From Linear to Circular Economy: Embracing Digital Innovations for Sustainable Agri-Food Waste Management among Farmers and Retailers

Abstract

Food waste is an issue throughout the food supply chain from production to consumption, especially in the later stages, such as retailing and final consumption. For the future of the developing world, changes in farming and retail practices are crucial. This study introduces a digital system for managing agricultural waste in Thailand that aims to encourage farmers and food retailers to sell their excess agricultural materials. The study’s objectives are as follows: (1) to explore factors that affect users’ behavioral intention to utilize an agriculture waste trading platform; (2) to compare the behavioral differences between farmers and retailers regarding their intention to use a digital platform for sustainable agriculture. Data were gathered from 570 fruit and vegetable sellers and farmers across five provinces in the northeastern region of Thailand. Structural equation modeling (SEM) was used to analyze the relationships between constructs based on the modified Unified Theory of Acceptance and Use of Technology (UTAUT2), and multigroup analysis (MGA) was employed to analyze differences in path coefficients across groups. The key findings revealed that social influence (SI) had a more significant impact on retailers compared to farmers, while facilitating conditions (FC), habits (HB), and privacy (PR) were necessary for both groups. Unlike retailers, farmers were also motivated by hedonic motivation (HM) from using the platform. Explicitly, retailers’ behavioral intentions were influenced by a more significant number of factors than those of farmers. This research suggests that policymakers should develop targeted marketing campaigns leveraging social influence for retailers, improve platform usability and security, and create incentives for habitual use to enhance platform adoption. Additionally, policymakers should promote engaging features for farmers, provide comprehensive education and training, and advocate for supportive policies and financial incentives. Strategic actions to facilitate the transition toward a circular economy will improve the environmental sustainability and economic resilience of the agri-food sector.

1. Introduction

Food waste has become a significant issue impacting the environment, economy, and local communities in recent years. Due to the increasing population and limited resources, there is a pressing need for changes in food production and distribution. Food waste happens throughout the supply chain, including production, post-harvesting, processing, and especially at the end of the supply chain, such as retail and consumption [1]. Studies have shown that the consumer sector (comprising retail, food services, and households) contributes to almost two-thirds of all food wastage globally. This results in an economic cost increase of USD 1 trillion, along with environmental costs amounting to USD 700 billion annually. Additionally, there is an estimated spending of USD 90 billion on this issue. The total global cost of food wastage stands at USD 2.6 trillion per year [2]. In the ASEAN, Thailand ranks second in waste generation, with an amount of 26.77 million tonnes [3]. The highest waste generation within ASEAN is attributed to Indonesia at 64 million tonnes annually, followed by Vietnam, which ranks third, with 22 million tonnes per year. The Sustainable Development Goals (SDGs) set by the United Nations set a target of a 50% decrease in worldwide per capita food wastage along with reducing waste in supply chains at the retail and consumer stages [4].Thailand is actively exploring ways to research and implement strategies to reduce food wastage and improve its management practices. The concept of the circular economy was developed to address the issue of waste within global food systems.

The global food waste problem has led the SDGs to endorse a circular economy as a solution for reducing food waste within individual countries. The concept of the circular economy, supported by the European Commission endorsement in 2015, aims to foster a zero-waste economy where materials circulate continuously through strategies such as recycling, innovative designs, and reusing materials and energy [4].The role of the circular economy is to promote the adoption of circular materials within the economy. In the agricultural industry, fruit and vegetable waste is produced during planting due to harvesting practices, resulting in the deterioration of raw materials. Fresh fruit and vegetables are commonly consumed, but excess fruit waste is produced from products such as fruit juice. These wastes are disposed of in landfills, impacting the environment negatively with issues such as unpleasant smells and uncleanliness [5]. Additionally, the methods used to prepare, peel, and extract seeds from fruits and vegetables before they are sold in stores result in the production of agricultural food waste. The central focus of this study pertains to the significance of the circular economy in mitigating the generation of waste stemming from fruit and vegetable production.

Converting agricultural waste into valuable resources has enormous potential to improve sustainability and resource efficiency. Thailand’s agricultural sector confronts considerable obstacles in shifting from a linear to a circular economy, particularly in waste valorization and the digital transformation of agri-food supply chains. Significant systemic adjustments in farming and retailing are required for a transition from the traditional linear paradigm of creating, consuming, and discarding to a circular economy that stresses recycling, reuse, and waste reduction (see Figure 1). The linear economy model encompasses a series of interconnected nodes, beginning with farming and followed by harvesting, packaging, distribution, and finally, retailing. This model has two primary stakeholders: farmers and retailers. The majority of waste is produced throughout the stages of harvesting, packaging, and retailing. In the circular economy model, farmers and retailers require significant adjustments to their routines to valorize food and agricultural wastes. Through digital technology for waste trading, sellers (farmers and retailers) and customers (value-added material manufacturers) can conduct their transactions online. Next, the recycle center intermediary travels to pick up agricultural wastes from farmers and retailers, transforms those wastes into value-added materials, such as bio-based leather and paper, and delivers those materials to the necessary manufacturer to produce the final products.

Based on a linear economic model, the reduction in agricultural waste is now a major study issue in Thailand, and the country’s agricultural industry is looking into the switch to a circular economy. Nevertheless, limited research has been conducted in emerging countries specifically pertaining to this subject matter Cane and Parra [6] stated that several research papers from the developed world have discovered solutions to reduce food waste using various technologies, such as food-sharing smartphone applications to exchange food that is close to expiration and cannot be consumed in time. In the developed world, digital platforms against food waste include food redistribution platforms, food sharing apps, food rescue apps, meal planning apps, and food waste tracking applications, which assist in minimizing food waste and promoting sustainability throughout the supply chain [6,7]. These services connect surplus food with those in need, allow individuals to exchange excess food, rescue unsold food at discounted prices, and provide effective meal planning and waste tracking tools. However, little research has been done into the precise elements that influence the design and implementation of new digital platforms for agricultural waste management in Thailand [8]. To address this gap, this study provides a conceptual model through the UTAUT2 framework to investigate the factors affecting farmers’ and agri-food retailers’ acceptance patterns and behavioral intentions as they utilize an agriculture waste trading platform.

The objective of this research is to fully comprehend the actions of real users (fruit and vegetable retailers vs. farmers) by analyzing and classifying their interactions with the agricultural waste platform. We intend to investigate the following research questions (RQs). RQ1: What factors affect users’ behavioral intention to use an agriculture waste trading platform that promotes a circular economy? RQ2: How are the behaviors of farmers and retailers differentiated regarding the intention to use an agriculture waste trading platform that supports a circular economy? Therefore, this study investigates the factors derived from the extended theory of UTAUT2 (performance expectation, effort expectancy, social influence, facilitating conditions, hedonic motivation, price value, and habit) [9] and two additional variables (trust and privacy) [8]. The study uses structural equation modeling (SEM) to analyze the causal relationship between the variables in the model. Explicitly, this study developed a predictive model to identify the factors influencing farmers’ and retailers’ behavioral intentions to use a digital platform for trading agricultural waste. Therefore, structural equation modeling (SEM) is a statistical technique suitable for analyzing the relationship between multiple variables [10,11]. Multigroup structural equation modeling analysis is appropriate for this study since it allows for a deeper assessment of farmer and retailer behaviors, revealing insight into the unique elements that influence their intention to utilize the circular economy digital technology. This research will provide vital insights into the creation of successful digital platforms and agricultural waste management solutions for Thailand’s transition to a circular economy.

In the following sections, we provide a comprehensive analysis of the existing literature related to the circular economy approach for effectively managing agricultural waste, as well as the developments made in the digital platforms field for this purpose. We also provide a review of the theoretical framework known as UTAUT2. The research technique used, which includes information on data collection and data analysis processes, is described in detail in the next section of the study. Subsequently, the following section presents the results obtained from the aforementioned analysis. The implications and limitations of this study are further discussed in the subsequent section, followed by the final conclusion section.

2. Literature Review and Hypothesis Development

The current study emphasizes the digital circular model in the context of waste management. This study employs the existing literature, theoretical frameworks, and relevant research to study the principles of circular economy and the development of a digital platform for waste management, using the Modified Unified Theory of Acceptance and Use of Technology (UTAUT2) with the SEM framework to identify relevant factors and achieve its objectives. Several studies have widely employed SEM models to investigate the circular economy and explain the relationships between users’ perceptions and behavioral intentions [8,12,13,14,15]. This section discusses the related literature and theories, leading to the development of research hypotheses.

2.1. Circular Economy and Digital Platforms for Agri-Food Waste Management

The circular economy approach aims to minimize waste and make the most of resources by creating a closed-loop system where materials are reused, recycled, and repurposed. In the context of agricultural waste management, this approach is essential for reducing environmental impact and enhancing resource efficiency.

As for Thailand, Naruetharadhol et al. [16] and Pienwisetkaew et al. [8] highlighted the growing focus on sustainable consumption and production, emphasizing initiatives to reduce food waste through bio–circular–green innovation. These initiatives encourage businesses to adopt practices that reduce carbon emissions, energy usage, and waste while using sustainable resources. For instance, Li et al. [17] investigated the effect of pickling on Daliuta long-flame coal’s surface composition and floatability, highlighting the role of surface engineering in material recyclability [17]. Similarly, Tang et al. [18] explored the removal of heterocyclic sulfur from coal using a potassium tert-butoxide and hydrosilane system, demonstrating a method to reduce pollutants and improve the quality of recycled fuel materials. Additionally, Pietwisetkakew et al. [8] suggested a digital platform that connects vegetable and fruit retailers with businesses needing materials transformed from food waste, showcasing an innovative approach to turning agricultural waste into valuable products such as bioplastic and vegan leather. These examples demonstrate how the combination of energy and materials engineering can strengthen the concepts of the circular economy by promoting effective resource utilization to meet sustainability goals through the adoption of resource efficiency, recycling, and waste-to-energy technologies.

However, transitioning to a circular economy presents challenges, particularly in Thailand’s agricultural sector. The draft 13th National Economic and Social Development Plan suggests boosting sector investments to improve production efficiency, reduce waste, promote recycling, and support entrepreneurs, aligning with the sufficiency economy philosophy and the United Nations’ Sustainable Development Goals (SDGs). The European Commission also promotes development and resource utilization to improve human life and create value in supply chains [19]. By focusing on reusing, reducing, and recycling food waste, the circular economy aims to conserve products, materials, and resources, impacting the economy, environment, and society by reducing storage costs, stabilizing prices, and enhancing waste management and resource conservation [20,21].

The development of digital platforms plays a crucial role in implementing circular economy principles in agricultural waste management. These platforms facilitate the efficient collection, processing, and redistribution of agricultural waste, transforming it into valuable products such as compost, bioenergy, and animal feed. Herrero et al. [22] suggest that technological and social innovation is necessary for the sustainable evolution of food systems. Digital platforms have gained popularity in various environments and activities [23], presenting opportunities for utilizing technologies in fruit waste trading services to combat food waste issues. By utilizing technology, a digital platform can be developed for farmers and sellers to sell agricultural waste to entrepreneurs, increasing the value of agricultural waste products. The role of technology in waste management involves facilitating waste reduction through reusing and recycling products [24]. Neligan et al. [25] suggests that integrating tools into networks can increase product quality stability and reduce uncertainties. Promoting sustainable food waste reduction efforts is possible through developing digital platforms that establish markets and economic systems for reused and recycled materials, facilitating trade relationships among platform users. For example, Kumar et al. [26] explained the framework of the sharing economy as a business entity or service facilitator connecting product or service suppliers with customers seeking goods and services that do not contribute to waste accumulation. The above studies are summarized in

Table 1. Summary of the literature review on circular economy and digital platforms for agri-food waste management.

Theme	Content	References
Circular Economy and Innovations	The circular economy approach minimizes waste and maximizes resources by creating a closed-loop system where materials are reused, recycled, and repurposed. This approach focuses on reducing food waste through bio–circular–green innovation, which is essential in reducing environmental impact and enhancing resource efficiency. Innovations derived from the circular economy process include, for instance, surface engineering to enhance material recyclability, methods to reduce pollutants in recycled fuels, and digital platforms transforming agricultural waste into valuable bioproducts.	[8,16,17,18]
Sustainable Practices and Challenges	This theme encourages practices that reduce carbon emissions, energy usage, and waste while using sustainable resources. It promotes resource utilization to improve human life and create value in supply chains. Challenges in transitioning to a circular economy suggest boosting investments to improve efficiency, reduce waste, promote recycling, and support entrepreneurs.	[19,20,21]
Development of Digital Platforms	The development of digital platforms plays a crucial role in implementing circular economy principles. These platforms facilitate the efficient collection, processing, and redistribution of agricultural waste, transforming it into valuable products like compost, bioenergy, and animal feed. They connect farmers, waste management companies, and other stakeholders to promote sustainable practices and reduce waste.	[22,23]
Technological and Social Innovation and Opportunities	Technological and social innovation are necessary for the sustainable evolution of food systems. Integrating tools into networks can increase product quality stability and reduce uncertainties. Opportunities and frameworks facilitate waste reduction through reusing and recycling products. They connect product and service suppliers with customers seeking non-wasteful goods and services, thereby enhancing the overall efficiency and sustainability of agricultural practices.	[24,25,26]

.

2.2. The Extended Unified Theory of Acceptance and Use of Technology (UTAUT2)

Several studies have utilized the Unified Theory of Acceptance and Use of Technology 2 (UTAUT2) model to explore how users in the agricultural sector plan to embrace digital platforms [8,27,28]. The Unified Theory of Technology Adoption and Utilization Theory (UTAUT2) was introduced by Venkatesh et al. [29] as an extension of UTAUT. UTAUT2 integrates eight theories on technology adoption. In their study, Venkatesh et al. [29] combined elements from relationship-based theories to identify the factors influencing technology adoption. These factors fall under four categories: performance expectation, effort expectancy, social influence, and facilitating conditions. Taherdoost [30] suggests that gender, age, degree of expertise, and voluntariness of usage are additional elements that enhance the reliability of predictions associated with the adoption and utilization of technology. Nevertheless, it has been established in prior studies that the Unified Theory of Technology Adoption and Utilization Theory (UTAUT) model has a deficiency in terms of comprehensiveness, as noted by Williams et al. [31]. This deficiency pertains specifically to excluding economic growth and the rise of novel technologies inside the model. Therefore, it is crucial to improve the UTAUT model by examining user context and adapting it to meet the changing needs of customers. In a development, Venkatesh et al. [32] extended the Unified Technology Adoption Theory (UTAUT2) by integrating three additional factors: hedonic motivation, price value, and habit.

The agricultural sector plays a role in boosting Thailand’s economy by leveraging its climate and geography for cultivation. Meanwhile, there is an issue of waste generation from the activities of farmers and sellers. To address this issue, researchers are interested in managing these wastes with the idea of a digital platform. This platform would serve as a bridge to gather information from farmers and sellers with materials they wish to sell. Subsequently, the platform operator would repurpose this waste into goods. However, this paper examines factors influencing users’ intentions to utilize the digital platform to trade agricultural waste from farmers and sellers. Drawing insights from the literature, the UTAUT2 model is deemed suitable for analyzing users’ intentions and behaviors concerning consumer needs in this context. Additionally, this research incorporates factors such as trust and privacy along with user categories, like farmers and sellers, which could impact user intentions and behaviors regarding digital platform usage.

2.2.1. Performance Expectancy (PE)

Based on research by Venkatesh et al. [29], performance expectation is defined as an individuals’ belief in the effectiveness of using technology to perform tasks, which ultimately leads to productivity and achieving desired outcomes. This factor plays a role in predicting user intentions and behaviors towards technology adoption, which can be beneficial across various industries [9,29,33]. Additionally, recent studies by Molina Maturano et al. [34], Shi et al. [28], and Han et al. [35] demonstrated that users perceive phone apps as beneficial for their daily activities. These apps provide local information, save time on tasks, and increase users’ willingness to use such technologies.

H1: 

The farmers’ and retailers’ performance expectations affect behavioral intentions to use a digital platform for the circular economy.

2.2.2. Effort Expectancy (EE)

Venkatesh et al. [29] defined effort expectancy as the expectation that users will be able to use technology without exerting significant effort. The perception of usability can be assessed by measuring features such as ease of learning, clarity, comprehensibility, and flexibility, which include user-friendly technology [29]. Additionally, Venkatesh et al. [9] examined the behavior of individuals’ acceptance and adoption of technology to shed light on the factors that affect users’ acceptance of technology. The study discoveries uncovered a positive correlation between effort expectancy and intention. Effort expectancy could potentially help ease users’ challenges in acquiring skills, as suggested by Molina Maturano et al. [34]. In a study by Muangmee et al. [36], it was found that effort expectancy has a notable impact on user behavior regarding the usage of food delivery mobile apps amidst the COVID-19 pandemic. However, the users’ behavior could be influenced by the quality and result in pleasure after the usage experience [37]. Therefore, the level of user satisfaction with the application experience may be influenced by the users’ interactions with those services. Users’ efforts may also indicate the complexity of ordering food through apps. Tak and Panwar [38] found that effort expectancy positively affects behavioral intentions toward using shopping applications. The outcomes revealed that people perceive shopping apps as convenient, effective, and easy to use, enabling them to save time and make purchases swiftly [38].

H2: 

The farmers’/retailers’ effort expectancy affects behavior and intention toward using a circular economy platform.

2.2.3. Social Influence (SI)

According to the research findings, social influence refers to how much a user values the opinions of individuals who suggest they adopt a specific technology [29]. Beza et al. [27] assert that close family or friends affect a person’s willingness to use technology through social influence. According to previous research, the study aimed to examine the relationship between social influence and behavioral intention, which showed a significant relationship between these two variables [31]. Additionally, research has indicated that social influence plays a role in predicting usage patterns in various domains, such as accounting digital platforms in Romania, smartphone apps for diet management, and mobile health services [39,40,41]. Therefore, farmers or retailers in the agricultural industry are most likely to adopt new technologies if they receive recommendations from peers, neighbors, advisors, or trusted ones.

H3: 

Social influence affects the farmers’/sellers’ behavior toward using a circular economy platform.

2.2.4. Facilitating Condition (FC)

Facilitating Conditions refer to being aware of the resources and support available to users of technologies, such as software systems and the technology specialists from collaborating platform providers [29]. It involves the individual’s confidence in the organizational infrastructure assistance during system usage. Previous research articles [42,43,44] on the demand for platforms have shown that user perception significantly influences their intention to use technology due to convenience. The organization should establish a system for both platforms. Sometimes, farmers may require assistance in using or have queries about the platform. Therefore, having an administrator who can assist farmers with platform issues is essential. Alongside this, external resources, including acquiring skills, knowledge, and professional techniques, are also necessary [28].

H4: 

Facilitating conditions affect the farmers’/sellers’ behavioral intention to use a circular economy platform.

2.2.5. Hedonic Motivation (HM)

Hedonic motivation refers to the joy and satisfaction derived from using devices [9]. It plays a role in determining users’ acceptance of technology, as highlighted by Brown and Venkatesh [9]. Previous studies have shown that, when users experience pleasure and happiness while using technologies, they are more likely to continue using them [45,46]. Therefore, using digital platforms to sell agricultural industry byproducts is expected to bring users pleasure and lead to the users’ acceptance.

H5: 

Hedonic motivation affects the farmers’/sellers’ behavioral intention to use a circular economy platform.

2.2.6. Price Value (PV)

The concept of price value entails weighing the benefits against the costs associated with using a system or technology. Users are more inclined to engage with platforms when they perceive the benefits to outweigh the costs [47]. The consideration of implications significantly influences customers’ willingness to adopt and accept new technologies [9]. Various studies indicate that price value positively influences technology adoption [48,49,50]. This study highlights how returns from platform usage enhance user revenue, contributing to increased acceptance and utilization of technologies. This is aligned with Shaw and Sergueeva [51], who found that price value impacts consumer intentions in mobile commerce settings. Additionally, Singh and Srivastava [15] discovered that the sample group in their study emphasized the significance of fees such as entrance fees, membership fees, and transaction charges. Thus, if compared to transactions through other avenues, excessively high fees could potentially deter consumers from utilizing this service. This implies that the service providers charge a price, and the user receives a satisfactory outcome, potentially enhancing user satisfaction and intention to use digital platforms. We therefore propose the following hypothesis:

H6: 

Price value affects the farmers’/retailers’ behavioral intention to use a circular economy platform.

2.2.7. Habit

Limayem et al. [52] argue that habit is the natural inclination to act based on past learning and practice, becoming a personal habit. According to Venkatesh et al. [9], the utilization of prior experience is considered a precondition for the effect of habits on technology usage, as individuals may develop different levels. Therefore, experience is expected to be effective in exploring the relationship between habit and intention in behavior. In the context of this study, habit means that the utilization of digital platforms as a primary instrument for farmers and sellers to trade agricultural waste is suggestive of their habitual practices. This repetitive occurrence assists individuals in developing a routine of utilizing the platform to sell goods and automate actions. Based on the findings of Tam et al. [53], the study recommended that service providers should develop/update the functionality of their mobile applications to encourage continued usage; this may contribute to an increase in users’ habits, followed by the user’s higher behavioral intentions.

H7: 

Habits affect farmers’ and sellers’ behavioral intentions to use a circular economy platform.

2.2.8. Trust

Trust is a factor in various fields; Parasuraman, Zeithaml, and Berry [54] define it as a positive belief in reliability, credibility, and honesty. This implies that consumers trust service providers who prioritize interests and avoid exploiting them. According to Mayer et al. [55], clients’ trust in interactions with trustees grows based on their perception of trust. Trust levels are influenced by an individual’s personality and are relatively stable based on their initial social practices. In this realm, the importance of trust in technology adoption is increasingly recognized. Previous studies examining transactions, food delivery apps, and mobile messaging services have highlighted the critical role of trust in shaping users’ intentions to utilize technology [27,56,57,58]. Consequently, successful online trading hinges on users’ confidence in the system’s functionality. Meeting user expectations leads to satisfaction levels and reinforces user–provider relationships.

H8: 

Trust affects the farmers’/sellers’ behavioral intention to use a circular economy platform.

2.2.9. Privacy

The issue of privacy is becoming increasingly important in the realms of marketing and business. When conducting transactions online, consumers often worry about the privacy implications associated with using social networks [59]. Previous studies found that concerns about privacy in online shopping are frequently tied to risks linked to sharing personal information such as date of birth and identification card [60,61], which holds significant importance for the e-commerce sector [62,63]. Privacy apprehensions also extend to media platforms with intricate features and advanced functionalities as well as the collection, storage, and utilization of personal data [64]. This aligns with research by Mutambik et al. [65] on how consumer privacy concerns impact the adoption of social media platforms. Their study revealed that disregarding these principles leads to privacy safeguards negatively affecting user behavior. Meanwhile, Merhi et al. and Mou et al. [66,67] found that perceived privacy positively influenced the behavioral intention of mobile banking usage, which is that engaging in transactions via online platforms has inherent risks associated with acquiring and safeguarding personal data. Furthermore, Trivedi and Yadav [68] conclude that third-party data privacy guarantees will increase online shopping intentions. Therefore, if the service provider clearly specifies a policy to protect privacy and data security for users. This will lead to increased adoption of technology and increased intention to use digital platforms [66].

H9: 

Privacy affects farmers’ and sellers’ behavioral intentions to use a circular economy platform.

2.2.10. Behavioral Intention

As stated by Venkatesh et al. [9,29], behavioral intention (BI) represents a user’s readiness to use and the likelihood of consumers accepting and displaying a tendency to adopt new technologies in the future. Behavioral intention is commonly employed to forecast user’s decisions regarding technology usage [9,29]. The perceived usefulness of utilizing platforms pertains to how employing technology systems can enhance work efficiency, provide clear information, streamline work processes, and build trust. This may also result in acceptance of the utilized technology. Therefore, the relationship between independent factors and behavioral intention indicates customers’ intention to utilize online platforms, highlighting the strength of these relationships [47].

2.2.11. Moderating Effect of Users’ Experience: Farmers vs. Agi-Food Retailers

The research study has chosen the Unified Theory of Acceptance and Use of Technology 2 (UTAUT2) as its theoretical framework. This section delves into how user experience influences technology usage. Exploring the connection between farmers’ and sellers’ awareness and adoption of technology and users’ perceptions of digital platforms opens up an interesting avenue for investigation. How well individuals can adapt to their tasks could also impact their experiences in various professions. The ease of use of technology is also influenced by an individual’s experience and ability to learn [69]. However, many farmers face limitations in participating in online markets, which reduces their bargaining power and options for selling at fair prices [70]. Additionally, most farmers lack business acumen, focusing solely on production activities that mirror those of their peers without considering market demands. From the study of farmers’ behavioral intentions towards accepting the use of technology [71], it was found that most farmers know about digital platforms from their close friends and co-workers; perceived usefulness and perceived ease of use of technology will drive them to become interested in technological innovation [71,72]. Additionally, Tajvidi and Tajvidi [73] delved into the core aspects of cyber entrepreneurship in today’s tech world with a specific focus on the impact of the COVID-19 pandemic. Their research revealed that entrepreneurs are increasingly comfortable utilizing platforms. Presently, sellers are also becoming more adept at using platforms. However, sellers believe that leveraging tools can enhance their productivity, streamline their workflow, and simplify their processes [74], which is consistent with the findings of Kumar et al. [75], who developed a framework to assess sellers experiences with e-commerce platforms. The research findings indicated that aspects such as registration, product display, pricing flexibility, and seller assistance play a role in the varying seller experiences across online marketplaces [75,76,77]. Consequently, it was deduced that professions influence user interactions, serving as a crucial element impacting moderation within the UTAUT2 model and showcasing an innovative academic aspect in this research. Moreover, the study’s model introduced trust and privacy as elements reinforcing the strong correlation between behavioral intent and technology adoption for digital waste trading platforms.

3. Research Methodology and Data Analysis

Many articles use the SEM approach in the context of behavioral research in food and agriculture [8,12,78]. Most of the results indicate the relationship between factors affecting the behavioral intentions of waste management among stakeholders such as farmers, retailers, and consumers [79,80,81]. Previous studies have also shown that SEM can help identify factors contributing to agricultural waste by modeling the relationship between economic and social factors, the environment, and various technologies [82]. Therefore, we used structural equation modeling (SEM) methods to examine the relationships among the variables. Figure 2 demonstrates the research’s conceptual framework derived from the literature review and hypothesis development based on the modified UTAUT2 model according to the literature review. Next, Figure 3 provides an overview of the SEM approach for this research, which is detailed in the next section.

3.1. Sampling and Data Collection

To calculate the sample size for confirmatory factor analysis (CFA), Hair et al. [11] advised against adopting a specific criterion due to the variability and complexity inherent in different research contexts. A standardized approach may not sufficiently accommodate the unique characteristics of individual studies, such as the number of constructs, model complexity, or anticipated effect sizes, all of which can influence the required sample size for reliable and valid CFA results [83]. Tabachnick and Fidell [84] note that the stability of confirmatory factor analysis (CFA) could be affected by the number of samples. A sample size of 200 should be used when there are less than 7 constructs in the structural model [85]. Wolf et al. [86] also recommended a minimum sample size of 200 when analyzing a model comprising 5 to 10 constructs. Therefore, the researchers aimed to gather by employing a structured questionnaire with a minimum sample size of 400 participants, which helped to improve the significance of the study.

Farmers and fruit and vegetable retailers in the northeastern region of Thailand were surveyed as part of the study. Five provinces (Nakhon Ratchasima, Ubon Ratchathani, Khon Kaen, Buriram, and Udon Thani) were chosen to represent 20 provinces in the area because the Northeastern holds 46.6 percent of the total 7.1 million workers in the agricultural sector of Thailand (National Statistical Office, 2022). By using Calculator.net, considering a population of 3.3 million laborers in the northeastern region, a sample size of 385 was deemed appropriate with a confidence level of 95% and a margin of error of 5%. These provinces were selected due to their population size for easier data collection. In this research, a minimum sampling requirement of 100 respondents (50 farmers and 50 retailers) was set in each province, making a planned sample of n = 500. As a result, the research team was able to gather the data from 570 respondents, exceeding their expectations.

In the data collection process, a structured questionnaire with two main sections (demographic information and user perceptions) was used to combine quota and purposive sampling techniques (see Appendix A). The demographic questionnaire contained three questions about age, gender, income, and occupation. The research questions in the user perception part are based on UTAUT2. The constructs were modified from Venkatesh et al. and Alam et al. [9,87]. The five-point Likert scale was used to measure each variable, ranging from “1” to “5”, i.e., from strongly disagree to strongly agree.

3.2. Data Analysis

Due to the large dataset we collected, containing data which may be from the same source, there is a risk that common method variance (CMV) may occur. According to Podsakov et al. [88], common method variance occurs when data predict variables and those standard variables are measured using the same measurement tool, sourced from the same data origin, and pose questions in a similar manner towards another question on the same level [89]. This situation, which becomes evident when utilizing the Likert scale, can lead to inaccurate outcomes in data analysis. Before scrutinizing the data, we use common method variance (CMV) to examine the variance. According to the results, there was a cumulative variation of 49.325 percent, which is less than the threshold of 50 percent.

The present study applied demographic data, including variables such as gender, age, and income, to perform a multivariate clustering analysis. A t-test was conducted between the two clusters to determine differences between clusters with different solutions. The statistical significance of the observed differences between clusters was determined by conducting a sequence of chi-square tests to determine their relation to demographic and psychographic segmentation. The present study utilized the cluster analysis approach. In the beginning, hierarchical cluster analysis is used to create clusters by maximizing both the similarities and differences within groups. Second, cluster centroid analysis is used to confirm the hierarchy of cluster solutions. In the end, the squared Euclidean distance reaches a minimum distance between observed and estimated values.

In this study, the data were analyzed using the structural equation modeling (SEM) technique with the AMOS statistics program. SEM employs statistical methods such as path analysis, confirmatory factor analysis (CFA), causal modeling with latent variables, analysis of variance, and multiple linear regression [90]. Structural equation modeling (SEM) was used to evaluate the model in three steps. At first, it involves conducting a CFA (confirmatory factor analysis) to establish the factor structures of the measurement models and assess data reliability and validity, including convergent and discriminant validity as well as goodness of fit. Next, structural equation modeling (SEM) is used to examine the goodness of fit of the structural model and evaluate the structural relationship between independent and dependent variables. In the last step of the SEM process, multigroup analysis (MGA) is utilized to examine the impact of differences in experiences between the two sample groups on the structural relationships proposed by the hypotheses. This step involves measuring invariances across sample groups to determine the extent to which dimensional characteristics of the data are consistent across different conditions [91,92]. The user’s occupation serves as a moderating factor for differences between groups.

4. Results

4.1. Multivariate Demographic Segmentation

Table 2. Descriptive statistics for demographic profiles.

		Farmers		Fruit and Vegetable Retailers		Total		Significant
Demographic Profile	Measure	n	%	n	%	n	%	Chi-Square Test
Segment size		279	48.9	291	51.1	570	100
Gender	Man	98	35.1	117	40.2	215	37.7	***
	Woman	181	64.9	174	59.8	355	62.3
Age	GenZ	19	6.8	23	7.9	42	7.4	***
	GenY	60	21.5	79	27.1	139	24.4
	GenX	61	21.9	91	31.3	152	26.7
	Baby Boomer	139	49.8	98	33.7	237	41.6
Income	Less than THB 15,000	197	70.6	132	45.4	329	57.7	***
	THB 15,001–20,000	57	20.4	89	30.6	146	25.6
	THB 20,001–25,000	12	4.3	39	13.4	51	8.9
	THB 25,001–30,000	6	2.2	16	5.5	22	3.9
	More than THB 30,000	7	2.5	15	5.2	22	3.9

Note: *** significant at <0.01; USD 1 dollar is roughly THB 37.90.

presents the results of the personal data analysis from respondents. The study found that 279 of the 570 participants surveyed identified themselves as farmers, accounting for 48.9%, while 291 identified themselves as fruit and vegetable sellers, accounting for 51.1%. There were 335 women (62.3%) and 215 men (37.7%). Moreover, the most common age range of respondents was between 57 and 76 years old (Baby Boomers), with 237 respondents representing 41.6%. The monthly income of both groups is less than 15,000 THB. Furthermore, the statistical analysis using the Chi-square test reveals that demographic factors such as gender, age, and income have a significant role in classifying individuals into the categories of farmers and sellers.

The behavioral characteristics of the two groups were compared using a t-test analysis based on the data shown in Table 2. In most cases, the mean scores of retailers are higher than farmers’ perceptions. The results in

Table 3. Result of the independent sample t-test.

Construct	Measure	Farmer		Fruit and Vegetable Retailer		Mean Diff	t	t-Test
		Mean	S.D	Mean	S.D
Performance Expectancy	PE1	3.74	0.87	3.80	0.78	−0.06	−0.869	0.385
	PE2	3.75	0.84	3.75	0.78	0.00	−0.039	0.386
	PE3	3.79	0.85	3.84	0.80	−0.05	−0.744	0.969
Effort Expectancy	EE1	3.67	0.91	3.75	0.83	−0.09	−1.182	0.238
	EE2	3.69	0.96	3.87	0.86	−0.18	−2.387	**
	EE3	3.70	0.92	3.86	0.85	−0.17	−2.272	*
Social Influence	SI1	3.53	0.90	3.75	0.83	−0.03	−0.437	0.662
	SI2	3.62	0.81	3.64	0.77	−0.02	−0.328	0.743
	SI3	3.59	0.85	3.64	0.80	−0.05	−0.764	0.445
Facilitating Condition	FC1	3.62	0.85	3.71	0.78	−0.10	−1.395	0.164
	FC2	3.69	0.98	3.90	0.72	−0.22	−3.016	**
	FC3	3.70	0.95	3.90	0.79	−0.20	−2.774	***
Hedonic Motivation	HM1	3.55	0.88	3.55	0.84	0.00	−0.067	0.946
	HM2	3.52	0.82	3.56	0.82	−0.05	−0.667	0.505
	HM3	3.66	0.86	3.80	0.84	−0.14	−1.966	*
Price Value	PV1	3.64	0.94	3.93	0.86	−0.30	−2.455	***
	PV2	3.64	0.92	3.82	0.85	−0.18	−1.574	*
	PV3	3.68	0.94	3.80	0.87	−0.12	−2.668	0.116
Habit	HB1	3.50	0.87	3.69	0.80	−0.19	−2.668	**
	HB2	3.51	0.89	3.50	0.87	0.01	0.146	0.884
	HB3	3.57	0.90	3.60	0.85	−0.02	−0.328	0.774
Trust	TR1	3.66	0.90	3.96	0.87	−0.30	−4.012	***
	TR2	3.61	0.96	3.78	0.83	−0.17	−2.294	*
	TR3	3.71	0.91	3.82	0.84	−0.11	−1.515	0.130
Privacy	PR1	3.64	0.93	3.77	0.95	−0.13	−1.657	0.098
	PR2	3.64	0.95	3.87	0.92	−0.23	−3.002	**
	PR3	3.73	0.95	3.81	0.91	−0.08	−0.986	0.325
Behavior Intention	BI1	3.62	0.93	3.82	0.81	−0.20	−2.769	**
	BI2	3.60	0.95	3.71	0.77	−0.11	−1.550	0.122
	BI3	3.66	0.91	3.74	0.81	−0.08	−1.108	0.268

Note: *** denotes significant at <0.001, ** at <0.01, and * at <0.05.

indicate that fruit and vegetable retailers experienced a higher level of expected performance than farmers when using the digital platform. Moreover, retailers have a higher level of expectations for their efforts on the platform than farmers. The finding also suggested that retailers exhibit more social influence than farmers. Retailers also were expecting the platform provider to assist users in solving issues while using the platform, compared to farmers. The perception of hedonic motivation among users indicates that retailers showed a higher level of hedonic motivation than farmers, while retailers also have a higher level of perceived price value-to-benefit from using the platform than farmers. The level of perception about habits while using the platform was slightly higher in retailers than in farmers. Additionally, the respondents reported having a greater level of perceived trust and privacy concerns. To conclude, it can be demonstrated that retailers have a higher level of perceived behavioral intention to utilize platforms than farmers.

4.2. Measurement Model (CFA)

Confirmatory factor analysis (CFA) was performed to scrutinize the measurement model. CFA deals specifically with measurement models, which is the relationship between observed measures (indicators) and latent variables (factors) [90]. Measurement model results can provide convincing evidence of the convergent and discriminant validity of theoretical constructs.

4.2.1. The Goodness of Fit (GOF)

The Goodness of Fit (GOF) measures and their corresponding thresholds are presented in

Table 4. The goodness of fit of the measurement model.

Fit Index	Value	Threshold	Assessment
p-value	0.00		Acceptable for complex model
CMIN/df	2.958	<3.000	Passed
TLI	0.954	>0.90	Passed
CFI	0.963	>0.90	Passed
IFI	0.963	>0.90	Passed
RMSEA	0.059	<0.10	Passed

Note: CMIN/df = Chi-square/degree of freedom; CFI = comparative fit index; IFI = incremental fit index; TLI = Tucker–Lewis index; RMSEA = root mean square error approximation.

. Based on research by Cheung and Rensvold [91], which suggested the criteria for goodness of fit that supported this structural model, the findings were positive. Specifically, the comparative fit index (CFI) achieved a value of 0.963, the Tucker–Lewis index (TLI) reached 0.954, the incremental fit index (IFI) obtained a score of 0.963, and the RMSEA was calculated to be 0.059.

4.2.2. Convergent Validity

Convergent validity is a subtype of construct validity used to test those two measures of related constructs. The model’s external validation relies on the assessment of convergent validity by examining Cronbach’s Alpha, composite reliability (CR), and the latent variables’ total confidence value, also known as the average variance extracted (AVE), is calculated. The thresholds for Cronbach’s Alpha, CR, and AVE are 0.70 and 0.50, respectively. From the conceptual framework consisting of 10 variables, the researcher examined the quality of the external model using convergent validity. This assessment indicated coefficient loading for each question.

Table 5. Convergent validity.

Construct	Indicator	Loading	p-Value	Cronbach’s Alpha	AVE	C.R.
Performance Expectancy	PE1	0.855	***	0.864	0.763	0.906
	PE2	0.890	***
	PE3	0.875	***
Effort Expectancy	EE1	0.828	***	0.900	0.782	0.915
	EE2	0.919	***
	EE3	0.903	***
Social Influence	SI1	0.897	***	0.891	0.793	0.920
	SI2	0.920	***
	SI3	0.853	***
Facilitating Condition	FC1	0.873	***	0.867	0.805	0.925
	FC2	0.907	***
	FC3	0.911	***
Hedonic Motivation	HM1	0.816	***	0.855	0.728	0.889
	HM2	0.810	***
	HM3	0.928	***
Price Value	PV1	0.890	***	0.929	0.825	0.934
	PV2	0.935	***
	PV3	0.900	***
Habit	HB1	0.850	***	0.898	0.761	0.905
	HB2	0.873	***
	HB3	0.893	***
Trust	TR1	0.894	***	0.937	0.848	0.944
	TR2	0.932	***
	TR3	0.936	***
Privacy	PR1	0.932	***	0.942	0.859	0.948
	PR2	0.934	***
	PR3	0.915	***
Behavior Intention	BI1	0.928	***	0.935	0.821	0.932
	BI2	0.892	***
	BI3	0.898	***

Note: AVE = average variance extracted, CR = composite validity, *** significant < 0.001.

demonstrates that the constructs of PE (performance expectancy), EE (effort expectancy), SI (social influence), FC (facilitating condition), HM (hedonic motivation), PV (price value), HB (habit), TR (trust), PR (privacy), and BI (behavioral intention) have successfully met the criteria for convergent validity. This is evident from comparing the calculated measures with their respective thresholds.

4.2.3. Discriminant Validity

Discriminant validity is another type of construct validity used to prove that measurements that should not be related are unrelated and to ensure that non-overlap factors do not overlap. The discriminant validity can be measured by the cross-loading of indicators, the Fornell–Larcker criterion, and the heterotrait–monotrait (HTMT) ratio of correlations [93].

Table 6. Discriminant validity.

Fornell–Larcker Criterion
Construct	PE	EE	SI	FC	HM	PV	HB	TR	PR	BI
PE	0.873
EE	0.746	0.884
SI	0.700	0.660	0.891
FC	0.704	0.746	0.616	0.897
HM	0.688	0.681	0.687	0.731	0.853
PV	0.700	0.755	0.616	0.769	0.771	0.908
HB	0.662	0.557	0.722	0.596	0.711	0.638	0.872
TR	0.693	0.758	0.662	0.736	0.775	0.838	0.673	0.921
PR	0.630	0.721	0.601	0.712	0.733	0.787	0.557	0.849	0.927
BI	0.703	0.685	0.726	0.751	0.777	0.751	0.733	0.760	0.776	0.906
HTMT Ratio Approach
	BI	PR	TR	HB	PV	HM	FC	SI	EE	PE
BI
PR	0.770
TR	0.754	0.849
HB	0.727	0.558	0.673
PV	0.745	0.787	0.838	0.638
HM	0.745	0.708	0.749	0.687	0.745
FC	0.753	0.720	0.744	0.602	0.778	0.714
SI	0.732	0.612	0.674	0.735	0.626	0.675	0.634
EE	0.680	0.722	0.759	0.557	0.756	0.658	0.755	0.672
PE	0.697	0.630	0.693	0.662	0.700	0.664	0.712	0.712	0.746

Note: PE = performance expectancy; EE = effort expectancy; SI = social influence; FC = facilitating condition; HM = hedonic motivation; PV = price value; HB = habit; TR = trust; PR = privacy; BI = behavioral intention.

shows the discriminant validity by comparing the square roots AVEs (values on the diagonal of Fornell-Larcker Criterion) with the correlations. The results show that all the constructs passed this validity check. Additionally, HTMT is used to assess discrete accuracy by pairwise comparisons of latent variables. The results can be calculated using the formula presented by Henseler in 2015. The value of the HTMT ratio must be less than 0.85 or 0.90 [94]. However, the results from the HTMT analysis demonstrated that all the values passed the HTMT threshold of 0.85.

4.3. Structural Model

After the measurement model was validated, we proceeded to develop the structure model by connecting all the constructs and running through SEM. The structural model shows how each construct is related to each other. Structural equation modeling (SEM) allows the researchers to check and examine the overall fit by generating the goodness of fit (GOF). Hair et al. [95] suggested that measurement indicators should have factor loadings of at least 0.50, including a preference for values greater than 0.70 for an acceptable measurement. The construct reliability (CR) must be greater than 0.70, and the average variance extracted (AVE) for each construct should be equal to or greater than 0.50. However,

Table 7. The goodness of fit of the structural model.

Fit Index	Value	Threshold	Assessment
p-value	0.00		Acceptable for complex model
CMIN/df	2.958	<3.000	Passed
TLI	0.954	>0.90	Passed
CFI	0.963	>0.90	Passed
IFI	0.963	>0.90	Passed
RMSEA	0.059	<0.10	Passed

Note: CMIN/df = Chi-square/degree of freedom; CFI = comparative fit index; IFI = incremental fit index; TLI = Tucker–Lewis index; RMSEA = root mean square error approximation.

shows the results from the goodness of fit of the structural model, CMIN/df, TLI, CFI, IFI, and RMSEA [96]. All the results passed the threshold.

In the model that was hypothesized, five paths were statistically significant (see

Table 8. Hypothesis test results from the structural model.

Path	Relationships	Standardized Estimate	p-Value	Result
H1	PE → BI	0.049	0.315	Rejected
H2	EE → BI	−0.047	0.371	Rejected
H3	SI → BI	0.161	***	Supported
H4	FC → BI	0.192	***	Supported
H5	HM → BI	0.124	0.029 **	Supported
H6	PV → BI	0.054	0.366	Rejected
H7	HB → BI	0.233	***	Supported
H8	TR → BI	−0.087	0.196	Rejected
H9	PR → BI	0.356	***	Supported

Note: *** denotes significant at ≤0.001 and ** at ≤0.05.

). However, when compared to the proposed hypotheses, it seems that performance expectation (H1) has no impact on users’ behavioral intention. H2 has also been rejected, demonstrating that behavioral intention to use digital platforms for waste trading was not influenced by effort expectancy. H3 was supported by the standardized factor loading of 0.161. The results from H3 show that social influence has a significant impact on users’ intentions to use waste trading platforms. According to H4, the facilitating condition positively affects users’ behavioral intention on the waste trading platform, with the factor loading at 0.192. Hypothesis 5 was also supported; hedonic motivation plays a role in users’ intention to use waste trading platforms. Hypothesis 6 was not supported. This means that the users’ intention to use the waste trading platform is not influenced by the price value. H7 is another accepted hypothesis, indicating that habit positively affects users’ behavioral intention to use waste trading platforms. H8 was also rejected. The result indicates that trust in online platforms does not impact users’ behavior intention. H9 suggests that privacy has an impact on users’ intentions when they interact with an agricultural waste trading platform.

4.4. Multigroup Moderation Analysis

This research utilized a confirmatory factor analysis (CFA) to investigate measurement invariance (MI). This analysis aimed to assess whether the relationships between two or more variables remained constant or varied across groups specifically categorized by age, gender, and experience. Three levels of invariance (configural, metric, and scalar) were evaluated using the measurement invariance approach. The comparative fit index (CFI), incremental fit index (IFI), and Tucker–Lewis index (TLI) values for invariance, metric invariance, and scalar invariance exceeded the threshold of 0.09. According to the study of Yuan and Chan (2016), these findings met the criteria for fit (see

Table 9. Measurement invariance.

Fit Index	Configural Invariance (Unconstrained Model)	Metric Invariance (Equal Factor Loading)	Scalar Invariance (Equal Intercept)	Threshold
p-value	0.000	0.000	0.000
CMIN/df	2.744	2.716	2.752	<3.000
CFI	0.937	0.936	0.932	>0.900
IFI	0.937	0.936	0.932	>0.900
TLI	0.922	0.924	0.922	>0.900
RMSEA	0.055	0.055	0.056	<0.10

Note: CFI = comparative fit index; IFI = incremental fit index; TLI = Tucker–Lewis index; RMSEA = root mean square error approximation.

) [92].

Table 10. The goodness of fit of the multigroup structural model.

Fit Index	Value	Threshold	Assessment
p-value	0.00		Acceptable for complex model
CMIN/df	2.744	<3.000	Passed
TLI	0.922	>0.90	Passed
CFI	0.937	>0.90	Passed
IFI	0.937	>0.90	Passed
RMSEA	0.055	<0.10	Passed

Note: CMIN/df = Chi-square/degree of freedom; CFI = comparative fit index; IFI = incremental fit index; TLI = Tucker–Lewis index; RMSEA = root mean square error approximation.

presents the measurement of goodness of fit (GOF) and the thresholds that apply to the multigroup structural model. The comparative fit index (CFI) reached a value of 0.937. Similarly, the incremental fit index (IFI) also scored 0.937, which meets the acceptable threshold. The Tucker–Lewis index (TLI) has a value of 0.922, which appears to be within a range. According to research data from this table, they are all above 0.900 [97], and the root mean square error of approximation (RMSEA) was calculated at 0.055, indicating a fit. All of these results met the criteria.

By comparing factor loadings between farmers and agricultural food retailers, a Z-test was performed with crucial ratio differences. The z-test can be used to identify a variety of critical ratios to analyze differences in the structural model. If the critical ratio is higher than 1.96, then the factor loading between groups would be different. The results presented in

Table 11. Multigroup analysis (MGA) results.

Path	Relationships	Farmers (1)			Fruit and Vegetable Retailers (2)			Critical Ratio Difference	Threshold
		Std. Est	p-Value	Result	Std. Est	p-Value	Result
H1	PE → BI	−0.106	0.107	Rejected	0.019	0.854	Rejected	1.019	|1.96|
H2	EE → BI	−0.073	0.332	Rejected	−0.008	0.930	Rejected	0.525	|1.96|
H3	SI → BI	0.015	0.798	Rejected	0.312	***	Supported	2.588 *	|1.96|
H4	FC → BI	0.211	***	Supported	0.275	0.006 **	Supported	0.494	|1.96|
H5	HM → BI	0.152	0.038 *	Supported	0.065	0.453	Rejected	−0.802	|1.96|
H6	PV → BI	0.110	0.195	Rejected	0.092	0.603	Rejected	−0.545	|1.96|
H7	HB → BI	0.302	***	Supported	0.146	0.030 *	Supported	−1.596	|1.96|
H8	TR → BI	0.027	0.806	Rejected	−0.233	0.012 *	Supported	−1.744	|1.96|
H9	PR → BI	0.376	***	Supported	0.402	***	Supported	−0.514	|1.96|

Note: *** denotes significant at ≤0.001, ** at ≤0.01, and * at ≤0.05. PE = performance expectancy; EE = effort expectancy; SI = social influence; FC = facilitating condition; HM = hedonic motivation; PV = price value; HB = habit; TR = trust; PR = privacy; BI = behavioral intention. (1) refers to hypotheses for farmers, (2) refers to hypotheses for fruit and vegetable retailers.

indicated that several hypotheses related to H1 (farmer) and (retailer), H2 (farmer) and (retailer), H3 (farmer), H5 (retailer), H6 (farmer) and (retailer), and H8 (farmer) were not statistically significant as their p-values were below 0.05, 0.01, or 0.001 thresholds. The absence of significant correlation implies that when the data are divided into two user groups, factors such as performance expectation, effort expectancy, facilitating conditions, and hedonic motivation do not significantly influence the behavioral intent to utilize an agriculture waste trading platform. However, H3 (retailer), H4 (farmer) and (retailer), H5 (farmer), H7 (farmer) and (retailer), and H8 (retailer) were found to have statistically significant results with p-values below the predefined significance threshold. The critical ratio of H8 is above the ratio threshold (|2.588| > |1.96|), which shows a significant path difference in the critical ratio. According to the statistical analysis, there is a significant difference between farmers and fruit and vegetable retailers, as evidenced by the finding of hypothesis 3.

5. Discussion

The UTAUT2 framework was used in the present study to examine the factors that influence farmers’ and retailers’ behavior intention when embracing waste trading platforms. Moreover, this study added two additional variables, namely trust and privacy, to the existing UTAUT2 model to improve understanding of users’ behavioral intentions. The data analyzed in this research helps to shed light on the factors that drive farmers and retailers toward adopting and utilizing technology. Developing agricultural waste management is a beneficial opportunity for technological advancement and innovative solutions in the field.

The statistical results validate that users’ behavioral intention is influenced by social influence, facilitating conditions, hedonic motivation, habit, and privacy. These factors are crucial for researchers aiming to encourage farmers and retailers to utilize waste trading digital platforms.

H1, H1(1), and H1(2): 

The farmers’ and retailers’ performance expectations affect behavioral intentions to use a digital platform for the circular economy.

According to the results (see Table 8), there is no connection between performance expectations and behavioral intentions, as shown by the p-value of 0.315, which is higher than the standard threshold of 0.05. It can be concluded that performance expectancy for waste trading platforms did not influence the users’ intentions. As it appears, the respondents did not provide their opinions on how they expected the platform to perform in managing waste. Additionally, they could have been unaware of the benefits of using this platform, such as improved efficiency and simplified processes in their workplace. Thus, respondents felt that using digital platforms did not improve their work productivity. The results presented contradict those of Beza et al. [27] who observed that farmers’ behavioral intentions were positively influenced by their performance expectations when using SMS for farm information.

Moreover, after conducting a multigroup analysis (MGA) (see Table 11 and Figure 4), it was revealed that there was no correlation between performance expectations and behavioral intentions among both farmers (p-value = 0.107) and retailers (p-value = 0.854). This suggests that neither group’s performance expectations had an influence on their intention to use digital platforms, contradicting previous research conducted by Molina-Maturano et al. [34].

H2, H2(1), and H2(2): 

The farmers’/retailers’ effort expectancy affects behavior and intention toward using a circular economy platform.

The structural equation modeling (SEM) analysis showed that there was no significant correlation between effort expectancy and behavioral intention when using a waste trading platform (p-value 0.371 > 0.05). This result aligns with the finding given by Najib et al. [98], which stated that anticipating business effort has no effect on behavior intention and adoption. It seems that the perception of the waste trading platforms as inconvenient services among the sample groups could be attributed to their limited familiarity with technology in their work processes. Furthermore, the complex interaction between service providers and users can be challenging for respondents to understand. Therefore, the level of effort expectancy among the sample groups in this study did not have any impact on their behavioral intentions toward using a digital platform for waste trading.

According to the multigroup analysis (MGA), the finding indicates that effort expectancy has no significant impact on behavioral intention for both (farmers and retailers). The fact that farmers’ p-value = 0.332 > 0.05 and retailers’ p-value = 0.930 > 0.05 are not significant demonstrates evidence for this statement.

H3, H3(1), and H3(2): 

Social influence affects the farmers’/sellers’ behavior toward using a circular economy platform.

This recent study discovered a significant relationship between social influence and behavioral intention. Previous research by Omar et al. [99] supports this finding. This implies that users will have a higher level of trust. If they have confidence in the advice and information from their peers or external sources, they are more likely to embrace a platform that promotes a circular economy.

Additionally, the results of the multigroup analysis (MGA) indicate that social influence did not significantly impact farmers’ intentions (loading = 0.015, p-value 0.789 < 0.05) but highly impacted retailers’ intentions (loading = 0.312, p-value ≤ 0.001). Table 11 reveals a ratio difference of |−2.011|, which highlights significant differences in behavior between farmers and retailers. The findings show that retailers are more interested in social influence when they receive positive feedback from peers, family members, and colleagues about using digital platforms, which leads to an increased intention to utilize digital platforms.

H4, H4(1), and H4(2): 

Facilitating conditions affect the farmers’/sellers’ behavioral intention to use a circular economy platform.

Our discovery revealed that the facilitating conditions had an impact on the users’ behavior and intention to use a digital platform to sell agricultural waste. Previous studies by Omar et al. [99] have shown that facilitating conditions affect farmers’ intentions to adopt agricultural applications.

Based on the results of the MGA analysis, it is evident that facilitating conditions had an influence on the behavioral intentions of both farmers (loading = 0.211, p-value = ***) and retailers (loading = 0.275, p-value = 0.006). Therefore, this discovery suggests that people in these groups need access to resources and technology infrastructure, such as reliable internet connectivity and compatible devices such as iOS and Android systems. By providing convenient access to these resources, users are able to adopt and utilize digital platforms.

H5, H5(1), and H5(2): 

Hedonic motivation affects the farmers’/sellers’ behavioral intention to use a circular economy platform.

The findings of the study show that hedonic incentive in using a platform for selling surplus fruit and vegetables has a positive impact on both behavior and the intention to use it. This aligns with the research by Naruetharadhol et al. [100], which suggests that enjoying the online ticket-purchasing process has an effect on consumers’ willingness to buy air tickets. Essentially, having fun while engaging in trading activities on a platform can boost users’ enthusiasm for utilizing sustainable platforms.

The results from the MGA indicate that hedonic motivation from using the platform influenced farmers’ behavior and intentions significantly (loading = 0.152, p-value = 0.038 < 0.05). The comparison between farmers and agri-food retailers suggests that hedonic motivation plays a more significant role for farmers than retailers.

H6, H6(1), and H6(2):

Price value affects the farmers’/retailers’ behavioral intention to use a circular economy platform.

The results of this study suggest that there is no statistically significant correlation between users’ perceptions of the pricing and their intention to use a digital platform for trading agricultural waste. The statistical analysis resulted in a p-value of 0.366, which is higher than the typical threshold of 0.05 for statistical significance. These findings align with a study by Dhiman et al. [101], indicating that the platform’s price does not strongly influence users’ intention to use it due to factors such as a lack of awareness among users and a preference for offline income sources over online channels. Furthermore, users may also be concerned about additional costs associated with joining the platform, such as registration, operation, and gross profit (GP) fees. This suggests the potential to improve the platform’s pricing value to increase the level of users’ intention to utilize a circular economy digital platform. Therefore, this observation shows a gap in the improvement of platform price value to increase the level of users’ behavioral intention.

The outcomes from the MGA showed that the price value did not have an impact on the behavioral intentions of farmers and fruit and vegetable retailers. The p-values of 0.195 for farmers and 0.603 for retailers indicated this, surpassing the significance level of 0.05. This study reveals that the willingness to utilize platforms remains unaffected by price value, which is in contrast with the results reported by Pienwisetkaew et al. [8].

H7, H7(1), and H7(2): 

Habits affect farmers’ and sellers’ behavioral intentions to use a circular economy platform.

Habit has a significant impact on behavioral intention, as demonstrated by the standardized loading coefficient of 0.233; the p-value is also less than 0.0001, which is equivalent to 0.000. The results derived from the SEM analysis align with research by Widodo et al. [102], indicating that habit significantly influences users’ behavioral intention toward digital wallet adoption. While collecting questionnaires, we provided respondents with information regarding the circular economy digital platform. This online platform becomes a part of users’ online interactions, making them familiar with its features and functionality. As a result, participants believe that will help them develop habits for using a digital platform to sell agricultural products.

The MGA analysis findings indicate that habit has a stronger influence on farmers’ behavioral intention (loading = 0.302) compared to retailers (loading = 0.146). Both segments agree that regular platform usage will improve their familiarity with the platform, resulting in a higher level of user intention.

H8, H8(1), and H8(2): 

Trust affects the farmers’/sellers’ behavioral intention to use a circular economy platform.

The study found no significant correlation between respondents’ trust level and their use of a digital platform for trading agricultural waste (p-value = 0.196 > 0.05); this finding is opposite to those of earlier studies by Merhi et al. [66]. The findings indicated that a level of trust in digital platform service providers plays a crucial role in influencing individuals’ behavioral intentions to utilize digital platforms.

The application of MGA in the research explained that trust substantially affects the behavioral intentions of agri-food sellers. The loading coefficient for trust was found to be −0.233, with a p-value of 0.012, which is lower than the significance level of 0.05. However, there was no statistically significant relationship seen between trust and behavioral intentions among farmers (loading = 0.027, p-value = 0.806 > 0.05).

H9, H9(1), and H9(2): 

Privacy affects farmers’ and sellers’ behavioral intentions to use a circular economy platform.

Privacy has a significant impact on users’ behavioral intentions to use a waste trading platform (loading = 0.356, p-value = ****). The effect of privacy in using digital platforms has been confirmed in the previous work of Alzaidi and Agag [103], who suggest this study examines the influence of privacy concerns on individuals’ behavioral intentions when utilizing social media platforms. According to the survey findings, users are concerned about the ethical handling of their data, compliance with privacy regulations, and clear policies for data collection, storage, and usage. If users feel that their data may be misused for purposes that are contrary to the agricultural waste platforms’ original intent, such as targeted advertising or shared information with unrelated entities, it can lead to a loss of trust and reduce the level of intention.

The MGA analysis revealed that while the privacy factor did not significantly affect users’ intentions toward farmers, it did have an impact on retailers. Retailers may be particularly concerned about safeguarding their business information, such as pricing strategies, customer data, inventory details, and competitive insights.

6. Implications, Limitations, and Future Research Directions

The objective of this study is to examine the various factors that impact farmers’ and fruit and vegetable retailers’ behavioral intentions in adopting a digital platform for trading agricultural waste, transitioning from the linear economy routine to circular economy innovation. Moreover, the aim of this study is to examine and compare the differences in farmers’ and retailers’ intentions toward adopting a digital platform to sell waste agricultural products. The research findings are discussed to reveal benefits in both theoretical and practical implications, as outlined below.

6.1. Theoretical Implication

The authors of this study propose a theoretical and conceptual framework that expands upon the UTAUT2 model, which has been extended by Venkatesh et al. [9], to provide a more thorough understanding of technology acceptance from the user’s perspective. Therefore, this study is crucial in developing a basic theoretical framework that aligns with the user context and identifies important factors in users’ behavioral intentions toward using an agricultural waste trading platform. In addition, Kilani et al. [104] conducted an empirical study on the effectiveness of UTAUT2 to explain the adoption behaviors by using e-wallets in Jordan. Furthermore, a study conducted by Alam et al. [87] highlighted the significance of perceived trust and privacy norms positively related to users’ intentions in using technology. Given the significance of the previous studies in expanding the utilization of UTAUT2, the present research combines two additional variables (including trust and privacy) to align with the specific parameters of this investigation.

Nevertheless, the findings elucidate that social impact, facilitating conditions, and hedonic motivation are significant factors influencing the users’ behavioral intentions in using agricultural waste trading platforms. This finding is consistent with Hassaan et al. and Migliore et al. [105,106]. However, no close relationship was found between behavioral intention and performance expectancy, effort expectancy, and price value, which is similar to the finding of Najib et al. [98]. This paper also modified UTAUT2 by adding the trust variable [104,107]. The research findings align with results from previous investigations, that show that trust positively impacts behavioral intentions [108,109]. Privacy was added as another variable in this study as well. From the results of this study, it was found that behavior intention is affected by perceived privacy, supporting previous research results of Najib et al. [98]. In summary, to clearly understand the factors that influence behavioral intentions and adoption of the platform, the researchers expanded the conceptual framework to appropriately address the factors affecting the intention to use the platform to trade agricultural wastes. This article also demonstrates the heterogeneity in farmers’ and merchants’ behavioral intentions to embrace a circular agriculture technology and decision-making process, which theoretically contribute to the recent circular economy studies based on the SEM approach [8,13,14,110].

6.2. Practical Implications

Due to the potential novelty of this platform in the agriculture business, the results from this research can be used to determine guidelines for developing platform functionalities that may facilitate Thailand’s agri-food sector to transform from the linear to circular economy paradigm. Those functionalities may include designing the platform for ease of use, building user trust, and increasing channels to generate income to be consistent with the critical factors specified in the UTAUT2 model. The development approach that meets the users’ specific needs and addresses users’ concerns would encourage users to adopt a circular economy digital platform to trade waste from agricultural products. This study discusses features affecting the behavioral intention to utilize a circular economy digital platform for fruit and vegetable waste trading across two categories of users using the UTAUT2 model: farmers and agri-food retailers. The findings of this study have significant implications that might be beneficial for platform development in the agriculture market. Examining the factors related to users’ behavioral intentions, SI, FC, HM, HB, TR, and PR appear to be the most important criteria for increasing the level of users’ intention to use a circular economy platform.

Hence, platform developers need to implement measures to enhance awareness of social impacts, including familial, peer, and community influencers. The developers should arrange convenient conditions for users to improve user satisfaction and promote user engagement (e.g., the system offers support for both iOS and Android, a comprehensive manual is available to assist new users in getting started, and an admin available to provide guidance and assistance during the usage of the platform). The results also indicate that hedonic motivation contributes to greater participation in the use of a circular economy digital platform. For this reason, platform developers should design a variety of user-friendly platform interfaces. This will help attract and increase interest among users, especially farmers. Furthermore, the findings suggest that farmers’ behavioral intentions are significantly influenced by familiarity or habit. This means that the platform developers designed the platform to be easy to use and not complicated, which will make farmers familiar with its daily use. Consequently, farmers are more likely to consider utilizing this platform whenever they need to sell their surplus vegetable and fruit waste. The research results also indicate that the farmers’ behavioral intentions are affected by the trustworthiness of the platform. Due to these reasons, platform providers must exhibit transparency, be honest, have appropriate policies during transactions, and focus on the benefits that users will receive. Moreover, it appears that the level of personal data privacy awareness of both farmer and retailer users has a strong positive impact on behavioral intentions to use a circular digital platform. As a result, the platform’s creators must develop an advanced security infrastructure to protect users’ data.

Based on the findings, policymakers should develop targeted marketing campaigns that leverage social influence to increase platform adoption among retailers. Improving platform usability, security, and creating incentives for habitual use are essential for both groups. Additionally, engaging features should be promoted for farmers, who are also motivated by hedonic aspects of the platform. Comprehensive education and training, along with supportive policies and financial incentives, will further enhance the adoption and effective use of these digital platforms in agricultural waste management.

6.3. Limitations and Future Research Directions

The study’s findings are limited to the Northeastern region of Thailand, which may limit their generalizability to broader populations or different geographic areas. In later studies, data collection could be expanded to include different geographic areas, particularly the central region, which are each characterized by distinct cultures and lifestyles. By expanding, farmers and retailers would be able to offer a range of perspectives on how new technologies are viewed and utilized.

The adoption of digital platforms for long-term waste management, as discussed in UTAUT2, has limitations due to its heavy focus on technology acceptance and behavioral intentions. Future research should advance this study, focusing on the factors that influence technological adoption and the sustained intention to use this platform for agricultural waste management over the long term, thereby enriching the understanding of the research findings. Additionally, sustainable innovation utilization requires an understanding of how digital waste management systems can be seamlessly integrated with existing technologies or practices in the agricultural sector. Longitudinal studies tracking user behavior over time may be necessary.

Based on the findings, policymakers should develop targeted marketing campaigns that leverage social influence to increase platform adoption among retailers. Improving platform usability, security, and incentives for habitual use are essential for farmers and retailers. Additionally, engaging features should be promoted for farmers, who are also motivated by the hedonic aspects of the platform. Comprehensive education, training, supportive policies, and financial incentives will further enhance the adoption and effective use of these digital platforms in agricultural waste management.

7. Conclusions

This study aimed to investigate the factors influencing farmers’ and retailers’ intentions to adopt a digital platform for selling agricultural waste products. Our findings revealed that various factors significantly affect the behavioral intentions of users engaging with waste trading platforms. Factors such as social influence (SI), facilitating condition (FC), hedonic motivation (HM), habit (HB), and concern about privacy were identified as significant in this context. Particularly noteworthy is the impact of influence on users’ decisions within these two user groups; participants are more likely to adopt a circular economy digital platform when influenced by people they trust and respect, such as family members, colleagues, and people with the same occupation. Additionally, the study suggests that users from both groups anticipate platform providers will prioritize user features that address diverse needs during their engagement with the platform, known as facilitating conditions. This includes ensuring compatibility with devices, such as iOS and Android, reliable internet connectivity, and providing guidance on effective utilization of the platform. Furthermore, this research has demonstrated that farmers’ behavior and intentions are influenced by hedonic motivation. The finding also shows that both users expected to improve their familiarity with a waste trading platform through interaction, which would result in their willingness to use the platform. Additionally, our findings suggested that both users may place a high value on the security of their personal information when utilizing the platform. In conclusion, platform developers and policymakers are responsible for fostering users’ behavioral intentions by leveraging their experience and learning capabilities. Ultimately, the adoption of waste trading platforms is expected to enhance users’ intentions to engage with such platforms.