDAs	0.130	2.393 *	2.438	S
H2m: SNs → intention to use GAFDAs	0.068	1.081	2.628	NS
H3m: PBC → intention to use GAFDAs	0.144	2.441 *	1.883	S
H4m: Trust → intention to use GAFDAs	0.184	3.809 **	2.068	S
H5m: eWOM → intention to use GAFDAs	0.078	1.778 *	1.081	NS
Females (R2 = 0.262)
H1f: Attitude → intention to use GAFDAs	0.105	1.896	1.278	NS
H2f: SNs → intention to use GAFDAs	0.121	2.261	1.422	S
H3f: PBC → intention to use GAFDAs	0.142	2.623	1.204	S
H4f: Trust → intention to use GAFDAs	0.021	0.522	1.243	NS
H5f: eWOM → intention to use GAFDAs	0.048	1.073	1.095	NS

Note: Subjective norms (SNs), perceived behavioral control (PBC), electronic word of mouth (eWOM), m = male, f = female, ** p < 0.01, * p < 0.05, S = Supported; NS = Not Supported. Source(s): Authors’ own work.

summarizes the gender-based multiple regression model, revealing differences in the predictors of intention to use GAFDAs between the male and female participants. For the males (R² = 0.556), attitude (β = 0.130, t = 2.393 *), PBC (β = 0.144, t = 2.441 *), and trust (β = 0.184, t = 3.809 **) significantly influenced their intention to use GAFDAs, while subjective norms and eWOM were not significant predictors. Conversely, for the female participants (R² = 0.262), subjective norms (β = 0.121, t = 2.261 *) and PBC (β = 0.142, t = 2.623 *) played significant roles, whereas attitude, trust, and eWOM did not. These results indicate that males rely more on personal attitudes and trust, whereas females are influenced by social norms and their PBC. The lower R² value for the females suggests that additional factors, beyond those tested, may drive their intention to adopt GAFDAs. These insights emphasize the importance of gender-specific marketing strategies to enhance adoption.

4.4. Moderation Analysis

To assess whether gender moderates the key relationships, we conducted a moderated multiple regression (MMR) analysis, similar to the work of Hong et al. [82].

Table 7. Moderation analysis by gender.

Relationship	β	t-Value	Sig.	Results
H6a: Attitude*Gender → intention to use	−0.025	−0.318	0.751	NS
H6b: SNs*Gender → intention to use	0.054	0.649	0.517	NS
H6c: PBC*Gender → intention to use	−0.002	−0.023	0.982	NS
H6d: Trust*Gender → intention to use	−0.163	−2.572	0.011	S
H6e: eWOM*Gender → intention to use	−0.030	−0.480	0.632	NS

Note: Subjective norms (SNs), perceived behavioral control (PBC), electronic word of mouth (eWOM), F = 53.071 (p < 0.001), S = Supported; NS = Not Supported. Source(s): Authors’ own work.

presents the results, evaluating gender’s influence on the associations between key predictors and the intention to use GAFDAs. The analysis indicate that gender does not significantly moderate the effects of attitude (β = −0.025, p = 0.751), subjective norms (β = 0.054, p = 0.517), PBC (β = −0.002, p = 0.982), or eWOM (β = −0.030, p = 0.632) on behavioral intention, suggesting that these factors influence men and women similarly in this context. However, trust (β = −0.163, p = 0.011) was found to be significantly moderated by gender, implying that men and women differ in the extent to which trust affects their intention to use GAFDAs. This suggests that its influence on behavioral intention varies by gender. Specifically, trust-related concerns appear to have a stronger impact on men compared to women.

5. Discussion

This is one of the few studies that promote food-related health behaviors through the adoption of GAFDAs. The aim of this research was to investigate the intention to use GAFDAs. The results showed that five hypotheses were supported, and the current research model was able to predict 55% of the variance in customers’ intention to use GAFDAs. Additionally, consistent with the assumptions of the TPB model [30] and in line with previous empirical evidence, attitude was shown to have a strong correlation with the intention to use food delivery apps [19,22,37,68,83]. Furthermore, Choe et al. [50] found that attitude (β = 0.693, p < 0.05) is the strongest predictor of the intention to use drone food delivery apps. In fact, attitudes not only influence intentions but have also been shown to affect the intention to continue using food delivery apps [4,36]. Customer attitudes toward GAFDAs are influenced by factors such as the opinions of others, eWOM about GAFDAs, perceived price fairness, trust, and other contextual factors.

In addition, our results reveal that subjective norms positively influence the intention to use GAFDAs. As evidenced by several studies, subjective norms, similarly to social influence, have a significant effect on the intention to use and reuse food delivery apps [3,21,22,37,42,50,56,69]. This suggests that customers rely on subjective norms when using GAFDAs. This could be due to the ability of subjective norms to provide useful information about green agri-food (such as suppliers, available applications, and green agri-food alternatives). In terms of eating behavior, customers often try to align their actions with reference groups (coherently and consistently), especially close groups [55,75,77].

Moreover, the results indicate that PBC is the strongest predictor of the intention to use GAFDAs. This suggests that customers who have the necessary resources, skills, and opportunities are more likely to use GAFDAs, compared to those who lack the resources, skills, or opportunities. Returning to the TPB, this result aligns with Ajzen’s [84] assertion that “a favorable attitude and a supportive subjective norm provide the motivation to engage in the behavior, but a concrete intention to do so is formed only when perceived control over the behavior is sufficiently strong” (p. 315). Lee et al. [1] found that the intention to use food delivery apps is influenced by ease of use, which in turn is affected by factors such as “easy registration” and “easy payment”. Previous studies have identified PBC as an important antecedent of the intention to use food ordering apps [37,50,63]. However, in the digital age, young customers seem to have good control over the use of mobile apps. In the context of food delivery, Belanche et al. [20] found that PBC is particularly important for older customer segments.

Similarly, the results show that trust plays a significant role in the intention to use GAFDAs. This suggests that customers who trust GAFDAs are more likely to use them. Many studies have supported a significant positive correlation between trust and willingness to use food delivery apps [19,21,22,38,47,73,82,83]. In their study, Su et al. [60] demonstrated that personalization, ease of use, usefulness, and information quality were related to trust in food delivery apps. However, in the era of digital transactions, earning customer trust has become more difficult, as it is closely tied to feelings of security and the absence of fear. For example, negative eWOM (or WOM) from a trusted source can undermine the trust in a particular brand, product, or app (e.g., [73,85]). Furthermore, a decline in trust toward a specific app is likely to result in users avoiding it and seeking out a more trustworthy alternative. In this context, a money-back guarantee can foster positive attitudes [37] and increase trust in GAFDAs. Therefore, we conclude that the success of GAFDAs depends on their ability to build strong trust among target customers.

Beyond the adoption of GAFDAs, trust has been widely recognized as a key determinant in various consumer decisions, including willingness to pay for environmentally friendly products [86] and the acceptance of artificial intelligence (AI) in service industries [87,88]. Recent studies highlight that consumers are more likely to engage in sustainability-related behaviors when they trust the service provider, product, technology, or regulatory framework [29,86,89,90]. In AI-based applications, trust influences user attitudes, thereby increasing adoption rates of automated services and AI technologies [91]. By integrating trust into the TPB framework, our study captures an essential psychological factor that enhances the model’s predictive power, offering insights into how consumer confidence in digital platforms influences green consumption behaviors.

Furthermore, our results showed that eWOM communications affect the intention to use GAFDAs. This means that positive eWOM about GAFDAs may encourage customers to use them, while negative eWOM can discourage use. Users’ experiences and satisfaction with GAFDAs play a significant role in influencing eWOM communications either positively or negatively. The positive correlation between eWOM and behavioral intention has been supported by previous empirical studies [39,40,64,73]. Additionally, Kumari and Sangeetha [92] found that positive eWOM dimensions (argument, credibility, and valence) positively affect the intention to book green hotels. In this sense, eWOM is a critical factor shaping consumer perceptions and influencing decision-making processes in online marketplaces. Given that digital platforms rely heavily on peer reviews, ratings, and social recommendations, eWOM has been found to significantly impact behavioral intentions in various domains, including travel intention [64], sustainable purchasing decisions [92], and AI-based service adoption [93]. Consumers tend to rely on eWOM to assess service quality, reduce uncertainty, and validate their choices, particularly when engaging with new or unfamiliar digital services [94]. Prior research suggests that positive eWOM fosters greater consumer trust and enhances the perceived credibility of green and AI-driven technologies, ultimately driving adoption [85]. By incorporating eWOM into the extended TPB model, our study acknowledges the growing importance of social influence in shaping consumer behavior and highlights its relevance in the sustainable agri-food sector.

Finally, the findings of this study indicate that gender does not moderate the relationship between the intention to use GAFDAs and its key determinants (i.e., TPB constructs and eWOM). While previous research, e.g., Refs. [66,75], suggests that gender plays a moderating role in the link between subjective norms and intention, as well as between attitude and intention in the context of healthy eating, this pattern does not hold in the present study. Additionally, Goel and Parayitam [65] provide evidence supporting gender’s moderating influence in consumer behavior, demonstrating how it can strengthen the association between intention and its predictors, such as perceived risk. However, our findings suggest that gender differences in decision-making may be diminishing in the context of GAFDA adoption, possibly due to increasing digital literacy and widespread exposure to online consumer experiences across both genders.

The results from the gender-based multiple regression analysis (Table 6) and the moderation analysis (Table 7) indicate that trust plays a more significant role in shaping men’s intention to use GAFDAs compared to women. Specifically, trust was found to be a significant predictor for males but not for females. Furthermore, the moderation analysis (Table 7) revealed that gender significantly moderates the relationship between trust and behavioral intention, reinforcing the idea that trust is a stronger determinant for men when adopting GAFDAs. These results are consistent with the findings of Wen et al. [38], who reported that trust exerts a stronger influence on males’ intentions to use food delivery apps. They also align with Teo et al. [74], whose research indicated that trust more significantly impacts males’ intentions to reuse e-commerce platforms. However, they contrast with Asaker’s [76] findings, which suggest that gender does not moderate the trust-intention relationship. Additionally, they diverge from Ting and Ahn’s [90] research, which revealed that trust exerted a stronger influence on sustainable consumer behavior among female customers. In fact, men may demonstrate greater risk sensitivity in digital environments [23], prompting them to prioritize trust-related factors such as security, privacy, and the reliability of customer service [74].

In contrast, women’s behavioral intentions appear to be driven more by social and psychological factors, such as subjective norms, platform design, fulfillment, and perceived ease of use [66,74,76]. This is consistent with previous literature indicating that women rely more on peer recommendations, social approval, perceived innovativeness, and their perceived ability to control app usage when making digital purchasing decisions [68]. The absence of a significant effect of trust among female consumers suggests that women may already possess a baseline level of trust in online food delivery systems or prioritize other factors over trust when considering GAFDA adoption. Alternatively, women may develop trust in these platforms through indirect cues, such as positive social influence, brand reputation, or previous experience, rather than explicit security or reliability assurances. According to Boldureanu et al. [73], female consumers tend to develop trust by prioritizing factors that enhance reliability and simplify usability.

One possible reason for the limited moderating effect of gender in this study could be the growing normalization of digital platforms across genders, reducing traditional differences in decision-making processes. Another explanation might be that gender’s influence is context-dependent, varying across industries, cultural settings, or specific consumer behaviors [71]. Unlike previous studies (e.g., [68,69,70]), which highlight gender’s moderating role in digital adoption, our findings suggest that gender differences may be diminishing in the GAFDA context, potentially due to increased familiarity with online platforms and evolving consumer behaviors across both men and women.

In addition to gender, prior research suggests that other demographic variables may play important moderating roles in the adoption of green technologies and digital platforms [66]. For instance, the original UTAUT identifies age, experience, and voluntariness of use, alongside gender, as key moderators in technology acceptance models [54]. Furthermore, research in social media marketing has shown that education level can positively moderate the relationships between “perceived quality”, “brand awareness”, and “brand loyalty”, and consumers’ intentions to continue engaging with a brand [95]. Similarly, age has been shown to influence the strength of relationships between attitude, PBC, perceived innovativeness, and various behavioral outcomes, such as WOM intentions [66,68]. These findings suggest that education, age, and digital experience may also moderate the effect of trust, PBC, and eWOM on consumers’ intention to use GAFDAs.

6. Conclusions

In the age of digital marketing, it has become essential for green agri-food marketers to embrace mobile apps to encourage customers to order their products. Furthermore, the success of GAFDAs is closely tied to customers’ acceptance of their use. However, previous studies have not sufficiently investigated the behaviors related to accepting and using GAFDAs. To address this gap, this research empirically examined Algerian customers’ intentions to use GAFDAs by integrating the TPB framework with trust and eWOM. The empirical results revealed that PBC, attitudes, subjective norms, trust, and eWOM are significant antecedents in forming positive intentions toward using GAFDAs. Moreover, the extended TPB model was able to explain 55% of the variance in intentions to use GAFDAs. These findings suggest (1) the validity and feasibility of adopting the TPB model in the context of online agri-food marketing and (2) the importance of incorporating new constructs into the TPB model to enhance its predictive ability.

6.1. Theoretical Contribution

This study makes several key theoretical contributions. First, it extends the TPB framework by integrating trust and eWOM to provide a more comprehensive understanding of consumers’ intentions to adopt GAFDAs. While the TPB has been widely used in consumer behavior research, prior studies have primarily focused on rational decision-making factors, namely, “attitude, subjective norms, and PBC”. However, digital consumption environments introduce additional psychological and social influences that shape consumer behavior [22,50,96]. By incorporating trust and eWOM, our study advances the theoretical discourse by illustrating how both cognitive and affective factors interact in the decision-making process. Specifically, trust functions as a psychological enabler that mitigates perceived risks associated with digital transactions, particularly in contexts where sustainability claims, food quality, and ethical sourcing may be questioned. Simultaneously, eWOM acts as a social validation mechanism, reinforcing consumer confidence through peer recommendations and online reviews. These findings enrich TPB-based models by demonstrating that, beyond individual attitudes and social norms, digital trust and peer influence play a fundamental role in the adoption of sustainable food technologies.

A second key contribution of this study lies in its gender-based perspective on digital adoption. While prior research suggests that women generally exhibit stronger pro-environmental attitudes than men, studies have also shown that men tend to demonstrate more consistent purchasing behaviors for green agri-food products. This study builds on the existing literature by revealing gender-based differences in the determinants of behavioral intentions, emphasizing the need for tailored strategies to encourage GAFDA adoption among different demographic groups.

Beyond gender, our study contributes to the broader discourse on digital trust and social influence in emerging markets. While much of the existing research on digital adoption has focused on technological benefits and economic incentives, this study shifts attention toward the psychosocial factors that influence consumer engagement with digital platforms. The findings highlight that trust is not merely a risk mitigation factor but also a key driver of digital service adoption, particularly in environments where institutional trust in digital transactions is still evolving. Additionally, eWOM emerges as a powerful credibility-enhancing mechanism, especially in markets where formal consumer protection policies may be less robust or inconsistently enforced. These insights refine existing theoretical models of digital adoption by demonstrating how trust and peer influence serve as critical levers in shaping online purchasing behavior, particularly in markets characterized by regulatory and technological uncertainties.

6.2. Managerial Implications

This study provides practical insights for various stakeholders, including producers, distributors, retailers, and app developers, aiming to promote the adoption of GAFDAs. First, since attitudes toward green agri-food play a crucial role in shaping users’ intention to adopt GAFDAs, it is essential for businesses to foster positive perceptions. This can be achieved by clearly communicating the benefits of green and healthy food, such as improved well-being and environmental sustainability. Marketing campaigns should be tailored to appeal to specific consumer segments—such as athletes, patients, health-conscious individuals, and higher-income or highly educated users—through persuasive messaging that highlights quality, safety, and nutritional value. Additionally, app interfaces should offer transparent information, including product origins and nutritional content, which can reinforce users’ confidence and encourage adoption.

Second, the study confirms that subjective norms significantly influence intention, especially among female users. Stakeholders should leverage social influence by involving credible figures such as nutritionists, doctors, and fitness trainers to endorse GAFDA use. Furthermore, reference groups and community institutions—such as schools, health associations, gyms, and religious centers—can play a proactive role in normalizing the use of GAFDAs and promoting healthy eating habits through trusted social channels.

Third, given that PBC emerged as the strongest predictor of behavioral intention, simplifying the user experience is critical. App developers should ensure that the platforms are user-friendly and quick to navigate and require minimal effort to complete tasks such as food selection, payment, and order tracking. Providing in-app tutorials or guides for new users, especially the elderly or those unfamiliar with technology, can reduce barriers to entry. Additionally, expanding distribution coverage and ensuring reliable access to products in various locations will help minimize user frustration and support continued app usage.

Fourth, trust plays a particularly important role—especially for male consumers—in influencing the decision to use GAFDAs. To strengthen trust, businesses should adopt transparent practices regarding food sourcing, implement robust data security measures, and highlight service reliability. Features such as real-time order tracking, visible customer support options, and secure payment gateways can further reassure users about the safety and reliability of the app experience.

Fifth, eWOM presents a powerful tool for encouraging adoption. Marketers should actively encourage satisfied customers to share their positive experiences through reviews, testimonials, and social media. Collaborating with influencers and content creators who resonate with the target audience can help enhance credibility and broaden reach. User-generated content that emphasizes authenticity, sustainability, and satisfaction can be featured across marketing channels to amplify trust and brand engagement.

Finally, the findings highlight important gender-based differences in user behavior, suggesting the need for gender-sensitive strategies. For male users, focusing on functional aspects such as trust, efficiency, and control will likely drive adoption. In contrast, female users may respond more to social proof, peer influence, and recommendations from trusted experts. Therefore, marketing approaches should be tailored accordingly—highlighting technical reliability and data safety for men, while emphasizing community engagement and peer validation for women. Simplifying app usability and maintaining high service quality are essential for both groups, but messaging and engagement tactics should reflect these gendered preferences.

By implementing these targeted strategies, stakeholders can better align their efforts with consumer behavior, overcome barriers to adoption, and support the sustainable growth of green agri-food delivery services.

6.3. Limitations and Future Research

Although this paper makes a significant contribution, it is not without limitations that present opportunities for future research. First, the extended TPB model explains 55% of the variance in customers’ intentions, meaning that 45% of the variance is influenced by other factors, such as moral norms, perceived green agri-food quality, health consciousness, and environmental knowledge. Therefore, future studies could extend the TPB by incorporating these or other factors. Additionally, our model can be applied to examine intentions to use GAFDAs in other countries, with the possibility of excluding some constructs and including others. Second, our study focused on examining direct effects but did not investigate the role of mediating variables (e.g., price sensitivity, willingness to pay, green skepticism, and perceived value) or moderating variables (e.g., age, educational level, income, and cultural factors). Future research could explore these roles further. Third, we explored customers’ intentions, but there may be a gap between intention and actual behavior [32]. Therefore, we recommend that future studies examine the factors influencing actual behavior in using GAFDAs, investigate re-use intentions, and measure customer satisfaction levels. Finally, given the role that the perceived benefits of healthy food (such as organic food) may play in influencing the willingness to use healthy food ordering apps, it will be important for researchers to explore this role in different cultural contexts, as this area remains largely unexplored.