Blockchain investment and dual-channel structure selection in fresh agricultural supply chains

Abstract

Fresh agricultural products are highly perishable and require strict preservation conditions. Information opacity amplifies quality uncertainty, which further weakens consumer trust and increases product loss. This not only raises operating costs but also leads to resource waste and environmental burdens. The introduction of blockchain technology can enhance transparency and credibility through immutable traceability information, but investment costs also affect channel profit distribution and channel structure selection. Therefore, this study incorporates blockchain investment level and investment costs into the model, compares pricing decisions and profit allocation under different channel modes with and without blockchain investment, and identifies the boundaries of channel structure selection. The results show that: (1) The supplier's channel structure selection is mainly affected by the direct-selling cost, and the introduction of blockchain changes the boundary of mode selection. When the direct-selling cost is low, the direct-selling structure has a greater advantage; as the direct-selling cost increases, the advantage of the e-commerce structure becomes stronger. After blockchain investment, the applicable region of the direct-selling structure expands within a certain range. (2) Whether the retailer invests in blockchain depends on the matching relationship among trust gain, the market size expansion effect, and investment costs. When demand improvement is significant and investment costs are relatively low, the retailer is more likely to invest in blockchain. Compared with the direct-selling structure, the feasible region for blockchain investment is larger under the e-commerce structure. (3) Blockchain investment affects supply chain decisions and profits through demand improvement and cost constraints. Within the effective investment interval, an increase in the blockchain investment level can increase offline demand and channel prices and improve member profits. However, excessively high investment costs or investment levels may weaken the advantage of the blockchain investment mode. Based on these findings, this study provides operational threshold criteria and managerial implications for channel structure optimization and rational blockchain investment in fresh agricultural dual-channel supply chains.

1 Introduction

Against the backdrop of global climate governance and the continued advancement of sustainable development goals, the green transformation of food systems has gradually become an important pathway for achieving coordinated economic growth and ecological protection. Relevant policies encourage enterprises to reduce resource consumption and environmental burdens in production and distribution through green technological innovation and improved environmental governance mechanisms. However, fresh agricultural products are highly perishable, heavily dependent on cold chain logistics, and subject to high return, exchange, and loss risks. As a result, their supply chains face more complex sustainability constraints than those of durable goods. The core issue lies in the transmission of quality information. If key information regarding origin, inspection, storage, and transportation cannot be delivered to consumers in a timely and credible manner, the resulting trust deficit will directly suppress purchase intentions, trigger demand contraction and transaction frictions, and further exacerbate product loss and resource waste. In this sense, removing barriers to information transmission is crucial for the green transformation of fresh product supply chains.

This issue becomes particularly prominent in the context of the rapid expansion of China's fresh product market. In 2024, the total retail scale of fresh products approached RMB 10 trillion, of which fresh e-commerce transactions reached RMB 736.79 billion, indicating significant growth in online channels (China Merchants Industry Research Institute, 2025; E-commerce Research Center NetEconomy, 2025). The virtual nature of online transactions prevents consumers from directly assessing product quality. Consequently, their reliance on quality information is much higher than in offline settings. Meanwhile, upstream suppliers, while retaining traditional offline channels, may cooperate with third-party fresh e-commerce platforms to expand online sales, or establish their own online direct-selling channels to reach consumers directly. This leads to two typical dual-channel structures. Under different channel structures, the intensity of competition, allocation of pricing power, and risk-bearing arrangements vary significantly. These differences directly affect the efficiency and credibility of information transmission, thereby influencing suppliers' long-term channel strategies and overall supply chain performance. Although existing studies have extensively examined pricing and competition in dual-channel supply chains, most focus on durable or standardized products and often assume a fixed channel structure. Systematic comparative research that accounts for the specific constraints of fresh products—such as high loss rates, strong time sensitivity, and frequent returns and exchanges—remains limited.

In this context, blockchain technology offers a new possibility for addressing information transmission challenges in fresh supply chains. Fresh product consumption is highly sensitive to food safety and traceability information. Information asymmetry and trust deficits significantly increase transaction costs and weaken consumers' willingness to engage in sustainable consumption. Blockchain technology, with its features of traceability, immutability, and verifiability, can generate credible records of key data, including origin, inspection, logistics, and transactions. This reduces information distortion and moral hazard, enhances consumers' trust in quality information, and improves purchasing decisions. This perspective aligns with the consensus in sustainable supply chain research: improving information transparency and reducing information asymmetry help strengthen consumer confidence and support more responsible consumption behavior. However, blockchain investment requires careful cost consideration. Implementation typically involves a fixed upfront investment and operating costs that increase with the level of investment. These costs reshape channel profit distribution through price and demand transmission mechanisms, thereby influencing the strategic choices of supply chain members.

Against this background, this study incorporates blockchain investment and application into the analytical framework of dual-channel structure selection for fresh agricultural products. It seeks to address the following questions: How should the supplier choose between the e-commerce structure and the direct-selling structure? Should the traditional retailer invest in blockchain technology, and what factors affect its investment decision? How does blockchain investment affect pricing decisions, demand changes, and profit allocation under different modes?

The main contributions of this study are as follows: (1) Within the framework of a fresh agricultural product dual-channel supply chain, this study characterizes the supplier's pricing decisions and profit allocation mechanisms under the e-commerce structure and the direct-selling structure, systematically compares different modes, and identifies the boundary conditions for the supplier's mode selection. (2) From the perspective of the traditional retailer, this study compares profit differences between blockchain investment and non-investment scenarios, characterizes the decision conditions for the retailer's blockchain investment, and identifies the feasible region for blockchain investment. (3) This study incorporates blockchain investment level, trust gain, the market size expansion effect, and investment costs into a unified analytical framework, revealing how blockchain investment affects pricing behavior, demand changes, and profit allocation outcomes under different dual-channel modes through demand improvement and cost constraints.

The remainder of this paper is organized as follows: Section 2 reviews the relevant literature and identifies research gaps; Section 3 presents the model description and basic assumptions; Sections 4 and 5 construct, solve, and compare the e-commerce and direct-selling dual-channel models without and with blockchain investment, respectively; Section 6 compares blockchain investment decisions under different channel structures; Section 7 provides numerical analysis; Section 8 concludes with key findings, managerial implications, and directions for future research.

2 Literature review

2.1 Research on general dual-channel supply chains and channel structure selection

Research on dual-channel supply chains originates from the practical context in which manufacturers (or upstream firms) introduce a direct-selling channel in addition to the traditional retail channel. Early studies mainly examined the motivation for adding the direct channel, pricing decisions, and member payoffs. From a strategic perspective, Chiang et al. (2003) argue that introducing a direct-selling channel does not necessarily harm retailers; under certain conditions, it can generate a win–win outcome. Dumrongsiri et al. (2008) develop a dual-channel model with coexisting direct selling and retailing, showing that channel expansion can increase system profit under centralized decision-making. Yan and Pei (2009) incorporate retail service into dual-channel competition and suggest that online direct selling may pressure retailers to improve service levels. Subsequent research extends the dual-channel setting by considering consumer preferences, contracts, and service competition. Hsiao and Chen (2014) emphasize heterogeneous motivations for introducing an Internet channel and the mechanism of channel preference. Zhao (2015) examines hybrid-channel pricing decisions when manufacturers open an online direct-selling channel. Fan and Liu (2016) analyze pricing and service competition strategies after the introduction of an online channel. From the perspective of customized products, Batarfi et al. (2016) show that a dual-channel structure can improve the performance of a centralized supply chain. Building on this stream, scholars further focus on power structures and channel structure selection. Sun et al. (2019) examine manufacturers' choices of online selling formats from the perspective of offline power structure. Chen et al. (2016), within an O2O hybrid-channel framework, analyze how power structure affects retailers' service decisions. (Joshi et al. 2023) introduce information-disclosure technology investment into multi-channel competition and reveal how such investment shapes channel choice and pricing. (Wang et al. (2023) further consider social media advertising and its impact on omni-channel competitive behavior. Lu et al. (2024) investigate blockchain adoption strategies in a dual-channel supply chain that includes both online and offline retailers, and compare optimal decisions across adoption modes and product types. Liang and Xiao (2023) additionally incorporate consumers' sensitivity to inspection time and blockchain verification costs, and characterize how blockchain adoption alters dual-channel pricing and channel selection outcomes. Gan and Huang (2022) discuss the governance value of blockchain for data integrity in dual-channel supply chains from the perspectives of information asymmetry and data credibility.

2.2 Research on dual-channel supply chains and channel structure selection for fresh agricultural products

Fresh agricultural products are characterized by short life cycles, high loss rates, and strong sensitivity to quality. These features make channel competition and coordination mechanisms significantly different from those of general products. Within a dual-channel framework, research has increasingly focused on preservation investment, quality competition, and logistics service coordination. Yang and Tang (2019) compare the applicability of retail, dual-channel, and O2O structures under preservation effort considerations. They show that the level of freshness investment significantly affects channel structure selection. Hu et al. (2022) examine sales mode selection and pricing strategies in a pre-sale context. Tian et al. (2022), based on a community fresh O2O setting, analyze how preservation investment influences demand, channel structure selection, and coordination mechanisms. Lin et al. (2023), within a “quality-price” competition framework, compare alternative selling modes in fresh e-commerce and evaluate their relative advantages. On the demand side, Xu et al. (2023) incorporate both consumers' low-carbon preference and freshness preference into the demand function of a dual-channel agri-food supply chain. They compare optimal pricing and effort decisions under different channel leadership structures, providing a useful reference for modeling the “preference-demand-decision” linkage in fresh product contexts. Ye et al. (2023) introduce cold-chain logistics service into the analysis and examine service and pricing decisions under different trade modes in fresh agricultural supply chains.

2.3 Research on blockchain applications in fresh agricultural supply chains

Blockchain technology, characterized by traceability, immutability, and smart contracts, is widely regarded as an effective tool for mitigating quality information asymmetry and trust deficits in supply chains Wang et al., (2025). Its role in enhancing transparency and credibility in fresh agricultural supply chains has been extensively examined. Liu and Li (2022) study blockchain investment decisions and coordination mechanisms under retailer risk aversion. Zheng et al. (2023), in the context of front-end warehouse operations for fresh products, compare different investors in traceability technology and evaluate their profit improvements. Sun et al. (2023) incorporate consumers' traceability preferences and analyze how blockchain adoption influences optimal pricing and competitive strategies.

From a broader perspective, Stranieri et al. (2021) discuss the role of blockchain in improving performance and information management in agri-food supply chains through case studies. Compagnucci et al. (2022) further summarize the application potential and key challenges of blockchain in agri-food supply chains from an interdisciplinary case-based perspective. de Carvalho et al. (2022), using highly perishable products such as fresh-cut flowers as examples, jointly optimize blockchain adoption and supply chain network design, revealing how blockchain affects the design and operation of perishable product supply chains. Zhang et al. (2025), in a fresh supply chain that includes a third-party cold-chain logistics provider, introduce different freight cost-sharing scenarios and compare blockchain adoption strategies and their effects on profit allocation under different cost-bearing structures. Li et al. (2023), from a dynamic optimization perspective, examine how the level of blockchain adoption affects preservation effort, advertising investment, and other operational decisions in fresh agricultural supply chains. Their study provides further evidence on the linkage among blockchain investment level, operational effort decisions, and supply chain performance.

2.4 Research on the impact of blockchain on consumer purchasing behavior

Information asymmetry has long been regarded as a key factor constraining efficiency in food markets. According to the “lemons market” theory, when product quality information cannot be effectively identified, consumers reduce their willingness to pay or even withdraw from the market, leading to market contraction and declining transaction efficiency. In the context of fresh agricultural products, this problem is more pronounced due to strong perishability, quality variability, and heavy reliance on cold-chain logistics. Consumers are highly sensitive to safety risks, which further intensifies the impact of information asymmetry. In recent years, a growing body of empirical research has shown that blockchain-based traceability can significantly enhance consumer trust and thereby influence purchasing decisions. Acciarini et al. (2023), through scenario-based experiments, find that consumers exhibit significantly higher purchase intention when informed that firms use blockchain technology to secure food labeling information. This suggests that blockchain-enabled information disclosure has a clear demand-enhancing effect. Doanh et al. (2026), in the context of agricultural e-commerce, develop a model linking “blockchain traceability–product trust–purchase intention.” Their results indicate that blockchain-based food traceability systems significantly increase online purchase intention by strengthening product trust. Based on survey data from pilot cities for food traceability in China, Zhai et al. (2022) found that the stronger consumers' perceived health risk, the higher their willingness to purchase blockchain-traceable fresh fruits, further indicating that blockchain traceability helps improve the market acceptance of fresh products.

Beyond improvements in purchase intention, consumers' valuation of traceability attributes further supports the potential demand-expansion effect of blockchain. Tran et al. (2024), based on a meta-analysis of multiple studies on food traceability, find that consumers are generally willing to pay a significant premium for traceable information. This implies that traceability enhances overall product value perception and shifts the demand curve outward. Jin et al. (2023) also report that consumers hold positive attitudes toward food traceability, and traceability labels significantly affect purchase choice. This effect is particularly strong in high-risk fresh categories such as aquatic products. He et al. (2026), through a discrete choice experiment on blockchain-traceable shrimp products, demonstrate that consumers show significant preference for blockchain traceability attributes and exhibit a substantial willingness to pay a premium.

Taken together, these studies suggest that blockchain technology not only improves transaction efficiency by restructuring information transparency, but also expands effective market demand by enhancing trust and willingness to pay. By reducing perceived risk, blockchain may attract potential consumers who would otherwise exit the market due to safety concerns, thereby enlarging the effective market size.

2.5 Research on blockchain applications in dual-channel fresh agricultural supply chains

Research that integrates blockchain with dual-channel fresh supply chains mainly focuses on pricing decisions, service investment, and channel structure selection. In the context of agricultural dual-channel traceability, Li and Zhao (2019) incorporate a traceability system into dual-channel decision analysis. With specific attention to blockchain, Jiang et al. (2023) systematically compare equilibrium outcomes, investment thresholds, and profit changes in fresh dual-channel supply chains before and after blockchain adoption. Tan and Zeng (2023) examine how blockchain information disclosure and unit usage costs influence two-stage pricing strategies in an O2O fresh product setting. From a sustainability perspective, Modak et al. (2024) propose management and coordination strategies for fresh dual-channel systems under blockchain implementation. Shao et al. (2025), within a low-carbon supply chain framework, construct a manufacturer-led dual-channel game model that jointly analyzes green investment and the e-commerce retailer's blockchain adoption decision. They identify boundary conditions for blockchain adoption under different market shares and trust preference scenarios. Beyond these studies, Liu et al. (2021) analyze sales mode selection in fresh agricultural product supply chains under blockchain by considering varying levels of channel competition intensity. Xu et al. (2024) further examine the interaction between platform-based blockchain participation and channel encroachment, revealing how blockchain may reshape channel strategies by altering information structures and competitive dynamics. In addition, Ma et al. (2025) investigate blockchain adoption strategies in dual-channel agricultural supply chains by considering differences in counterfeiting behavior across online and offline channels. They identify conditions under which blockchain adoption can lead to Pareto improvements, thereby providing direct support for the integration of food safety governance and blockchain technology.

2.6 Research review and gap identification

In summary, research on general dual-channel supply chains and mode selection has become relatively mature, forming fairly clear analytical approaches in areas such as channel competition, coordination mechanisms, power structures, and mode selection. However, most of these studies are based on the context of general products and provide relatively limited characterization of the distinctive features of fresh agricultural products, such as high perishability, high loss rates, opaque quality information, and strong consumer sensitivity to trust. Although existing studies on fresh-product dual-channel supply chains have further considered factors such as freshness-keeping effort, presale strategies, cold-chain logistics services, quality and price competition, and consumer preferences, and have identified decision conditions for different sales modes, the mechanism through which digital technologies further influence mode selection via changes in demand response still requires deeper investigation.

In research on blockchain and fresh-product dual-channel supply chains, the existing literature generally holds that blockchain can alleviate information asymmetry and improve supply chain coordination efficiency by enhancing transparency, traceability, and consumer trust. Relevant studies have mainly focused on within-channel decisions under given channel structures, emphasizing the effects of blockchain investment on pricing, freshness preservation, service, advertising, cost sharing, and coordination mechanisms, while further discussing its influence on profit distribution and Pareto improvement. Overall, existing research has provided a relatively sufficient explanation of how blockchain affects operational performance under given structures, but it has paid less attention to systematically comparing different dual-channel structures within a unified framework and further examining how blockchain investment changes the relative advantages and disadvantages among modes as well as their selection boundaries. Compared with the existing literature, the focus of this paper is not merely to discuss whether blockchain improves operational performance under a given structure, but rather to further examine how blockchain investment affects dual-channel mode selection. To this end, within the framework of a fresh-product dual-channel supply chain, this paper distinguishes between two channel structures, namely the e-commerce structure and the direct-selling structure, and further introduces two strategies for the traditional retailer, namely non-investment in blockchain and investment in blockchain, thereby forming four modes. Within a unified model, the paper compares the differences in pricing, demand, and profits across these four modes, and derives the threshold conditions for mode selection.

As can be seen from Table 1 and the corresponding literature, existing studies do not adopt a unified setting regarding the party that bears blockchain costs, and the research focus and conclusions also differ across different cost-bearing arrangements. When the supplier bears the cost, relevant studies pay more attention to whether the upstream party can recover the technology cost through wholesale price adjustment, sales expansion, or coordination mechanisms. When the retailer bears the cost, the research focus is placed more on blockchain information disclosure, the enhancement of consumer trust, and whether the investment returns can cover constraints such as fixed costs and variable costs. When the platform participates in or bears the relevant costs, blockchain adoption is more closely associated with the platform's mode of participation, platform sales mode selection, and channel entry decisions. This indicates that different cost-bearing arrangements not only change the allocation of technology investment, but also further affect the constraints on blockchain adoption, the path through which returns are realized, and the profit distribution among supply chain members.

Compared with the existing literature, which mostly discusses the adoption conditions of different cost-bearing parties under given channel structures, this paper sets the traditional retailer as the blockchain investor. On the one hand, the offline retailer directly faces end consumers and therefore bears more direct responsibility for food safety, quality disclosure, and trust maintenance. On the other hand, the focus of this paper is not to compare which cost-bearing party is superior, but to examine how, when the retail side bears blockchain costs, the trust gains and demand improvements brought about by blockchain investment further alter the relative returns of the e-commerce structure and the direct-selling structure, as well as the boundaries of mode selection, through the transmission mechanisms of prices and profits.

In addition, existing studies also differ in their specification of demand functions. As can be seen from Table 1, the existing literature mainly adopts two forms: linear demand functions and utility-based functions, while linear demand functions can be further divided into additive and multiplicative forms. Among them, the additive form directly incorporates factors such as blockchain-induced trust and quality inspection time into the demand function as additional terms, so as to characterize their marginal improvement in demand level. The multiplicative linear form, by contrast, treats factors such as freshness and traceability level as multiplicative factors acting on the linear core demand, thereby reflecting their overall amplifying or reducing effect on basic market demand. Utility-based functions, on the other hand, analyze the effects of consumer strategic behavior, quality disclosure decisions, and blockchain information transparency on purchase choices. This paper is more concerned with the issue of mode selection under blockchain investment. If the demand increase induced by blockchain were directly modeled in a multiplicative form acting on the entire demand function, both the price term and the cross-price term would be magnified simultaneously, which would be unfavorable for maintaining the relative stability of the dual-channel competitive structure in comparative analysis. Therefore, the following analysis adopts a linear framework and summarizes the blockchain effect as a trust gain effect and a market size expansion effect, so as to more clearly identify how these factors influence the comparative results across the four modes.

3 Problem description and basic assumptions

3.1 Model framework and notation

This study considers two common supplier-led dual-channel supply chain structures for fresh agricultural products. In the first structure, the supplier sells products through a traditional offline retailer and a third-party e-commerce platform (hereafter referred to as the e-commerce structure). In the second structure, the supplier sells products through a traditional offline retailer and a self-operated direct-selling channel (hereafter referred to as the direct-selling structure). Due to the multiple intermediaries involved in traditional retail channels, problems such as information opacity and traceability difficulties often arise. Therefore, this study considers the possibility that the traditional retailer adopts blockchain technology to improve transparency and operational efficiency in the traditional channel, thereby enhancing consumer trust. Based on whether the traditional retailer invests in blockchain technology, the two supply chain structures are further divided into four modes: the e-commerce structure without blockchain investment (NW), the direct-selling structure without blockchain investment (ND), the e-commerce structure with blockchain investment (YW), and the direct-selling structure with blockchain investment (YD), as shown in Figure 1. The definitions of the four modes are provided below, and the corresponding decision process is illustrated in Figure 2.

Figure 1

Figure 2

(1) NW Mode (No Investment—E-commerce Structure): The supplier cooperates with both the traditional retailer and a fresh e-commerce platform. The supplier wholesales agricultural products to the traditional retailer and the fresh e-commerce platform at given wholesale prices. The traditional retailer and the fresh e-commerce platform independently determine their retail prices and conduct sales.

(2) ND Mode (No Investment—Direct-Selling Structure): The supplier cooperates with the traditional retailer while simultaneously operating a self-owned direct-selling channel. The supplier wholesales agricultural products to the traditional retailer at a given wholesale price and directly sets the retail price in the direct-selling channel. The traditional retailer independently determines its retail price and conducts sales.

(3) YW Mode (Investment—E-commerce Structure): Based on the NW mode, the traditional retailer invests in blockchain technology to enhance information traceability and ensure food safety.

(4) YD Mode (Investment—Direct-Selling Structure): Based on the ND mode, the traditional retailer invests in blockchain technology to enhance information traceability and ensure food safety.

This study aims to analyze channel structure selection in a fresh agricultural dual-channel supply chain. Specifically, the supplier first decides the channel structure (W or D). Based on this structure, the traditional retailer determines whether to invest in blockchain technology (N or Y). To model these decisions, different short-term pricing games are considered. Under the NW and YW modes, the decision process follows a three-stage Stackelberg game. Under the ND and YD modes, it follows a two-stage Stackelberg game. All models are solved using backward induction. The main symbols and parameter definitions are presented in Table 2.

3.2 Basic assumptions and demand functions

Assumption 1: Considering the market share of the dual channels and the channel substitution intensity, the demand functions of the offline and online channels are defined as follows (Chen et al., 2016):

Where: a denotes the market size; β denotes the market share of the offline channel, reflecting consumers' preference for the offline channel; and b denotes the channel substitution intensity, characterizing the extent to which price changes in the online and offline channels affect each other's demand. To ensure that the equilibrium solution has a clear economic meaning, the substitution intensity between channels must be weaker than the effect of a channel's own price, and therefore 0 < b < 1 is assumed. The closer b is to 1, the more intense the competition between the two channels; as b approaches 0, the two channels become relatively independent.

Assumption 2: In a dual-channel fresh agricultural product supply chain, the traditional retailer directly faces end consumers and is responsible for quality management and regulatory compliance, and therefore has greater advantages in information collection and quality control (Jiang et al., 2023). On this basis, this paper takes the traditional retailer's undertaking of blockchain investment as the benchmark scenario in order to improve information transparency and consumer trust.

Assumption 3: After the traditional retailer invests in blockchain, blockchain affects demand through two aspects: offline trust gain and market size expansion. On the one hand, traceability information improves the transparency and reliability of products in the offline channel, thereby enhancing existing consumers' trust in the offline channel and converting it into an increase in offline demand. Let the blockchain investment level be τ and the trust gain coefficient be δ. Then, the increase in offline demand caused by blockchain investment is δτ (Liang and Zhang, 2020; He et al., 2022). Here, δ represents the intensity of the trust gain brought about by blockchain investment. A larger δ indicates a stronger demand-enhancing effect in the offline channel. On the other hand, by improving information transparency and credibility, blockchain may attract potential consumers into the market, thereby expanding the effective market size (Acciarini et al., 2023). Therefore, this study specifies the potential market size as a(τ) = a(1+μτ), where μ≥0 is the market size expansion coefficient. After simultaneously considering the offline trust gain and the market size expansion effect, the demand functions are as follows:

Assumption 4: The blockchain investment cost includes a one-time fixed investment cost c_f, and a unit variable cost c_v(τ) that increases convexly with the investment level. To reflect the characteristic of increasing marginal cost, the cost function is assumed to be (Liang and Zhang, 2020; He et al., 2022), where k represents the sensitivity of the variable cost to the blockchain investment level.

Assumption 5: To highlight the main focus of the analysis, and following the relevant literature (Liang and Xiao, 2023), this paper assumes that production cost is zero. At the same time, related distribution costs are not modeled separately but are incorporated uniformly into the direct-selling cost, which mainly includes expenses arising in the direct-selling channel such as cold-chain fulfillment, logistics distribution, packaging, and sorting.

4 Model construction and solution under dual-channel modes without blockchain investment

4.1 E-commerce structure without blockchain investment (NW mode)

This section examines the e-commerce structure (NW), in which the traditional retailer does not invest in blockchain and the supplier cooperates with a third-party fresh e-commerce platform. Under this mode, the supplier supplies products to both the traditional retail channel and the fresh e-commerce channel through offline and online wholesale arrangements. The traditional retailer and the fresh e-commerce platform then sell the agricultural products to end consumers through their respective channels. In this setting, the demand functions of the offline and online channels are given by Equation 1.

Based on this demand setting, the supplier's revenue is derived from wholesale income from both channels, while the traditional retailer and the fresh e-commerce platform earn profits through sales in their respective channels. Therefore, the profit functions of the three parties are given by Equations 2–4, respectively.

According to the game setting of the NW mode, the supplier first determines the wholesale prices for the two channels, then the traditional retailer determines the offline retail price, and finally the fresh e-commerce platform determines the online selling price. Therefore, this mode forms a three-stage Stackelberg game.

Using backward induction, in Stage 3, the fresh e-commerce platform chooses to maximize its profit and obtains the optimal online price. Next, substituting this optimal pricing decision into Stage 2, the traditional retailer chooses to maximize its profit and obtains the optimal offline price. Finally, substituting the downstream optimal pricing decisions into Stage 1, the supplier solves the joint first-order conditions with respect to and to obtain the optimal wholesale prices. By substituting the optimal decisions of each stage backward, the equilibrium wholesale prices, equilibrium retail prices, equilibrium demands, and equilibrium profits of all members under the NW mode can be obtained.

We further examine the second-order conditions of the profit functions at each stage: ,. After substituting the downstream optimal pricing decisions, the Hessian matrix of the supplier's profit function with respect to is:

Under the condition 0 < b < 1, we have and is negative definite. Therefore, Lemma 1 can be obtained. Under the NW mode, the three-stage Stackelberg game has a unique equilibrium. The equilibrium wholesale prices, equilibrium retail prices, and the corresponding equilibrium demands and profits are shown in Equations 5–13:

4.2 Direct-selling structure without blockchain investment (ND mode)

This section examines the direct-selling structure (ND), in which the traditional retailer does not invest in blockchain and the supplier establishes its own online channel. Unlike the e-commerce structure, under this mode the supplier wholesales agricultural products to the traditional retailer through the offline channel while also selling directly to end consumers through a self-operated online channel. In this setting, the demand functions of the offline and online channels are given by Equation 14.

Based on this demand setting, the supplier's revenue comes partly from wholesale income obtained from selling to the traditional retailer and partly from sales revenue generated through the online direct-selling channel, while also bearing the direct-selling cost associated with that channel. The traditional retailer, by contrast, earns profit through sales in the offline channel. Accordingly, the profit functions of the two parties are given by Equations 15, 16, respectively.

According to the game setting of the ND mode, the supplier first determines the wholesale price and the online direct-selling price, and then the traditional retailer determines the offline retail price. Therefore, this mode forms a two-stage Stackelberg game.

Using backward induction, in Stage 2, the traditional retailer chooses to maximize its profit and obtains the optimal offline price. Next, substituting this optimal pricing decision into Stage 1, the supplier solves the joint first-order conditions with respect to ω^ND and to obtain the optimal wholesale price and the optimal online direct-selling price. By substituting the optimal decisions of each stage backward, the equilibrium wholesale price, equilibrium retail price, equilibrium direct-selling price, and the corresponding equilibrium demands and profits under the ND mode can be obtained.

We further examine the second-order condition of the profit function at each stage: . After substituting the traditional retailer's optimal pricing decision into the supplier's profit function, the Hessian matrix of the supplier's profit function with respect to is: . Under the condition 0 < b < 1, we have is negative definite. Therefore, Lemma 2 can be obtained. Under the ND mode, the two-stage Stackelberg game has a unique equilibrium. The equilibrium wholesale price, equilibrium retail price, direct-selling price, and the corresponding equilibrium demands and profits are shown in Equations 17–23:

4.3 Comparative analysis between ND and NW modes

In the absence of blockchain investment, the supplier faces two typical dual-channel structures. One is the e-commerce structure (NW), in which the supplier cooperates with a third-party fresh e-commerce platform. The other is the direct-selling structure (ND), in which the supplier establishes its own online channel. By comparing pricing decisions and profit outcomes under these two structures, the following conclusions are obtained.

Proposition 1: In the e-commerce structure without blockchain investment, the wholesale price of each channel is positively correlated with its corresponding channel market share.

Proof:

When , ; When , ; When , .

Proposition 1 indicates that when the market share of a given channel increases, its contribution to the supplier's profit rises, and the supplier tends to capture more upstream profit by raising the wholesale price for that channel. Conversely, a decrease in market share weakens the channel's bargaining power, and the corresponding wholesale price is adjusted downward.

Proposition 2: Under the direct-selling structure without blockchain investment, the wholesale price is independent of the direct-selling cost. The retail price and the direct-selling price are positively correlated with the direct-selling cost c. The retailer's profit increases with the direct-selling cost, whereas the supplier's profit decreases as the direct-selling cost increases.

Proof:

When , ,

Proposition 2 indicates that, under the direct-selling structure, an increase in the direct-selling cost raises both the retail price and the direct-selling price. Meanwhile, the direct-selling price is more sensitive to the direct-selling cost, and part of the demand in the online channel shifts to the offline channel, thereby increasing the traditional retailer's profit. As the supplier bears the direct-selling cost, it not only incurs higher cost but also faces channel conflict, resulting in a decrease in the supplier's profit as the direct-selling cost increases.

Proposition 3: In the absence of blockchain investment, when 0 < c<c₁, ; , , where: .

Proposition 3 indicates that the direct-selling cost is a key factor affecting the supplier's channel structure selection. The threshold represents the upper cost limit that the direct-selling structure can bear, and it is affected by market size, channel demand allocation, and channel substitution intensity. Specifically, a larger market size increases the terminal returns that can be obtained under the direct-selling structure. A higher potential demand share of the online channel makes the profit-internalization advantage of the direct-selling structure more pronounced. A higher channel substitution intensity strengthens the price linkage between the online and offline channels, thereby affecting the relative returns of the two structures. When the direct-selling cost is low, the disintermediation benefit can cover the additional cost, so the ND mode is more advantageous. As the direct-selling cost increases, the cost pressure on the direct-selling side gradually weakens its profit advantage, and the supplier is more likely to shift to the NW mode.

5 Model construction and solution under dual-channel modes with blockchain investment

5.1 E-commerce structure with blockchain investment (YW mode)

This section examines the e-commerce structure (YW), in which the traditional retailer invests in blockchain and the supplier cooperates with a third-party fresh e-commerce platform. Compared with the NW mode, the channel structure and decision sequence remain unchanged in the YW mode. The only difference is that the traditional retailer bears the blockchain investment cost and enhances offline demand and consumer trust through credible information disclosure. Let the blockchain investment level be τ, and the trust gain coefficient be δ. Then, compared with the case without investment, the offline demand increases by δτ. At the same time, blockchain investment enhances supply chain transparency and credibility, attracting consumers who would otherwise withdraw from the market due to information asymmetry to re-enter the market, thereby expanding the potential market size. The demand functions and profit functions under the YW mode are given in Equations 24 and 25, respectively, and the equilibrium solutions are shown in Equations 26–34:

Under the YW mode, in addition to earning channel profit, the traditional retailer must bear the blockchain investment costs, including the fixed cost c_f and the variable cost c_v. In this case, the profit functions of the supplier, the traditional retailer, and the fresh e-commerce platform become:

Following the NW mode and using backward induction, the equilibrium solutions are obtained as follows:

where:

5.2 Direct-selling structure with blockchain investment (YD mode)

This section examines the direct-selling structure (YD), in which the traditional retailer invests in blockchain and the supplier establishes its own online channel. Compared with the ND mode, the channel structure and decision sequence remain unchanged in the YD mode. The difference is that the traditional retailer bears the blockchain investment cost, and blockchain-enabled credible disclosure enhances the traceability and reliability of the offline channel. This generates a trust-driven increase δτ in offline demand and expands the potential market size. The demand functions and profit functions under the YD mode are given in Equations 35 and 36, respectively, and the equilibrium solutions are shown in Equations 37–43:

Under the YD mode, the supplier's profit function, based on the ND mode, still includes the unit direct-selling cost c. The traditional retailer's profit function, based on the ND mode, deducts the blockchain fixed cost c_f and the variable cost c_v.

Following the ND mode and using backward induction, the equilibrium solutions are obtained as follows:

where:

5.3 Comparative analysis between YD and YW modes

Under the condition that the traditional retailer invests in blockchain, this section compares the equilibrium outcomes and profit differences between the e-commerce structure (YW) and the direct-selling structure (YD). The following conclusions are obtained.

Proposition 4: Under the e-commerce structure with blockchain investment, the wholesale price of each channel is positively correlated with its corresponding channel market share.

Proof:

When , When , When , ,

When , .

Proposition 5: Under the direct-selling structure with blockchain investment, the wholesale price is independent of the direct-selling cost. The retail price and the direct-selling price are positively correlated with the direct-selling cost. The retailer's profit increases with the direct-selling cost, whereas the supplier's profit decreases as the direct-selling cost increases.

Proof:

When , ,

Proposition 6: Under blockchain investment, when 0 < c<c₂, ; when , , where , and c₂>c₁.

Proposition 6 indicates that, under blockchain investment, the supplier's channel structure preference is still constrained by the direct-selling cost. Compared with the direct-selling cost threshold without blockchain investment, the threshold under blockchain investment is higher. This is because blockchain can improve demand conditions by enhancing transparency, strengthening consumer trust, and expanding the effective market size, thereby increasing the profit advantage of the direct-selling structure to some extent. However, these benefits do not automatically offset the operational pressure of the direct-selling channel. When the direct-selling cost is low, the supplier can still internalize terminal returns more fully through a self-operated online channel, and therefore tends to prefer the YD mode. When the direct-selling cost increases to a certain level, the additional cost of the direct-selling channel weakens this internalization advantage, and the supplier is more likely to shift to the YW mode.

Further analyze the impact of the blockchain investment level on dual-channel decision-making.

Proposition 7: Under the two blockchain investment modes (YW and YD), the offline wholesale price increases monotonically with τ within the effective investment interval . Meanwhile, the offline retail price and the online price also increase with τ.

Proof:

when , ,.

;

.

Proposition 7 indicates that, within the effective investment interval, a higher level of blockchain investment affects dual-channel pricing decisions through demand-side improvements. On the one hand, blockchain enhances the transparency and traceability of quality information, which helps strengthen consumer trust and expand effective market demand. As a result, supply chain members gain greater pricing flexibility, and both the offline retail price and the online price generally increase with the level of blockchain investment. On the other hand, although a higher level of blockchain investment also leads to an increase in variable costs, within the effective investment interval the effects of demand expansion and trust enhancement still outweigh the cost pressure. Therefore, the offline wholesale price also increases as the level of blockchain investment rises. This shows that the impact of blockchain investment on prices reflects the combined effects of demand improvement and cost constraints, but within the effective investment interval, the demand-improvement effect remains dominant.

6 Analysis of the impact of blockchain investment on dual-channel decisions

6.1 Comparative analysis between NW and YW modes

Proposition 8: Under the e-commerce structure, blockchain investment increases the profit of the fresh e-commerce platform. When the variable cost satisfies or and the fixed cost satisfies , the retailer can increase its profit through blockchain investment. When the variable cost satisfies or , the supplier's profit increases.

Proof:

For simplicity, let A = −2b+(−4+b(2+b))β, B = 2+(−2+b)β.

(1) Comparison of Fresh E-commerce Platform Profits

Therefore, under the model setting of this study, the profit of the fresh e-commerce platform increases under blockchain investment.

(2) Comparison of Retailer Profits

By taking the difference between the equilibrium profits, we obtain , where Δ_R(τ, c_v) is defined as the incremental operational benefit:

For the retailer to invest in blockchain, the operating condition must satisfy . This condition can be equivalently expressed as an interval constraint on c_v, with the two boundary values given by:

Thus, or holds.

Under the condition that the variable cost requirement is satisfied, the retailer's profit increases if and only if .

(3) Comparison of Supplier Profits

where

At this time, the condition for the supplier's profit to increase, ΔΠ_S≥0, is equivalent to M(c_v)≥0, and M(c_v)≥0 is a quadratic function that opens upward. Let

When N≥0, the two roots of the quadratic equation M(c_v) = 0 are:

Thus, or holds.

Proposition 8 indicates that, under the e-commerce structure, the introduction of blockchain leads to significant differences in benefit distribution among supply chain members.

(1) The fresh e-commerce platform can benefit. Although the blockchain investment is borne by the offline retailer, the resulting trust gain and market size expansion reshape the demand structure of the dual channels. Since the fresh e-commerce platform does not bear the blockchain fixed cost or unit variable cost, yet can share the gains from demand expansion and price increases, it can benefit under the YW mode compared with the NW mode.

(2) Whether the retailer benefits depends on the intensity of the variable cost. Although blockchain investment helps expand market demand and increases pricing potential, the retailer bears both the unit variable cost and the fixed cost. The retailer's profit increment can be decomposed as “incremental operational benefit minus fixed cost.” When the variable cost is excessively high, the cost pressure offsets the trust-driven gains through equilibrium pricing and demand transmission mechanisms. Only when the incremental operational benefit is sufficient to cover the fixed cost will the retailer achieve a net increase in profit.

(3) The supplier's profit also exhibits a cost-threshold characteristic. The trust gain induced by blockchain expands effective demand and increases pricing potential, which generates a positive effect for the supplier. However, the variable cost borne by the retailer is transmitted upstream through wholesale price and demand adjustment mechanisms, thereby compressing the supplier's profit margin. Therefore, whether the supplier's profit increases still depends on c_v.

6.2 Comparative analysis between ND and YD modes

Proposition 9: Under the direct-selling structure, when the variable cost satisfies and , the retailer can increase its profit through blockchain investment; when the variable cost satisfies , the supplier's profit increases.

By taking the difference between the equilibrium profits, we obtain , where

From , the threshold value of the variable cost can be derived as:

where E = (a(τ)²−a²)(2+(−1+b)β(4+(−3+b)β))+2(a(τ)−a)(b²−1)c(2+(−2+b)β)+2a(τ)(2b+(b²−1)β)δτ+2(b²−1)bcδτ+(1+b²)δ²τ².

At this time, the condition for the supplier's profit to increase, ΔΠ_S>0, is equivalent to

Solving yields: .

Proposition 9 indicates that, under the direct-selling structure, the introduction of blockchain also leads to differences in benefits across members.

(1) For the retailer, blockchain investment affects profit through two opposite channels. On the one hand, it increases product demand and sales revenue by enhancing consumer trust and expanding the potential market size. On the other hand, the retailer bears both the unit variable cost and the fixed cost. Therefore, whether the retailer's profit increases depends on the trade-off between the demand expansion effect and the investment cost. Blockchain investment is profitable only when the unit variable cost is below the critical level implied by the demand increment and the fixed cost does not exceed the incremental operational benefit.

(2) For the supplier, blockchain does not directly alter its cost structure. However, the unit variable cost borne by the retailer is indirectly transmitted through the equilibrium wholesale price and the demand structure, thereby affecting the supplier's profit. Therefore, an increase in the supplier's profit also requires that the feasible threshold condition for c_v be satisfied.

7 Numerical analysis

The numerical analysis parameters are set based on a real-world context. In terms of market parameters, this study takes the Zespri fruit supply chain in the Yangtze River Delta as an example. The brand enters the regional market through distribution nodes such as Shanghai Nangang and Jiaxing Haiguangxing Fruit Market, while also operating through offline retail terminals and online e-commerce platforms. Based on the regional circulation scale of premium fruits in the Yangtze River Delta and industry studies on the sales ranges of comparable leading fresh product brands, the average monthly sales volume is set as a = 120 tons. The offline sales share is set as β = 0.6. According to public disclosure by Zespri executives, about 40% of its products in the Chinese market are sold through e-commerce structures, indicating that offline channels still play a dominant role for fresh product brands (Foodaily, 2023). Referring to the cost structure of the fresh e-commerce industry and considering comprehensive costs in online channels, such as cold-chain fulfillment, logistics distribution, packaging, and sorting, the unit direct-selling cost is set as c = 15 (10,000 yuan/ton). This represents a medium-to-high cost scenario when the supplier undertakes direct-selling operations. Regarding channel substitution intensity, existing studies generally assume that this parameter is weaker than the own-price effect of the corresponding channel, so as to ensure that the dual-channel equilibrium solutions have clear economic meanings (Zhang et al., 2019; Yue et al., 2016). Therefore, this study sets b = 0.6, indicating a moderately strong substitution relationship between online and offline channels.

In terms of blockchain investment, the fixed cost is set as c_f = 90 (10,000 yuan), with reference to the budget level of government procurement projects for agricultural product blockchain traceability platforms (Fuping County Market Supervision Administration, 2025) and the investment range of firms of comparable size, namely 80–150 (10,000 yuan). The standardized blockchain investment level is set as τ = 0.65, representing a medium-to-high level of technology deployment by the retailer. The variable cost sensitivity is set as k = 26, which is used to capture the extent to which costs related to blockchain operation and maintenance, data verification, node coordination, and information uploading increase with the investment level. Under the benchmark investment level, the sum of fixed investment and related variable costs falls within the industry investment range, indicating that the blockchain cost parameters have a certain real-world basis.

In terms of consumer response behavior parameters, existing studies show that the quality of agricultural product traceability information, the credibility of traceability information sources, and restrictions on the use of traceability labels positively affect consumers' purchase intention, with perceived risk control playing a partial mediating role (Gao and Xiang, 2025). Meanwhile, information source trust further affects purchase intention by enhancing consumers' confidence in traceable food consumption (Meng et al., 2019). Based on these findings and the model assumptions, this study sets δ = 0.5 to characterize the extent to which the unit trust improvement brought about by blockchain investment affects market demand. Meanwhile, the market size expansion coefficient is set as μ = 0.55, reflecting the activation effect of trust reconstruction on the potential market. The robustness of these parameter settings is further verified through sensitivity analysis.

7.1 Impact analysis of dual-channel structure parameters

Figure 3 shows that the offline market share β and the channel substitution intensity b significantly influence the profit levels of supply chain members under different modes. As either β or b increases, the profit differences among the four modes gradually widen, which in turn affects the optimal structural choices of supply chain members.

Figure 3

For the supplier, when the offline market share becomes larger or the substitution intensity between channels increases, the trust enhancement and demand expansion effects generated by blockchain technology are more likely to be transformed into upstream profits. As a result, the profit improvement under the blockchain investment modes YD and YW becomes more pronounced. In a relatively wide parameter range, the direct-selling structure allows the supplier to internalize more downstream profits, which strengthens the relative advantage of the direct-selling channel.

For the retailer, as β orbincreases, the profit improvement under the blockchain investment modes—especially the YW mode—becomes more significant. This indicates that when the offline demand base is relatively strong or when the substitution relationship between online and offline channels is stronger, blockchain investment is more likely to generate net profit gains. Conversely, when β and b are relatively small, the demand expansion and trust premium brought by blockchain technology are limited, and the investment benefits may not fully offset the associated costs. Under such conditions, retailers are more likely to maintain non-blockchain modes.

Figure 4 shows that the optimal supply chain structure varies across different combinations of b and β. When blockchain is not adopted, the supplier prefers the direct-selling structure ND in most regions of the parameter space. Only in a limited region characterized by a relatively large offline market share and weak channel competition does the supplier prefer the e-commerce structure NW.

Figure 4

This result suggests that when the supplier can effectively internalize downstream profits through the direct-selling channel, the direct-selling structure is generally more attractive. However, when the offline market share is sufficiently large and channel competition is relatively weak, cooperation with e-commerce platforms may become a more profitable alternative.

After the introduction of blockchain technology, the supplier's optimal structure is mainly concentrated in the YD region, with YW appearing only in a relatively small part of the parameter space. This indicates that blockchain investment does not fundamentally change the supplier's preference for the direct-selling structure but further strengthens the advantage of the direct-selling mode through demand expansion effects.

From the retailer's perspective, without blockchain investment the optimal regions are mainly distributed between ND and NW. After blockchain investment, however, the optimal region shifts primarily to YW, with NW or YD appearing only in a limited area. This suggests that when the offline market base is relatively strong and the competition between channels is intense, retailers are more inclined to invest in blockchain under the e-commerce structure in order to benefit from trust enhancement and demand expansion.

From a practical perspective, the offline market share and the degree of channel substitution essentially reflect consumers' shopping habits and the competitive relationship between online and offline channels. When the offline channel has a stronger market base, the retailer is more likely to translate blockchain investment into a credibility advantage. When the substitution relationship between online and offline channels is stronger, the trust improvement brought about by blockchain is more likely to be reflected in profit differences under channel competition. Therefore, this set of results suggests that the effect of blockchain investment depends not only on the technology itself, but also closely on the channel competitive environment in which the firm operates.

7.2 Impact analysis of direct-selling cost

Figure 5 shows that the direct-selling cost c significantly influences both pricing decisions and profit distribution within the supply chain. In terms of pricing decisions, an increase in the direct-selling cost mainly affects the ND and YD modes, which involve direct-selling activities, while the pricing decisions in the e-commerce structures NW and YW remain relatively stable.

Figure 5

Under the ND and YD modes, both the direct-selling price and the retail price increase as the direct-selling cost rises. This indicates that the supplier tends to transfer part of the direct-selling cost to the end market through price adjustments.

In terms of profits, the supplier's profits under ND and YD decrease as the direct-selling cost increases, while the profits under NW and YW remain relatively stable. This suggests that the direct-selling cost is a key factor influencing the supplier's structural choice between the direct-selling channel and the e-commerce structure. When the direct-selling cost is relatively low, the supplier prefers the direct-selling structure in order to internalize more downstream profits. As the direct-selling cost increases, however, the supplier is more likely to switch to the e-commerce structure.

For the retailer, profits under ND and YD increase with the direct-selling cost. This is because higher direct-selling costs weaken the competitiveness of the supplier's direct channel, causing a portion of demand to shift toward the offline retail channel and thereby improving the retailer's sales volume and profitability.

This result indicates that whether the supplier is suitable for adopting the direct-selling structure depends largely on its fulfillment, delivery, and after-sales capabilities. If the supplier has relatively weak self-operated direct-selling capabilities, an increase in direct-selling costs will quickly weaken its profit advantage. In this case, cooperating with a third-party fresh e-commerce platform is usually a more prudent choice. Conversely, when the supplier can operate the online direct-selling channel at a relatively low cost, establishing a self-operated direct-selling channel is more conducive to capturing terminal returns.

7.3 Analysis of the effects of blockchain investment level and consumer response parameters

Figure 6 shows that the blockchain investment level τ has a significant impact on pricing decisions and profit levels under blockchain-enabled supply chain structures. As τ increases, the wholesale price, retail price, and direct-selling price under the blockchain modes YW and YD generally rise.

Figure 6

This result indicates that blockchain technology enhances consumer trust and expands the effective market demand, which provides supply chain members with greater pricing flexibility.

In terms of profits, the profits under the non-blockchain modes ND and NW remain almost unchanged as τ varies, since these structures do not involve blockchain investment. In contrast, under the blockchain investment modes YD and YW, profits gradually increase with higher blockchain investment level. The supplier benefits from demand expansion and can capture part of the increased value through price adjustments, while the retailer, although bearing the blockchain investment cost, may still experience profit improvements within certain investment ranges due to increased demand.

Overall, the effect of blockchain investment level on supply chain profits reflects the joint influence of the demand expansion effect and the cost pressure effect. When the blockchain investment level is relatively low, its benefits may be insufficient to cover the associated costs. However, once the investment level reaches a certain threshold, the profit advantage of the blockchain investment mode gradually becomes more evident.

In addition, to further examine the robustness of the conclusions under different consumer response scenarios, this study analyzes the effects of the trust gain coefficient and the market size expansion coefficient on profits under different modes. The results are shown in Figure 7. As shown in Figure 7, as the trust gain coefficient increases, the profits of the supplier and the retailer under the blockchain investment modes, namely YW and YD, generally show an upward trend, while the profits under the non-investment modes, namely NW and ND, remain basically stable. This indicates that the demand-enhancing effect brought about by blockchain becomes more evident as consumer trust increases. As the market size expansion coefficient increases, the profit improvement under the blockchain investment modes also becomes more pronounced. This suggests that blockchain can further improve the returns of supply chain members by attracting potential consumers into the market and expanding effective demand. When the trust gain coefficient or the market size expansion coefficient is at a relatively low level, the profit improvement under the blockchain investment modes is relatively limited, and the feasible region for investment may narrow accordingly. Overall, consumer response parameters mainly affect the magnitude of profit improvement and the boundary position under blockchain investment modes, but they do not change the direction of the main conclusions obtained above.

Figure 7

7.4 Joint analysis of blockchain investment and direct-selling cost

Figure 8 shows that the feasibility of blockchain investment depends jointly on the unit variable cost c_v and the fixed cost c_f. The left panel compares the YW and NW modes under the e-commerce structure, while the right panel compares the YD and ND modes under the direct-selling structure.

Figure 8

The results indicate that when the blockchain unit variable cost c_v is relatively low and the fixed cost c_f remains within a relatively low range, blockchain investment modes generate higher profits. However, as either type of cost increases, the non-blockchain modes gradually become the preferred choice.

Furthermore, the feasible investment region under the e-commerce structure is noticeably larger than that under the direct-selling structure. This suggests that blockchain investment is more likely to be profitable in the e-commerce structure, where the benefits from trust enhancement and demand expansion can more easily offset the associated costs. This further suggests that the effect of blockchain investment depends not only on whether the technology itself can enhance transparency and trust, but also on its ability to transmit benefits within a given channel structure. For fresh-product firms, whether to invest in blockchain should therefore be judged comprehensively in light of the channel structure and cost conditions they face.

To more directly examine the impact of blockchain cost sensitivity on the conclusions, this study further analyzes the changes in profits under different modes as parameter k varies, as shown in Figure 9. The results indicate that as k increases, the profits of both the supplier and the retailer under the blockchain investment modes, YW and YD, decline overall, while the profits under the non-blockchain modes, NW and ND, remain basically stable. This suggests that the faster the marginal cost associated with blockchain investment increases, the more easily the profit space of the blockchain investment modes is squeezed.

Figure 9

Further comparison shows that an increase in k weakens the profit advantage of the blockchain investment modes. From the supplier's perspective, the profit improvement of the YD mode relative to the non-blockchain modes is compressed. From the retailer's perspective, although the YW mode still maintains a relatively high profit level, its relative advantage diminishes as k increases, and the profit under the YD mode declines more rapidly. Overall, changes in k mainly affect the profit gaps among different modes and the feasible range of blockchain investment, but they do not alter the basic conclusion that the profitability of blockchain investment is constrained by cost conditions and that the e-commerce structure is more likely to satisfy the investment requirements.

Figure 10 shows that the direct-selling cost c and the blockchain investment level τ jointly influence the equilibrium supply chain structure. From the supplier's perspective, when the direct-selling cost is relatively low, the supplier prefers the direct-selling structure with blockchain (YD). As the direct-selling cost increases, the supplier gradually shifts toward the e-commerce structure with blockchain (YW), indicating that the direct-selling cost remains a key determinant of the supplier's structural decision even after blockchain investment.

Figure 10

From the retailer's perspective, when the blockchain investment level is relatively low, the non-blockchain modes ND or NW are more likely to be optimal because the demand expansion and trust effects generated by blockchain technology are insufficient to compensate for the associated investment costs. As τ increases beyond a certain threshold, the blockchain investment modes gradually become more advantageous.

This result suggests that blockchain investment exhibits an effective investment range: only when the benefits generated by blockchain investment are sufficient to cover both the fixed and variable costs will retailers choose to adopt blockchain technology.

The numerical simulation results for the key parameters provide further support for the propositions derived above. Changes in the direct-selling cost verify the conclusions of Propositions 3 and 6, namely that the supplier's choice between the e-commerce structure and the direct-selling structure is consistently constrained by the direct-selling cost. When the direct-selling cost is low, the direct-selling structure has a greater advantage; as the direct-selling cost increases, the advantage of the e-commerce structure becomes stronger. Changes in the blockchain investment level, trust gain, and market size expansion effect support the conclusion of Proposition 7, indicating that blockchain can affect pricing decisions and profit allocation by improving demand conditions, and can enhance the profit performance of blockchain-based modes within the effective investment interval. Changes in the fixed cost, unit variable cost, and cost sensitivity coefficient are consistent with Propositions 8 and 9, showing that the profitability of retailer blockchain investment depends on whether the benefits from demand improvement can cover the investment costs, and that the investment conditions are more easily satisfied under the e-commerce structure. Overall, changes in key parameters affect profit gaps, critical boundaries, and feasible investment regions, but they do not alter the basic directions revealed by the propositions, suggesting that the proposition analysis in this study is relatively robust.

8 Conclusion

Based on two common dual-channel structures for fresh agricultural products, this study develops a theoretical model to analyze the supplier's channel structure selection problem. It further considers the retailer's blockchain investment decision and reveals how blockchain-induced trust enhancement and cost constraints affect pricing strategies, profit distribution, and mode selection. The main conclusions are as follows:

First, the direct-selling cost is a key factor affecting the supplier's channel structure selection. When blockchain is not adopted, a lower direct-selling cost enables the supplier to capture terminal profits through the direct-selling structure. As the direct-selling cost increases, the cost pressure of the direct-selling structure becomes stronger, and the relative advantage of the e-commerce structure increases. After blockchain investment, trust gain and the market size expansion effect can improve demand conditions and expand the applicable region of the direct-selling structure within a certain range. However, the direct-selling cost remains an important boundary for the supplier's mode selection.

Second, whether the retailer invests in blockchain depends on the comparison between incremental benefits and investment costs. When the demand increase and price improvement brought by blockchain can cover the fixed cost and unit variable cost, blockchain investment is more beneficial for the retailer. When the trust gain is insufficient, the market size expansion effect is limited, or the investment cost is too high, the investment return will be weakened. Compared with the direct-selling structure, the retailer can more easily satisfy the profit improvement condition for blockchain investment under the e-commerce structure.

Third, blockchain investment jointly affects supply chain profits through demand improvement and cost constraints. Within the effective investment interval, a higher blockchain investment level can enhance consumer trust, increase offline demand, and drive up the wholesale price, offline retail price, and online price. The stronger the trust gain, the more sufficient the profit improvement; the larger the variable cost sensitivity coefficient, the greater the investment cost pressure. Blockchain investment helps the supplier expand wholesale revenue and can also improve the retailer's operating profit, but the retailer bears both the fixed cost and the unit variable cost. Only when the demand improvement effect can cover the investment cost does the blockchain investment mode become more advantageous.

Based on these conclusions, this study proposes the following managerial implications. First, suppliers should regard the direct-selling cost as an important basis for channel structure selection and dynamically adjust the channel structure according to boundary changes after blockchain introduction. When blockchain is not adopted, suppliers should carefully estimate direct-selling costs, including cold-chain fulfillment, packaging and sorting, online operations, after-sales service, and return handling. When the direct-selling cost is low, the direct-selling structure helps suppliers capture terminal profits. As the direct-selling cost increases, the e-commerce structure becomes more effective in alleviating the cost pressure of self-operated channels. After blockchain investment, suppliers should further reassess trust gain, demand improvement, and profit changes, determine whether the applicable region of the direct-selling structure has expanded, and adjust the channel structure in a timely manner when the direct-selling cost exceeds the new mode selection boundary.

Second, retailers should comprehensively weigh trust gain, the market size expansion effect, investment level, and cost constraints, rather than investing in blockchain solely because of its technological advantages. Blockchain investment is more likely to improve profits only when consumers respond strongly to traceability information, offline demand improves significantly, and both the fixed cost and unit variable cost remain within a controllable range. Before investing, retailers should estimate the relationship among demand increase, price improvement, and cost expenditure. When trust enhancement is limited or investment costs are high, retailers may reduce the blockchain investment level, postpone investment, or adopt a phased implementation approach. Compared with the direct-selling structure, the feasible region for blockchain investment is wider under the e-commerce structure, and retailers may moderately increase their willingness to invest.

Third, supply chain members should integrate channel structure selection with blockchain investment decisions to improve the matching between technology investment and channel structure. Blockchain investment can improve supply chain returns through trust enhancement, demand increase, and price improvement, but fixed costs, unit variable costs, and the variable cost sensitivity coefficient may weaken the conversion of these benefits into profits. When the supplier chooses the direct-selling structure, competition between online and offline channels may intensify, which can affect the retailer's investment returns. When the supplier chooses the e-commerce structure, the feasible region for retailer blockchain investment is generally wider. Therefore, both parties should determine the blockchain investment level within the range where demand improvement can cover investment costs, and improve the benefit conversion efficiency of blockchain investment by sharing demand data, jointly displaying traceability information, optimizing wholesale pricing, and coordinating promotional arrangements. In this way, blockchain investment can not only improve profit conversion efficiency, but also enhance traceability-based information governance, reduce quality uncertainty and transaction frictions, and support more sustainable operation of fresh agricultural product supply chains.

This study still has certain limitations. To highlight the core mechanism linking blockchain investment and dual-channel structure selection, this paper adopts a linear deterministic demand function, a setting that has been widely used in game-theoretic studies of dual-channel supply chains. However, fresh agricultural products are characterized by significant demand uncertainty, quality variability, and time-sensitive perishability, all of which may affect the pricing and investment decisions of supply chain members. Specifically, first, demand uncertainty may lead risk-averse decision makers to adopt more conservative pricing strategies, and the deterministic model used in this paper may therefore underestimate the critical threshold of the channel substitution parameter. Second, the perishability of fresh products implies that the value of unsold inventory decays over time, which may strengthen the retailer's reliance on the offline channel because of its immediacy in sales, thereby affecting the marginal returns to blockchain investment. Third, quality variability may increase consumers' demand for traceability information, in which case the trust gain coefficient assumed in this paper may be underestimated. Therefore, the conclusions of this study are more applicable to settings in which demand fluctuations for fresh products are relatively small and supply chain coordination is relatively strong. Future research may further test the robustness of these conclusions by introducing stochastic demand functions or quality decay functions, while also incorporating consumer heterogeneity, product differentiation, and differences in implementation contexts. In addition, differences in blockchain cost-bearing arrangements may also affect the applicable boundaries of the related conclusions. Existing studies show that different cost-bearing parties can influence blockchain investment conditions, profit distribution, and member incentives. Consistent with studies in which the retailer bears the cost, this study argues that the feasibility of retailer blockchain investment depends on whether the benefits from demand improvement can cover the investment costs. The difference is that this study further links blockchain investment decisions with the supplier's channel structure selection and analyzes their effects on the boundaries of mode selection. If the costs are borne by the supplier, the platform, or multiple parties, the corresponding investment constraints and mode selection thresholds may change. Future research may further extend the analysis in this direction.

References

• 1

  AcciariniC.CappaF.Di CostanzoG.PriscoM.SardoF.StazzoneA.et al. (2023). Blockchain technology to protect label information: the effects on purchase intentions in the food industry. Comput. Ind. Eng.180:109276. doi: 10.1016/j.cie.2023.109276

  • CrossRef
  • Google Scholar
  • View reference in article

• 2

  BatarfiR.JaberM. Y.ZanoniS. (2016). Dual-channel supply chain: a strategy to maximize profit. Appl. Math. Model.40, 9454–9473. doi: 10.1016/j.apm.2016.06.008

  • CrossRef
  • Google Scholar
  • View reference in article

• 3

  ChenX.WangX.JiangX. (2016). The impact of power structure on the retail service supply chain with an O2O mixed channel. J. Oper. Res. Soc.67, 294–301. doi: 10.1057/jors.2015.6

  • CrossRef
  • Google Scholar
  • View reference in article

• 4

  ChiangW. K.ChhajedD.HessJ. D. (2003). Direct marketing, indirect profits: a strategic analysis of dual-channel supply-chain design. Manag. Sci.49, 1–20. doi: 10.1287/mnsc.49.1.1.12749

  • CrossRef
  • Google Scholar
  • View reference in article

• 5

  China Merchants Industry Research Institute (2025). Market Size and Trend Forecast of China's Fresh Agricultural Products Industry 2025. Shenzhen: China Merchants Industry Research Institute.

  • Google Scholar
  • View reference in article

• 6

  CompagnucciL.LeporeD.SpigarelliF.FrontoniE.BaldiM.Di BerardinoL. (2022). Uncovering the potential of blockchain in the agri-food supply chain: an interdisciplinary case study. J. Eng. Technol. Manag.65:101700. doi: 10.1016/j.jengtecman.2022.101700

  • CrossRef
  • Google Scholar
  • View reference in article

• 7

  de CarvalhoP. R. V.Naoum-SawayaJ.ElhedhliS. (2022). Blockchain-enabled supply chains: an application in fresh-cut flowers. Appl. Math. Model.110, 841–858. doi: 10.1016/j.apm.2022.06.011

  • CrossRef
  • Google Scholar
  • View reference in article

• 8

  DoanhD. C.TramT. B.VuP. Q.LinhL. T. N.HuongN. T. T. (2026). Blockchain-based food traceability system and green brand image: enhancing product trust and purchase intentions for online agricultural products. Clean. Responsib. Consump.20:100368. doi: 10.1016/j.clrc.2025.100368

  • CrossRef
  • Google Scholar

• 9

  DumrongsiriA.FanM.JainA.MoinzadehK. (2008). A supply chain model with direct and retail channels. Eur. J. Oper. Res.187, 691–718. doi: 10.1016/j.ejor.2006.05.044

  • CrossRef
  • Google Scholar

• 10

  E-commerce Research Center NetEconomy (2025). Annual Report on China's Fresh E-commerce Industry 2024. Hangzhou: E-commerce Research Center, NetEconomy.

  • Google Scholar
  • View reference in article

• 11

  FanX.LiuY. (2016). Pricing and service competition strategies in a dual-channel supply chain with manufacturer online channel introduction. Chin. J. Manag. Sci.24, 143–148. doi: 10.16381/j.cnki.issn1003-207x.2016.07.017

  • CrossRef
  • Google Scholar

• 12

  Foodaily (2023). Zespri CEO: 40% of Our Business in China is Conducted Through E-commerce. Foodaily. Available online at: https://www.foodaily.com/newsflashes/6989 (Accessed June 3, 2026).

  • Google Scholar

• 13

  Fuping County Market Supervision Administration (2025). Procurement Intention for the Fuping Persimmon Cake Blockchain Traceability Intelligent Platform Project. Fuping County Market Supervision Administration. Available online at: https://www.sxggzyjy.cn/jydt/001001/001001004/001001004006/20251105/8a69c4f59a375604019a53690785713e.html (Accessed June 3, 2026).

  • Google Scholar

• 14

  GanW.HuangB. (2022). Exploring data integrity of dual-channel supply chain using blockchain technology. Comput. Intell. Neurosci.2022:3838282. doi: 10.1155/2022/3838282

  • CrossRef
  • Google Scholar
  • View reference in article

• 15

  GaoC.XiangC. (2025). Mechanism of the influence of agricultural product traceability information characteristics on consumers' purchase intention. J. Huazhong Agric. Univ.2025, 121–133. doi: 10.13300/j.cnki.hnwkxb.2025.01.011

  • CrossRef
  • Google Scholar

• 16

  HeD.YuY.XiangZ.ZuoM. (2026). Chinese consumer preferences and willingness to pay for blockchain-traceable aquatic products. Front. Sustain. Food Syst.9:1728028. doi: 10.3389/fsufs.2025.1728028

  • CrossRef
  • Google Scholar
  • View reference in article

• 17

  HeY.ChenZ.LiaoN. (2022). Impact mechanism of government subsidy modes on emission reduction decisions of manufacturers in green supply chains. Chin. J. Manag. Sci.30, 87–98. doi: 10.16381/j.cnki.issn1003-207x.2019.1854

  • CrossRef
  • Google Scholar
  • View reference in article

• 18

  HsiaoL.ChenY. J. (2014). Strategic motive for introducing internet channels in a supply chain. Prod. Oper. Manag.23, 36–47. doi: 10.1111/poms.12051

  • CrossRef
  • Google Scholar
  • View reference in article

• 19

  HuH.CaoY.WuK. (2022). Sales mode selection and pricing of a fresh produce supply chain based on presale. Oper. Res. Manag. Sci.31, 128–134.

  • Google Scholar
  • View reference in article

• 20

  JiangY.LiuC.BaiS.LiH. (2023). Optimal decision-making of a dual-channel fresh agricultural product supply chain using blockchain. Syst. Eng.41, 63–72.

  • Google Scholar
  • View reference in article

• 21

  JinS.CaoY.JonesG.LiW.FrewerL. J. (2023). Consumers' purchase intentions towards traced foods: a comparative analysis between the United Kingdom and China. Food Control152:109828. doi: 10.1016/j.foodcont.2023.109828

  • CrossRef
  • Google Scholar
  • View reference in article

• 22

  JoshiR.BasuS.JonnalageddaS.AvittathurB. (2023). Multichannel retailer's channel choice and product pricing: influence of investment in fit-disclosing technology by competing retailers. Int. J. Prod. Econ.262:108895. doi: 10.1016/j.ijpe.2023.108895

  • CrossRef
  • Google Scholar
  • View reference in article

• 23

  LiW.ZhaoS. (2019). Decision-making of a dual-channel supply chain for high-quality agricultural products considering traceability system application. Oper. Res. Manag. Sci.28, 98–108.

  • Google Scholar
  • View reference in article

• 24

  LiY.TanC.IpW. H.WuC.-H. (2023). Dynamic blockchain adoption for freshness-keeping in the fresh agricultural product supply chain. Expert Syst. Appl.217:119494. doi: 10.1016/j.eswa.2022.119494

  • CrossRef
  • Google Scholar
  • View reference in article

• 25

  LiangX.XiaoJ. (2023). Pricing and channel selection of a dual-channel supply chain based on blockchain and consumer sensitivity. Chin. J. Manag. Sci.31, 29–38. doi: 10.16381/j.cnki.issn1003-207x.2020.1755

  • CrossRef
  • Google Scholar
  • View reference in article

• 26

  LiangX.ZhangY. (2020). Pricing and emission reduction strategies in a low-carbon dual-channel supply chain based on consumer preferences. Oper. Res. Manag. Sci.29, 107–117.

  • Google Scholar

• 27

  LinQ.ChenL.LinX. (2023). Sales mode selection of fresh e-commerce under “quality–price” competition. Oper. Res. Manag. Sci.32, 22–28.

  • Google Scholar

• 28

  LiuL.LiF. (2022). Blockchain technology investment decision and coordination in fresh supply chains considering retailer risk aversion. J. Ind. Eng. Eng. Manag.36, 159–171. doi: 10.13587/j.cnki.jieem.2022.01.014

  • CrossRef
  • Google Scholar

• 29

  LiuY.MaD.HuJ.ZhangZ. (2021). Sales mode selection of fresh food supply chain based on blockchain technology under different channel competition. Comput. Ind. Eng.162:107730. doi: 10.1016/j.cie.2021.107730

  • CrossRef
  • Google Scholar

• 30

  LuQ.LiaoC.ChenX. (2024). The blockchain adoption strategies of online retailer in a dual-channel supply chain. Int. J. Prod. Econ.274:109322. doi: 10.1016/j.ijpe.2024.109322

  • CrossRef
  • Google Scholar

• 31

  MaY.PengH.FengX.JinT. (2025). Blockchain adoption strategy in a dual-channel agricultural product supply chain considering counterfeiting behavior. J. Ind. Eng. Eng. Manag.39, 178–191. doi: 10.13587/j.cnki.jieem.2025.06.013

  • CrossRef
  • Google Scholar

• 32

  MengX.LiuZ.ZhangL.LiY. (2019). The influence of information source trust on consumers' willingness to purchase traceable pork: empirical evidence from 1009 consumer questionnaires in Shanghai and Jinan. Acta Agric. Shanghai35, 107–114. doi: 10.15955/j.issn1000-3924.2019.02.20

  • CrossRef
  • Google Scholar

• 33

  ModakN. M.SenapatiT.SimicV.PamucarD.SahaA.Cárdenas-BarrónL. E. (2024). Managing a sustainable dual-channel supply chain for fresh agricultural products using blockchain technology. Expert Syst. Appl.244:122929. doi: 10.1016/j.eswa.2023.122929

  • CrossRef
  • Google Scholar

• 34

  ShaoS.LiY.ChengX.QuJ. (2025). Green investment strategies and pricing decisions in a supply chain considering blockchain technology. Sustainability17:6491. doi: 10.3390/su17146491

  • CrossRef
  • Google Scholar

• 35

  StranieriS.RiccardiF.MeuwissenM. P.SoregaroliC. (2021). Exploring the impact of blockchain on the performance of agri-food supply chains. Food Control119:107495. doi: 10.1016/j.foodcont.2020.107495

  • CrossRef
  • Google Scholar

• 36

  SunS.PuX.HanG. (2019). Manufacturer online sales mode selection considering offline power structure. Chin. J. Manag. Sci.27, 119–129. doi: 10.16381/j.cnki.issn1003-207x.2019.05.013

  • CrossRef
  • Google Scholar

• 37

  SunY.SongX.JiangY.GuoJ. (2023). Strategy analysis of fresh agricultural enterprises in a competitive circumstance: the impact of blockchain and consumer traceability preferences. Mathematics11:1090. doi: 10.3390/math11051090

  • CrossRef
  • Google Scholar

• 38

  TanC.ZengY. (2023). Two-stage pricing strategy of an O2O fresh agricultural product supply chain based on blockchain technology. Chin. J. Manag.20, 904–915.

  • Google Scholar

• 39

  TianY.DanB.LiuM.MaS. (2022). Mode selection and coordination of community fresh O2O supply chains with demand influenced by fresh-keeping investment. Chin. J. Manag. Sci.30, 173–184. doi: 10.16381/j.cnki.issn1003-207x.2019.1602

  • CrossRef
  • Google Scholar

• 40

  TranD.SchoutetenJ. J.GellynckX.De SteurH. (2024). How do consumers value food traceability? A meta-analysis. Food Control162:110453. doi: 10.1016/j.foodcont.2024.110453

  • CrossRef
  • Google Scholar

• 41

  WangD.ZhuM.DongH. (2025). Supply chain operation decisions based on commercial credit financing under blockchain environment. J. Ind. Eng. Eng. Manag.39, 243–255. doi: 10.13587/j.cnki.jieem.2025.01.017

  • CrossRef
  • Google Scholar

• 42

  WangY.MeiS.XuR.YangD.ZhongW. (2023). Channel competition in omni-channel supply chain considering social media advertising. Manag. Decis. Econ.44, 3354–3366. doi: 10.1002/mde.3883

  • CrossRef
  • Google Scholar

• 43

  XuJ.XiongS.CuiT.ZhangD.LiZ. (2023). Incorporating consumers' low-carbon and freshness preferences in dual-channel agri-foods supply chains: an analysis of decision-making behavior. Agriculture13:1647. doi: 10.3390/agriculture13091647

  • CrossRef
  • Google Scholar

• 44

  XuY.WangJ.CaoK. (2024). Interaction between joining platform blockchain technology and channel encroachment for fresh agricultural product firms. Int. Trans. Oper. Res.31, 3565–3591. doi: 10.1111/itor.13266

  • CrossRef
  • Google Scholar

• 45

  YanR.PeiZ. (2009). Retail services and firm profit in a dual-channel market. J. Retail. Cons. Serv.16, 306–314. doi: 10.1016/j.jretconser.2009.02.006

  • CrossRef
  • Google Scholar

• 46

  YangL.TangR. (2019). Comparisons of sales modes for a fresh product supply chain with freshness-keeping effort. Transp. Res. E Logist. Transp. Rev.125, 425–448. doi: 10.1016/j.tre.2019.03.020

  • CrossRef
  • Google Scholar

• 47

  YeJ.GuB.FuY. (2023). Cold chain logistics service and pricing decisions in fresh agricultural product supply chains under different trade modes. Chin. J. Manag. Sci.31, 95–107. doi: 10.16381/j.cnki.issn1003-207x.2020.0751

  • CrossRef
  • Google Scholar

• 48

  YueL.LiuY.ZhuG. (2016). Research on coordination contracts in retailer-led fresh dual-channel supply chains. Soft Sci.30, 123–128, 144. doi: 10.13956/j.ss.1001-8409.2016.08.27

  • CrossRef
  • Google Scholar

• 49

  ZhaiQ.SherA.LiQ. (2022). The impact of health risk perception on blockchain traceable fresh fruits purchase intention in China. Int. J. Environ. Res. Public Health19:7917. doi: 10.3390/ijerph19137917

  • CrossRef
  • Google Scholar

• 50

  ZhangQ.TangW.ZaccourG.ZhangJ. (2019). Should a manufacturer give up pricing power in a vertical information-sharing channel?Eur. J. Oper. Res.276, 910–928. doi: 10.1016/j.ejor.2019.01.054

  • CrossRef
  • Google Scholar

• 51

  ZhangW.WuL.JiL. (2025). Blockchain technology adoption strategies for the shipping costs bearer in the fresh product supply chain. Front. Sustain. Food Syst.9:1550985. doi: 10.3389/fsufs.2025.1550985

  • CrossRef
  • Google Scholar

• 52

  ZhaoL. (2015). Pricing decision of a hybrid channel supply chain with manufacturer online direct selling. Chin. J. Manag. Sci.23, 557–563. doi: 10.13587/j.cnki.jieem.2023.04.014

  • CrossRef
  • Google Scholar

• 53

  ZhengQ.FanT.HuB.SongY. (2023). Traceability technology investment decision in fresh agricultural product supply chains under the front warehouse model. J. Ind. Eng. Eng. Manag.37, 165–178.

  • Google Scholar