Manufacturer’s Channel Strategy and Demand Information Sharing in a Retailer-Led Green Supply Chain

Abstract

In the rapidly evolving landscape of e-commerce, companies are increasingly focusing on their channel strategies to gain sustainable development. However, asymmetric demand information poses challenges to these decisions. This paper explores the interplay between a retailer’s information sharing strategy and a manufacturer’s channel strategy in a retailer-led green supply chain, where the manufacturer may establish an online channel to sell its green products directly. The dominant retailer has private demand information about the market and decides whether to share private information with the manufacturer. By establishing a game model, we analyze the impacts of information sharing and the manufacturer’s channel strategy on the payoffs for all the supply chain members, considering sustainability aspects such as the environmental benefits of green products and the efficiency of supply chain operations. The results show that information sharing benefits both the retailer and the manufacturer, irrespective of the establishment of an online channel. When the retailer shares demand information, opening an online channel benefits the manufacturer but benefits the retailer under certain conditions. Furthermore, through a numerical approach, we examine the strategic preferences of the firms and derive the equilibrium strategy. Interestingly, the manufacturer consistently prefers the scenario involving both an online channel and information sharing. The retailer’s preference, however, depends on the direct selling cost; it favors information sharing with or without an online channel based on this cost. Ultimately, our findings suggest that the equilibrium strategy can either be sharing information with an online channel or not sharing information without an online channel, which is contingent upon the direct selling cost and the forecast signal precision. These insights provide actionable strategies for enhancing the sustainability of supply chain operations.

1. Introduction

With the ongoing deterioration of the environment and the depletion of natural resources, consumers are becoming increasingly aware of environmental protection and are beginning to prioritize purchasing environmentally friendly products. This shift in consumer behavior has promoted many manufacturers to produce green products to attract these environmentally conscious consumers and capture more market share. Companies like Haier have responded by launching green product lines. Moreover, in the era of the rapid development of the internet and logistics, many upstream manufacturers have begun implementing online channels to complement their existing retail channels or expand market demand [1]. Since the COVID-19 epidemic in 2020, more and more manufacturers have turned to open online channels in addition to retail channels. Establishing an online selling channel can provide a platform to attract more consumers and increase upstream manufacturers’ incomes. As many large retailers emerge (e.g., Walmart and Best Buy), some researchers have studied the supply chain where the retailer is the leader [2,3]. While opening online channels can reduce double marginalization and benefit firms in manufacturer-led supply chains [4,5,6], it may harm retailers in retailer-led supply chains [7]. These conflicting findings highlight a gap in the current research: the need to understand the nuanced effects of online channel strategies in retailer-led supply chains, particularly in the context of green product markets.

In reality, the upstream manufacturer, who is away from end consumers, cannot obtain market demand information, while the downstream retailer, who is close to consumers, has better access to market conditions (e.g., the retailer has point-of-sale data and insight into consumer preferences). Knowing the true market demand information can help the manufacturer make better decisions regarding channel strategies, product green levels, and pricing. With the rapid development of information technology, it is more common for downstream retailers to share demand information with upstream manufacturers. According to a Chinese bank report, 61% of the surveyed firms said that information sharing is essential for their business success. The annual surveys by Bearing Point reported rapid growth in the frequency and scope of communication between downstream retailers and upstream manufacturers in the United States [8]. For example, Walmart shares its weekly sales information with its suppliers to coordinate their supply chains.

Practical examples and existing research indicate that downstream retailers often share private demand information with upstream manufacturers. However, this decision requires careful consideration, especially if the manufacturer might establish an online channel and determine the green level of products. In a retailer-led supply chain, sharing demand information allows the manufacturer to better decide on channel strategies and product green levels. Demand information sharing is crucial for manufacturers in deciding whether to open an online channel, potentially leading to a win–win outcome for both parties [9]. Additionally, it also enables manufacturers to make informed decisions about product green levels based on true market demand. For example, Haier leverages market demand data to sell products on e-commerce platforms like JD.com and Tmall.com, improving product green levels and benefiting retailers. Thus, retailers must balance the potential gains from improved product green levels against the potential losses from the manufacturer opening an online channel. Simultaneously, manufacturers must consider the impact of establishing an online channel. Therefore, we study the interplay between the retailer’s information sharing strategy and the manufacturer’s channel strategy, providing the motivation for our research. Based on the aforementioned considerations, this paper investigates the interaction between the retailer’s information sharing strategy and the manufacturer’s channel strategy in a retailer-led green supply chain and aims to address the following questions: (i) How do the manufacturer’s channel strategy and information sharing affect the profits of supply chain members and decisions regarding product green level? (ii) Under the combined effects of channel strategy and information sharing, what are the equilibrium strategies for the dominant retailer and the manufacturer?

To address these questions, we initially construct game models to analyze the interactions between the retailer and the manufacturer and derive equilibrium decisions under different strategies. The results show that information sharing always improves the green level of products and benefits both the manufacturer and the retailer, regardless of whether the manufacturer establishes an online channel. However, the impact of opening an online channel on the optimal green level and the profits of the supply chain members depends on specific parameter conditions. When the retailer shares demand information, opening an online channel always improves the green level and benefits the manufacturer, but it benefits the retailer only when the direct selling cost is high. In contrast, when the retailer does not share demand information, the impact of opening an online channel depends on the direct selling cost and the demand forecast signal precision. Furthermore, we explore the strategy preferences of the manufacturer and retailer, deriving equilibrium strategies for both firms through numerical analysis, and suggest that the manufacturer prefers the scenario of opening an online channel with information sharing, while the retailer prefers information sharing regardless of the manufacturer’s channel decision. Importantly, the retailer prefers to share demand information to inspire the manufacturer to establish an online channel when the direct selling cost is relatively high and to withhold demand information to prevent the manufacturer from opening an online channel when the direct selling cost is relatively low and the forecast signal precision is relatively high.

In summary, our main contributions are as follows. First, this paper provides a comprehensive analysis of the strategic interactions between the retailer and the manufacturer in a green supply chain context, highlighting the dual roles of information sharing and channel strategies. Second, the equilibrium strategies derived from the model provide a nuanced understanding of when and how information sharing and online channel establishment should be pursued, adding valuable knowledge to the field of supply chain management. Finally, we offer practical insights for the manufacturer and the retailer on optimizing their strategies to achieve a win–win outcome, considering factors such as the direct selling cost and forecast signal precision.

This paper is organized as follows. We review the related literature in Section 2. Section 3 describes the basic model. In Section 4, we derive the equilibrium decisions and further analyze the equilibrium outcomes. Section 5 extends our model to change the game sequence. Finally, we conclude the paper with a summary and provide some potential research directions in Section 6. Proofs of all the propositions are arranged in Appendix A.

2. Literature Review

Our work is closely related to three main streams of literature: green products, the manufacturer’s decision to open an online channel, and information sharing.

2.1. Green Products

With increasing consumer environmental awareness, manufacturers are beginning to invest in emission-reduction technologies [10], focusing on the production of green products. A substantial body of literature extensively studies green products, which can be classified into two types. The first type focuses on manufacturers selling green products only through the traditional retail channel. For instance, Zhang and Liu [11] explore the pricing decision of green products in a three-level supply chain with a nonlinear demand function, while Zhang et al. [12] investigate pricing decisions for both green and non-green products in a supply chain with a linear demand function. Ghosh and Shah [13], Zhu and He [14], and Chen and Ulya [15] also study product pricing and green level decisions in supply chains with linear demand functions. Fang and Xu [16] and Heydari et al. [17] study a two-level SC where demand is a function of the selling price and the product’s green quality. Ma et al. [18] examine the impact of information superiority, green optimism, and power structures on the decision making and performance of a green supply chain comprising a manufacturer and a retailer.

The second type examines manufacturers selling green products through both traditional retail and online channels. For example, Li et al. [19] study the supplier who decides whether to introduce an online direct pattern in addition to the existing off-line retail pattern when the green level is endogenous; in addition, Yang et al. [20] also investigate this issue by considering the green level as endogenous and exogenous, respectively. Heydari et al. [21], Ranjan and Jha [22], and Aslani and Heydari [23] study the pricing and coordination strategies in supply chains in which the supplier sells green products through both an off-line retail channel and an online channel. Zhang et al. [24] explore the greening issues and dynamic pricing strategy in a two-echelon dual-channel supply chain. Pal and Sarkar [25] consider retailer promotional efforts for a green supply chain with a dual-channel structure. Li et al. [26] study how the manufacturer makes joint decisions on green product development, pricing, and the online selling format in a dual-channel supply chain with green products.

The researchers above assume information symmetry among green supply chain members. However, in real life, the green supply chain often faces the challenge of information asymmetry. Xia and Niu [27], Yu et al. [28], and Cai et al. [29] have studied the game-theoretical models in green supply chains with information asymmetry. However, they do not consider the impact of the channel strategy and information sharing strategy on pricing and green level decisions. Our research fills this gap, and our findings provide green supply chain members with useful guidelines on the channel strategy and information sharing strategy.

2.2. Manufacturer’s Decision to Open an Online Channel

Our research is associated with the manufacturer’s decision to introduce an online channel (referred to as “encroachment”), which has two main findings. On one hand, most of the literature shows that opening an online channel may reduce the retailers’ market share and lead to channel conflict [30,31]. On the other hand, some scholars indicate that opening an online channel can reduce the double marginalization effect and benefit retailers [4,6,32,33,34,35]. For example, Chiang et al. [4] find that the retailer may benefit from encroachment due to a wholesale price reduction. Tsay and Agrawal [32] show that encroachment benefits the retailer when the manufacturer promotes the product. Arya et al. [6] demonstrate that encroachment can benefit both firms when the manufacturer is an ineffective retail competitor. On this basis, Li et al. [33] further study supplier encroachment under asymmetric demand information and show that encroachment can lead to win–win, win–lose, lose–win, or lose–lose outcomes. Yoon [34] indicates that the encroachment is not always detrimental to the retailer when the manufacturer invests in the model. Li et al. [35] discuss how manufacturers offering substitutable green products through dual channels adopt different encroachment strategies.

Most research on encroachment assumes manufacturer/supplier leadership. However, with the rise of large retailers like Amazon and Walmart, some studies explore encroachment in retailer-led supply chains. For instance, Zhang et al. [7] examine retailer strategies to prevent manufacturer encroachment in a retailer-led supply chain, considering private direct selling costs and retailer incentives for investing in a retail service. Motivated by Zhang et al. [7], our paper investigates the manufacturer’s decision to open an online channel in a retailer-led supply chain. However, the difference is that we explore the impact of the retailer’s information sharing on the manufacturer’s decision to open an online channel.

2.3. Information Sharing

The present paper is also related to the research on information sharing. In marketing management and operational research, information sharing in supply chains has been extensively studied and can be classified into three types. The first type is bilateral information sharing, where both the upstream and downstream supply chain members have demand information [36,37,38]. The second type is a scenario where only the upstream member has information [39,40]. The third type involves a scenario where only the downstream member has information [9,41,42], which is the most related to our work. Our work is closed related to the research on the interplay between supplier/manufacturer encroachment and demand information sharing. For example, Tsunoda and Zennyo [43] consider a setting where a supplier wants to sell online via a platform and explore how the platform’s information sharing policy affects the supplier’s selection of a selling mode. Ha et al. [44] focus on the supplier-led scenario and examine how an e-tailer’s information sharing affects the supplier’s incentive to encroach through an agency channel. Tang et al. [45] explore how the leadership scenario influences the supplier’s encroachment channel selection by considering the e-tailer’s information sharing. Liu et al. [46] study the e-retailer’s information sharing strategy with supplier encroachment in an e-tailing supply chain for fresh produce. Unlike the prior literature, which assumes supplier/manufacturer-led supply chains with information sharing, this paper examines information sharing in a retailer-led green supply chain where the manufacturer has the incentive to establish an online channel.

Table 1. The literature review.

References	Green Product	Power Structure	Information Sharing	Channel Entry
Aslani and Heydari [23]	Yes	Supplier-led	No	No
Zhang et al. [25]	Yes	Manufacturer-led	No	No
Li et al. [26]	Yes	Manufacturer-led	No	No
Yu et al. [28]	Yes	Supplier-led	Yes	No
Li et al. [35]	Yes	Manufacturer-led	No	Yes
Zhang et al. [7]	No	Retailer-led	No	Yes
Tsunoda and Zennyo [43]	No	Supplier-led	Yes	Yes
Ha et al. [44]	No	Supplier-led	Yes	No
Tang et al. [45]	No	Supplier-led	Yes	No
Liu et al. [46]	No	Supplier-led	Yes	Yes
This paper	Yes	Retailer-led	Yes	Yes

summarizes the critical difference between our work and the most closely related studies mentioned above.

3. Model

3.1. Problem Description and Assumptions

We consider a supply chain consisting of a manufacturer (he, denoted as $m$) and a dominant retailer (she, denoted as $r$). The retailer, acting as a Stackelberg leader with private demand information, purchases a green product from the manufacturer at a wholesale price $w$ and then resells it with a retail price ($p_r$) to the end consumers. The manufacturer, being the follower, produces a green product with the green level ($g$) and may establish an online channel to directly sell his green product to the end consumer at an online selling price ($p_m$) (i.e., opening an online channel, denoted as $O$). Following Arya et al. [6] and Li et al. [33], we normalize both the manufacturer’s production cost and the retailer’s selling cost to zero. Additionally, we assume that the manufacturer incurs an additional per-unit online selling cost ($c$) due to a potential lack of marketing experience and skills.

Table 2. Notations and descriptions.

Notations	Descriptions
$\eta$	retail margin, decided by the retailer
$w$	wholesale price, decided by the manufacturer
$p_m$	online selling price, decided by the manufacturer
$g$	green level of product, decided by the manufacturer
$p_r$	$\mathrm{retail} \mathrm{price} \mathrm{of} \mathrm{product}, p_r=w+\eta$
$c$	per-unit online selling cost
$\beta$	signal precision
$\theta$	consumer’s acceptance on the online channel.
$a\in \{H,L\}$	market base (high or low state of demand)
$Y\in \{h,l\}$	demand signal (high or low forecast signal)
${\pi }_m$	manufacturer’s profit
${\pi }_r$	retailer’s profit
${\Pi }_m$	manufacturer’s expected profit
${\Pi }_r$	retailer’s expected profit
$N,I$	no information sharing, information sharing
$\overline{O},O$	without online channel, with online channel

summarizes the notations adopted in this paper.

In addition, it is supposed that the manufacturer’s cost for achieving greening improvement follows a quadratic function of the green level $g$, represented as $C\left(g\right)={\displaystyle \frac{1}{2}}\mathrm{k}{\mathrm{g}}^2$. This assumption is widely accepted in the literature [47,48]. Here, $k$ represents an investment scaling parameter, which is sufficiently large and satisfies $k>{\displaystyle \frac{2}{3}}a$.

3.2. Demand

Let $v$ denote the consumers’ reservation value, which also represents the consumers’ willingness to pay for the product. Each consumer in the market has at most one unit demand for the product. We assume that the consumers’ reservation value $v$ is uniformly distributed in [0, 1], which is common in the literature [49,50]. In line with Qi et al. [51], the consumers’ utility in the retail channel is given by $U_r=v+g-p_r.$ If the manufacturer does not establish an online channel, a consumer will buy through the retailer if $U_r>0$ is satisfied. Consequently, the fraction of consumers in the retail channel when the manufacturer does not have an online channel can be obtained as follows:

$${\lambda }_r^{\overline{O}}=1-p_r+g,$$

(1)

where the superscript “$\overline{O}$” denotes the case in which the manufacturer does not establish an online channel.

When the manufacturer establishes an online channel, the utility of a consumer buying green products through the online channel is $U_m=\mathsf{\theta }\left(v+g\right)-p_m$. The parameter $\mathsf{\theta }\in \left(\mathrm{0,1}\right)$ represents the discount factor of consumer utility obtained from products sold through the online channel and can be explained as the consumer’s acceptance on the online channel. Usually, consumers strategically choose between retail and online channels to maximize their utility. Based on the utility functions, a consumer will buy from the retailer when the following conditions are satisfied:

$$\left\{\begin{array}{l}{U}_r>0\Rightarrow v>p_r-g \\ U_r>U_m\Rightarrow v>{\displaystyle \frac{p_r-p_m+\left(\theta -1\right)g}{1-\theta }} \end{array}\right.$$

Similarly, a consumer will purchase via the online channel when the following conditions are satisfied:

$$\left\{\begin{array}{l}{U}_m>0\Rightarrow v>{\displaystyle \frac{p_m-\theta g}{\theta }} \\ U_m>U_r\Rightarrow v<{\displaystyle \frac{p_r-p_m+\left(\theta -1\right)g}{1-\theta }} \end{array}\right.$$

Similar to Mantin et al. [52], this paper focuses on $\mathsf{\theta }>{\displaystyle \frac{p_m}{p_r}}$ to guarantee the coexistence of products in both the retail and direct channels. Thus, when the manufacturer establishes an online channel, we can obtain the following consumers’ fractions in the retail and online channels, respectively

$${\lambda }_r^O=1-{\displaystyle \frac{p_r-p_m+\left(\theta -1\right)g}{1-\theta }},$$

(2)

$${\mathsf{\lambda }}_m^O={\displaystyle \frac{p_r-p_m+\left(\mathsf{\theta }-1\right)g}{1-\mathsf{\theta }}}-{\displaystyle \frac{p_m-\mathsf{\theta }g}{\mathsf{\theta }}},$$

(3)

where the superscript “$O$” represents the scenario in which the manufacturer opens an online channel.

The total consumer market demand, denoted as $a$, is uncertain and random. Following Iyer et al. [53], we assume $a$ has two states:$a=H$ and $a=L$, which occur with equal probability, i.e., $\mathrm{Pr}\left(a=H\right)=\mathrm{Pr}\left(a=L\right)={\displaystyle \frac{1}{2}}$. Both the retailer and the manufacturer know the prior distribution of $a$. We normalize the low state of demand as $L=1$ and denote the high state of demand as $H>1$. $H$ represents the degree of demand uncertainty.

The retailer, who is closer to the end consumers, can observe a private demand forecast signal $Y$ about uncertain market demand, which is common in the literature regarding information sharing [46,54]. The forecast signal $Y$ has two possible values: $Y=h$ represents a high forecast signal, implying high demand, and $Y=l$ represents a low forecast signal, implying low demand. Similar to Guo and Iyer [55] and Li and Zhang [56], each signal $Y\in \{h, l\}$ is independently generated from the actual demand state $a\in \{H, L\}$ with probability $\beta \in \left({\displaystyle \frac{1}{2}},1\right]$, i.e., $\mathrm{Pr}\left(h|H\right)=\mathrm{Pr}\left(l|L\right)=\beta$, which has been widely used in the literature [39,53]. Parameter $\beta$ measures the forecast signal precision. When $\beta =1$, the signal is perfectly reliable; when $\beta ={\displaystyle \frac{1}{2}}$, the signal is uninformative. Based on the aforementioned information, we derive the following properties:

(1)

The probability of the signal $Y\in \{h, l\}$:

$$\mathrm{Pr}\left(h\right)=\mathrm{Pr}\left(H\right)\mathrm{Pr}\left(h|H\right)+\mathrm{Pr}\left(L\right)\mathrm{Pr}\left(h|L\right)={\displaystyle \frac{1}{2}}, \mathrm{Pr}\left(l\right)=\mathrm{Pr}\left(H\right)\mathrm{Pr}\left(l|H\right)+\mathrm{Pr}\left(L\right)\mathrm{Pr}\left(l|L\right)={\displaystyle \frac{1}{2}}.$$

(2)

The updated probabilities of demand states conditional on the signal $Y\in \{h, l\}$:

$$\mathrm{Pr} \left(H|h\right)=\mathrm{Pr}\left(L|l\right)=\beta , \mathrm{Pr}\left(H|l\right)=\mathrm{Pr}\left(L|h\right)=1-\beta .$$

(3)

The expected $a$ conditional on the signal $Y\in \{h, l\}$ (i.e., $\mathrm{E}\left[a|Y\right]=a_Y$):

$$a_h=\mathrm{E}\left[a|h\right]=1+\left(H-1\right)\beta ,{ a}_l=\mathrm{E}\left[a|l\right]=H-\left(H-1\right)\beta .$$

3.3. Timeline of the Game

The sequence of events is shown in Figure 1. First, before observing actual demand information, the retailer decides ex ante whether to share the demand information (denoted by $I$) or not (denoted by $N$).

Once the retailer commits to the information sharing policy, which is binding and irreversible, she must adhere to it. Otherwise, she will withhold the demand information privately. This setting is widely adopted in the literature [9,57]. Second, the retailer observes the private demand information before the selling season and decides whether to share it with the manufacture. Then, the manufacturer determines whether to establish an online channel based on the retailer’s information sharing decision, which is similar to that in Huang et al. [9] ($\overline{O}$ represents the case without an online channel; $O$ represents the case with an online channel). Third, the retailer decides her retail margin $\eta$ (i.e., $\eta =p_r-w$). Finally, the manufacturer determines the green level ($g$) and the wholesale price ($w$). If an online channel exists, the manufacturer also decides the online selling price ($p_m)$. To consider the scenario in which the retailer shares demand information and the manufacturer opens an online channel, this paper considers four possible cases: Case $N\overline{O}$, Case $NO$, Case $I\overline{O}$, and Case $IO$.

4. Analysis

In this section, we formulate the retailer’s and the manufacturer’s optimal decisions and profits under Case $N\overline{O}$, Case $NO$, Case $I\overline{O}$, and Case $IO$ by backward induction and then compare and analyze them. Furthermore, we derive the firms’ strategy preferences and the equilibrium strategy.

4.1. Equilibrium Solutions

When the retailer does not share demand information, the manufacturer can infer demand from the retailer’s retail margin, leading to a signaling game with two types of equilibria: separating and pooling. Under separating equilibria, the low-type retailer (who observes a low market size) and the high-type retailer (who observes a high market size) set a distinct retail margin for each market size. Under pooling equilibria, the retailer sets the same retail margin whether demand is high or low, and the manufacturer cannot obtain any demand information according to the retailer’s retail margin; thus, the manufacturer makes decisions relying only on the prior information. In this paper, we refine the multiple equilibria using the lexicographically maximum sequential equilibrium (LMSE) concept [58], which has been widely used in the literature [39,59]. Essentially, the rationale of LMSE is that the retailer selects the most profitable equilibrium from the separating equilibria and the pooling equilibria.

4.1.1. Case $\mathrm{N}\overline{\mathrm{O}}$

When the retailer does not share information and the manufacturer does not open an online channel, the retailer first decides her retail margin ($\eta$) based on her private forecast signal $Y\in \{h, l\}$; then, the manufacturer decides the green level ($g$) and the wholesale price ($w$) of the product according to the demand information inferred from the retailer’s retail margin. The expected profits of the retailer and manufacturer are

$$\mathrm{E}\left[{\pi }_r|Y\right]=\eta {\lambda }_r^{\overline{O}}\mathrm{E}\left[a|Y\right],$$

(4)

$$\mathrm{E}\left[{\pi }_m|\eta \right]=w{\lambda }_r^{\overline{O}}\sum _{Y=h, l}\mathrm{Pr}\left(Y|\eta \right)E\left[a\right|Y]-{\displaystyle \frac{1}{2}}\mathrm{k}{\mathrm{g}}^2.$$

(5)

By backward induction, we summarize the unique LMSE equilibrium solutions in the following Lemma 1.

Lemma 1. 

Under case $N\overline{O}$,

(a)

There exists a unique LMSE outcome, and the LMSE outcome is pooling.

(b)

The optimal retail margin, green level of product, and wholesale price, respectively, are

$${\eta }^{N\overline{O}}={\displaystyle \frac{1}{2}}, g^{N\overline{O}}={\displaystyle \frac{1+H}{2\left(4k-H-1\right)}}, w^{N\overline{O}}={\displaystyle \frac{k}{\left(4k-H-1\right)}}.$$

(c)

The manufacturer’s and the retailer’s ex ante expected profits are as follows

$${\Pi }_m^{N\overline{O}}={\displaystyle \frac{k\left(1+H\right)}{8\left(4k-H-1\right)}}, {\Pi }_r^{N\overline{O}}={\displaystyle \frac{k\left(1+H\right)}{4\left(4k-H-1\right)}}.$$

4.1.2. Case $\mathrm{N}\mathrm{O}$

When the retailer does not share information and the manufacturer opens an online channel, the retailer first decides her retail margin ($\eta$) based on her private demand forecast signal $Y$; then, the manufacturer decides the green level of product ($g$), wholesale price ($w$), and online selling price ($p_m$) by inferring the demand information. The expected profits of the retailer and manufacturer can be expressed as follows:

$$\mathrm{E}\left[{\pi }_r|Y\right]=\eta {\mathsf{\lambda }}_r^O\mathrm{E}\left[a|Y\right],$$

(6)

$$\mathrm{E}\left[{\pi }_m|\eta \right]=w{\mathsf{\lambda }}_r^O\sum _{Y=h, l}\mathrm{Pr}\left(Y|\eta \right)E\left[a\right|Y]+\left(p_m-c\right){\mathsf{\lambda }}_m^O\sum _{Y=h, l}\mathrm{Pr}\left(Y|\eta \right)E\left[a\right|Y]-{\displaystyle \frac{1}{2}}\mathrm{k}{\mathrm{g}}^2.$$

(7)

Similarly, we summarize the unique LMSE equilibrium solutions for the case $NO$ in the following Lemma 2.

Lemma 2. 

Under case $NO$,

(a)

Only the separating equilibrium exists.

(b)

The most profitable separating equilibrium for the retailer is: the low-type retailer sets${\eta }_l^{NO}={\displaystyle \frac{2k\left(1-\theta \right)+(2k-a_l)c}{2(2k-\theta a_l)}}$, and the high-type retailer sets  ${\eta }_h^{NO}={\displaystyle \frac{2k\left(1-\theta \right)+\left(2k-a_h\right)c+\sqrt{\Delta }}{2\left(2k-\theta a_h\right)}}$.

Here, $\mathsf{\Delta }={\displaystyle \frac{2k\left(1-\mathsf{\theta }\right)\left(2k-a_h\right)\left(2k-a_l\right)\left(a_h-a_l\right)\left(2\mathsf{\theta }-c\right)c+4k^2{\left(1-\mathsf{\theta }\right)}^2\left(a_h-a_l\right)\left[\left(2k-a_h\mathsf{\theta }\right)+\mathsf{\theta }\left(2k-a_l\right)\right]}{\left(2k-a_l\right)\left(2k-\mathsf{\theta }{a}_l\right)}}$.

According to Lemma 2, we can see that under case $NO$, there only exists the separating equilibrium, where it is optimal for the low-type retailer to set the first-best retail margin, while the high-type retailer has to distort her retail margin upward to separate herself.

Based on the Lemma 2, we further obtain the corresponding equilibrium outcomes in Lemma 3.

Lemma 3. 

Under case $NO$,

(a)

When  $Y=l$, the optimal green level of the product, wholesale price, and online selling price are given as follows: 

$$g_l^{NO}={\displaystyle \frac{a_l\left[2k\left(1+\mathsf{\theta }-c\right)-\left(2\mathsf{\theta }-c\right)a_l\right]}{2\left(2k-a_l\right)\left(2k-a_l\mathsf{\theta }\right)}}, w_l^{NO}={\displaystyle \frac{k\left[2k\left(1+\mathsf{\theta }-c\right)-\left(2\mathsf{\theta }-c\right)a_l\right]}{2\left(2k-a_l\right)\left(2k-\mathsf{\theta }{a}_l\right)}},p_{ml}^{NO}={\displaystyle \frac{3\mathsf{\theta }c{a}_l^2+8\left(\mathsf{\theta }+c\right)k^2-2\left(2c+\mathsf{\theta }+3c\mathsf{\theta }+{\mathsf{\theta }}^2\right)k{a}_l}{4\left(2k-a_l\right)\left(2k-\mathsf{\theta }{a}_l\right)}};$$

when $Y=h$, the optimal green level of the product, wholesale price, and online selling price are given as follows: $g_h^{NO}={\displaystyle \frac{a_l\left(1-{\eta }^s\right)}{2k-a_l}}$, $w_h^{NO}={\displaystyle \frac{k\left(1-{\eta }^s\right)}{2k-a_l}}$, $p_{mh}^{NO}={\displaystyle \frac{\theta \left(2k-{\eta }^s{a}_l\right)+\left(2k-a_l\right)c}{2\left(2k-a_l\right)}}$.

(b)

The manufacturer’s and the retailer’s ex ante expected profits, respectively, are

$$\begin{array}{ll}{\Pi }_m^{NO}=& {\displaystyle \frac{1}{32}}\left[{\displaystyle \frac{\left(4-3\mathsf{\theta }\right)a_l}{\mathsf{\theta }\left(1-\mathsf{\theta }\right)}}{c}^2+{\displaystyle \frac{3{\left(2k-a_{l}c\right)}^2}{2k-\mathsf{\theta }{a}_l}}+{\displaystyle \frac{4k\left(2a_l-3k\right)}{2k-a_l}}\right]+\\ & {\displaystyle \frac{a_l\left[\left(2k-a_l\right)c^2+2k\mathsf{\theta }\left(1-\mathsf{\theta }\right)\left(1-2{\eta }^s\right)+\mathsf{\theta }\left(2k-\mathsf{\theta }{a}_l\right){\left({\eta }^s\right)}^2-2\mathsf{\theta }c\left(2k-a_l\right){\eta }^s\right]}{8\mathsf{\theta }\left(1-\mathsf{\theta }\right)\left(2k-a_l\right)}},\\ {\Pi }_r^{NO}=& {\displaystyle \frac{\left(1+H\right){\left[2k\left(1-\mathsf{\theta }\right)+\left(2k-a_l\right)c\right]}^2}{16\left(1-\mathsf{\theta }\right)\left(2k-a_l\right)\left(2k-\mathsf{\theta }{a}_l\right)}}.\end{array}$$

4.1.3. Case $\mathrm{I}\overline{\mathrm{O}}$

When the retailer shares information and the manufacturer does not open an online channel, both the retailer and the manufacturer know the actual demand signal $Y$ and make decisions according to the signal $Y$. The expected profits of the retailer and manufacturer are

$$\mathrm{E}\left[{\pi }_r|Y\right]=\eta {\lambda }_r^{\overline{O}}\mathrm{E}\left[a|Y\right],$$

(8)

$$\mathrm{E}\left[{\pi }_m|\eta \right]=w{\lambda }_r^{\overline{O}}\mathrm{E}\left[a|Y\right]-{\displaystyle \frac{1}{2}}\mathrm{k}{\mathrm{g}}^2.$$

(9)

By backward induction, we summarize the optimal decisions of the supply chain members in the following lemma.

Lemma 4. 

Under case $I\overline{O}$,

(a)

For  $Y=\{h, l\}$, the optimal retail margin, green level of product, and wholesale price, respectively, are

$${\eta }^{I\overline{O}}={\displaystyle \frac{1}{2}}, g^{I\overline{O}}={\displaystyle \frac{a_Y}{2\left(2k-a_Y\right)}}, w^{I\overline{O}}={\displaystyle \frac{k}{2\left(2k-a_Y\right)}}.$$

(b)

The ex ante expected profits of the manufacturer and the retailer are

$${\Pi }_m^{I\overline{O}}={\displaystyle \frac{k\left[k\left(1+H\right)-\beta \left(1-\beta \right){\left(H-1\right)}^2-H\right]}{8\left[2k+\left(H-1\right)\beta -H\right]\left[2k-\left(H-1\right)\beta -1\right]}}, {\Pi }_r^{I\overline{O}}={\displaystyle \frac{k\left[k\left(1+H\right)-\beta \left(1-\beta \right){\left(H-1\right)}^2-H\right]}{4\left[2k+\left(H-1\right)\beta -H\right]\left[2k-\left(H-1\right)\beta -1\right]}}.$$

4.1.4. Case $\mathrm{I}\mathrm{O}$

When the retailer shares information and the manufacturer opens an online channel, the retailer first determines her retail margin based on the signal $Y$; then, the manufacturer decides the green level, wholesale price, and online selling price based on the signal $Y$. Thus, the expected profits of the retailer and manufacturer can be expressed as follows

$$\mathrm{E}\left[{\pi }_r|Y\right]=\eta {\mathsf{\lambda }}_r^O\mathrm{E}\left[a|Y\right],$$

(10)

$$\mathrm{E}\left[{\pi }_m|Y\right]=w{\mathsf{\lambda }}_r^O\mathrm{E}\left[a|Y\right]+\left(p_m-c\right){\mathsf{\lambda }}_m^O\mathrm{E}\left[a|Y\right]-{\displaystyle \frac{1}{2}}\mathrm{k}{\mathrm{g}}^2.$$

(11)

Lemma 5. 

Under case $IO$,

(a)

For  $Y\in \{h,l\}$, the optimal retail margin, green level of product, wholesale price, and online selling price are as follows

$$\begin{array}{l}{\eta }^{IO}={\displaystyle \frac{2k\left(1-\theta \right)+\left(2k-a_Y\right)c}{2\left(2k-\theta a_Y\right)}}, g^{IO}={\displaystyle \frac{a_Y\left[2k\left(1+\theta -c\right)-\left(2\theta -c\right)a_Y\right]}{2\left(2k-a_Y\right)\left(2k-\theta a_Y\right)}},\\ w^{IO}={\displaystyle \frac{k\left[2k\left(1+\theta -c\right)-\left(2\theta -c\right)a_Y\right]}{2\left(2k-a_Y\right)\left(2k-\theta a_Y\right)}}, p_m^{IO}={\displaystyle \frac{3c\theta a_Y^2+8\left(c+\theta \right)k^2-2k\left(2c+\theta +3c\theta +{\theta }^2\right)a_Y}{4\left(2k-a_Y\right)\left(2k-\theta a_Y\right)}}.\end{array}$$

(b)

The manufacturer’s and the retailer’s ex ante expected profits are

$$\begin{array}{l}{\Pi }_m^{IO}={\displaystyle \frac{1}{32}}\left[{\displaystyle \frac{\left(4-3\mathsf{\theta }\right)\left(1+H\right)}{\mathsf{\theta }\left(1-\mathsf{\theta }\right)}}{c}^2+{\displaystyle \frac{3{\left(2k-c{a}_h\right)}^2}{2k-\mathsf{\theta }{a}_h}}+{\displaystyle \frac{3{\left(2k-c{a}_l\right)}^2}{2k-\mathsf{\theta }{a}_l}}-{\displaystyle \frac{4k\left(3k-2a_h\right)}{2k-a_h}}-{\displaystyle \frac{4k\left(3k-2a_l\right)}{2k-a_l}}\right],\\ {\Pi }_r^{IO}={\displaystyle \frac{{\left[2k\left(1-\mathsf{\theta }\right)+\left(2k-a_h\right)c\right]}^2{a}_h}{16\left(1-\mathsf{\theta }\right)\left(2k-a_h\right)\left(2k-\mathsf{\theta }{a}_h\right)}}+{\displaystyle \frac{{\left[2k\left(1-\mathsf{\theta }\right)+\left(2k-a_l\right)c\right]}^2{a}_l}{16\left(1-\mathsf{\theta }\right)\left(2k-a_l\right)\left(2k-\mathsf{\theta }{a}_l\right)}}.\end{array}$$

4.2. Comparison and Analysis

In this subsection, we focus on the impacts of the channel strategy and information sharing on the optimal green level of product and the profits of the supply chain members, respectively. By calculation, we can derive the following propositions.

Proposition 1. 

(i)

$\mathrm{E}\left[g^{N\overline{O}}\right]<\mathrm{E}\left[g^{I\overline{O}}\right]$ and $\mathrm{E}\left[g^{NO}\right]<\mathrm{E}\left[g^{IO}\right]$;

(ii)

${\Pi }_m^{I\overline{O}}>{\Pi }_m^{N\overline{O}}$, ${\Pi }_r^{I\overline{O}}>{\Pi }_r^{N\overline{O}}$, ${\Pi }_m^{IO}>{\Pi }_m^{NO}$ and ${\Pi }_r^{IO}>{\Pi }_r^{NO}$.

It follows from Proposition 1 that information sharing always improves the green level of product and benefits both the manufacturer and the retailer whether an online channel is opened or not. One might intuitively think that the retailer would be reluctant to share her information with the manufacturer. However, as shown in Proposition 1(ii), the retailer prefers to share information. Here is why: without an online channel, if the retailer does not share information, she selects a pooling equilibrium. This means that the manufacturer, relying only on prior information, cautiously makes the green level of product decision. To avoid potential losses when the actual market size is lower without information sharing compared to when information is shared, the manufacturer sets a low green level. Consequently, the retailer benefits from information sharing due to the positive effect on the green level. For the manufacturer, when the retailer shares demand information, he can make optimal decisions regarding the green level of product and pricing based on the shared information. Additionally, the retail margin remains $\eta ={\displaystyle \frac{1}{2}}$ regardless of whether the retailer shares information, which does not impact the manufacturer. Therefore, without an online channel, the manufacturer also benefits from information sharing.

With an online channel, a separating equilibrium exists when the retailer does not share demand information. In this scenario, the retailer distorts her retail margin upwards, resulting in lower expected profits compared to a situation where she shares information and can set the optimal retail margin. Consequently, both the green level of product and the manufacturer’s profit are also lower without information sharing. In summary, information sharing always enhances the green level of product and benefits both the manufacturer and the retailer, regardless of whether the manufacturer establishes an online channel.

Next, we aim to analyze the effects of the manufacturer’s channel strategy on the green level of product and the ex ante expected profits of the supply chain members. As the equilibrium solutions under case $NO$ are complex, we apply numerical analysis to study the effects of the manufacturer’s channel strategy.

Proposition 2. 

When the retailer shares information,

(i)

${\displaystyle \frac{\partial g^{IO}}{\partial c}}<0$and$g^{I\overline{O}}<g^{IO}$;

(ii)

Opening an online channel is always beneficial to the manufacturer but is beneficial to the retailer when ${\displaystyle \frac{{\varphi }_1+2\sqrt{{\Delta }_1}}{{\varphi }_2}}<c<\theta$. Here, ${\varphi }_1={\displaystyle \frac{4k\left[\theta a_l{a}_h-k\left(1+H\right)\right]}{\left(2k-\theta a_l\right)\left(2k-\theta a_h\right)}}$, ${\varphi }_2={\displaystyle \frac{1+H}{1-\theta }}-{\displaystyle \frac{a_l^2}{2k-\theta a_l}}-{\displaystyle \frac{a_h^2}{2k-\theta a_h}},$ and ${\Delta }_1={\displaystyle \frac{k\left(1+H\right)\left[k\left(1+H\right)-\theta a_h{a}_l\right]}{\left(1-\theta \right)\left(2k-\theta a_l\right)\left(2k-\theta a_h\right)}}$.

Proposition 2(i) implies that the green level of product decreases with the increase in the manufacturer’s direct selling cost $c$ under Case $IO$. This implies that as the direct selling cost rises, the benefits of the online channel diminish, prompting the manufacturer to reduce the green level of product to avoid intense competition and to cut costs. Additionally, the green level of product can be improved when the manufacturer opens an online selling channel. With an online channel, demand in both channels can be expanded by increasing the green level of product, whereas without an online channel, only demand in the retail channel can be expanded. Thus, the manufacturer prefers to improve the green level of product by opening an online channel to expand demand.

Proposition 2(ii) shows that when the retailer shares information, the manufacturer always benefits from opening an online channel, while the retailer only benefits from opening an online channel under certain conditions. For the manufacturer, opening an online channel can expend total demand by improving the green level of product ($g^{I\overline{O}}<g^{IO}$) and reducing the retail margin (${\eta }^{I\overline{O}}>{\eta }^{IO}$), which leads to more profit for the manufacturer. For the retailer, opening an online channel not only creates competition between the manufacturer and the retailer, but also increases the fraction of consumers for the retail channel (${\mathsf{\lambda }}_r^O>{\mathsf{\lambda }}_r^{\overline{O}}$), creating a balance between negative competition effects and positive retail demand increases. When the manufacturer’s direct selling cost is high, the retailer benefits from an online channel as the positive retail demand increase effect outweighs the negative competition effect.

Real-life examples such as Amazon, Zara, and Nike illustrate these findings. For example, Nike’s direct-to-consumer (DTC) strategy highlights the dual benefits of information sharing and online channel integration. Nike shares demand data with retail partners to ensure optimal stock levels and product placement, enhancing overall green initiatives. This approach reflects the conditions in Proposition 2(ii), where Nike benefits from increased direct sales and enhanced consumer insights, while retailers benefit from improved product offerings and reduced competition due to data-driven strategies.

Furthermore, under the case without information sharing, we perform a numerical example to study the effect of the manufacturer’s channel strategy due to the complex equilibrium solutions. We show some results in Figure 2 and Figure 3, where the default values of the parameters are $H=1.5$, $\beta =0.6$, $k=1.5$, $\mathsf{\theta }=0.5$, and $c=0.1$.

Figure 2 shows the impact of opening an online channel on the green level of product. In Figure 2, we can find that the effect of opening an online channel on the green level of product depends on the direct selling cost ($c$) and the forecast signal precision ($\beta$). Specifically, when both the direct selling cost and the forecast signal precision are relatively low, opening an online channel can improve the green level of product. Otherwise, the green level of product cannot be improved by establishing an online channel. This result is different from the case where the retailer shares information (as shown in Proposition 2(i)), which indicates that opening an online channel can improve the green level of product.

Figure 3 shows the effect of opening an online channel on the ex ante profits of the manufacturer and the retailer when the retailer does not share information. From Figure 3, we can see that without information sharing, the effects of opening an online channel on the manufacturer and the retailer depend on the direct selling cost ($c$) and the forecast signal precision ($\beta$). Specifically, for the manufacturer, when $\beta$ and $c$ are relatively high, the manufacturer is worse off with the opening of an online channel. Otherwise, the manufacturer is always better off with the opening of an online channel. For the retailer, when $\beta$ is relatively low and $c$ is relatively high, the retailer is better off with the opening of an online channel. Otherwise, the retailer is always worse off with the opening of an online channel. This observation is different from that with information sharing (as shown in Proposition 2(ii)).

There is a significant difference in the impacts of opening an online channel under the cases with and without information sharing from the results in Figure 3 and Proposition 2(ii). The manufacturer always prefers to open an online channel under the case of information sharing, whereas he may not open an online channel under the case without information sharing. In particular, the manufacturer prefers not to open an online channel when $\beta$ is relatively large (as shown in Figure 3a). This result means that the retailer may deter the manufacturer from opening an online channel via withholding the private demand information. The reason is as follows. Opening an online channel can lead to a high signaling cost of the retailer, which means that the high-type retailer must distort the retail margin upwards to separate herself from the low-type. Such distortion is detrimental not only to the retailer but also to the manufacturer. Thus, the manufacturer may not open an online channel when the forecast signal precision and the direct selling cost are relatively high, and the retailer is also harmed by the opening of an online channel.

4.3. Numerical Study

As mentioned above, there are four possible strategy combinations in this paper; these are $\{N,\overline{O}\}$, $\{N, O\}$, $\left\{I,\overline{O}\right\},$ and $\{I, O\}$, respectively. In this subsection, we examine the manufacturer’s and the retailer’s strategy preferences through the numerical approach and further derive the equilibrium strategy.

First of all, we characterize the strategy preferences of the manufacturer and the retailer in terms of the four strategy combinations, and the corresponding results are shown in Figure 4 and Figure 5. In Figure 4 and Figure 5, the regions of invalid zones represent the case where a dual-channel selling case is not satisfied. Figure 4 shows the manufacturer’s strategy preference in terms of the four strategy combinations. In Figure 4, we can see that the manufacturer prefers the strategy combination $\{I, O\}$, which represents that the retailer shares her private demand information and that the manufacturer establishes an online channel (i.e., case $IO$). Under this strategy combination, both the improvement of the green level of product and the decreases in the retail margin increase the demand for the green products, which, in turn, leads to more profit for the manufacturer. In addition, when the forecast signal precision ($\beta$) is relatively low, the manufacturer always chooses to open an online channel regardless of whether the retailer shares her private demand information or not, as shown in region ① of Figure 4. When $\beta$ is relatively high, in region ③ of Figure 4, the manufacturer chooses to open an online channel if the retailer shares the demand information, while he chooses not to open an online channel if the retailer does not share the demand information. In contrast, in region ② of Figure 4, when $\beta$ and $c$ are relatively high, or $\beta$ is relatively high and $c$ is low, the manufacturer chooses to establish an online channel if the retailer chooses not to share the demand information.

Figure 5 describes the strategy preference of the retailer. From Figure 5, we can see that the retailer’s strategy preference depends on the direct selling cost ($c$). Specifically, when the direct selling cost is relatively high, the retailer prefers the strategy combination $\{I, O\}$. Otherwise, the retailer prefers the strategy combination $\{I,\overline{O}\}$. This verifies the results in Proposition 2(ii). Furthermore, according to Figure 5, we can observe that when $c$ is relatively low, i.e., regions ② and ③ of Figure 5, the retailer does not want the manufacturer to establish an online channel, regardless of whether she shares the demand information or not. However, when $c$ is relatively high, i.e., region ① of Figure 5, the retailer wants the manufacturer to open an online channel when she shares the demand information; otherwise, the retailer does not want the manufacturer to open an online channel when she does not share the demand information.

Based on the above analysis, we further derive the equilibrium strategy, i.e., $\{N,\overline{O}\}$ or $\{I, O\}$, as shown in region ① or ② of Figure 6. That is, the retailer can stimulate the manufacturer to establish an online channel by sharing her private demand information, or the retailer can prevent the manufacturer from opening an online channel via withholding her demand information. In other words, when the retailer alters her information sharing strategy, the manufacturer will change his original channel strategy accordingly. Specifically, when the direct selling cost $c$ is relatively high, the retailer prefers to share the demand information, and the manufacturer prefers to establish an online channel due to both the positive retail margin reducing effect and the positive channel addition effect, as shown in ① of Figure 6. When $c$ is relatively low and $\beta$ is relatively high, the retailer chooses not to share the demand information and the manufacturer also chooses not to open an online channel, as shown in region ② of Figure 6. Under this case, when the retailer does not share information, opening an online channel can distort the retail margin upwards, which harms the manufacturer and the retailer. Therefore, without information sharing, the manufacturer may be deterred from opening an online channel.

5. Extensions

In the previous section, for ease of exposition, we made restrictions on the game sequence of the basic model. In this section, we examine an alternative scenario where the manufacturer decides whether to open an online channel before the retailer shares the demand information [60], to further investigate the retailer’s incentive to share the demand information in anticipation of the manufacturer’s channel strategy.

First, we study the retailer’s optimal information sharing decision given that the manufacturer has already made his decision to establish an online channel. If the manufacturer decides to open an online channel, the retailer’s ex ante expected profits under no information sharing and information sharing are ${\tilde{\Pi }}_r^{ON}={\displaystyle \frac{\left(1+H\right){\left[2k\left(1-\mathsf{\theta }\right)+\left(2k-a_l\right)c\right]}^2}{16\left(1-\mathsf{\theta }\right)\left(2k-a_l\right)\left(2k-\mathsf{\theta }{a}_l\right)}}$ and ${\tilde{\Pi }}_r^{OI}={\displaystyle \frac{{\left[2k\left(1-\mathsf{\theta }\right)+\left(2k-a_h\right)c\right]}^2{a}_h}{16\left(1-\mathsf{\theta }\right)\left(2k-a_h\right)\left(2k-\mathsf{\theta }{a}_h\right)}}+{\displaystyle \frac{{\left[2k\left(1-\mathsf{\theta }\right)+\left(2k-a_l\right)c\right]}^2{a}_l}{16\left(1-\mathsf{\theta }\right)\left(2k-a_l\right)\left(2k-\mathsf{\theta }{a}_l\right)}}$, respectively. The fact that ${\tilde{\Pi }}_r^{OI}>{\tilde{\Pi }}_r^{ON}$ implies that the retailer will share her demand information. This is different from the results in Li et al. [33], which show that the retailer conditionally shares her demand information given the supplier’s online channel.

If the manufacturer decides not to open an online channel, the retailer’s ex ante expected profits under no information sharing and information sharing are ${\tilde{\Pi }}_r^{\overline{O}N}={\displaystyle \frac{k\left(1+H\right)}{4\left(4k-H-1\right)}}$ and ${\tilde{\Pi }}_r^{\overline{O}I}={\displaystyle \frac{k\left[k\left(1+H\right)-\beta \left(1-\beta \right){\left(H-1\right)}^2-H\right]}{4\left[2k+\left(H-1\right)\beta -H\right]\left[2k-\left(H-1\right)\beta -1\right]}}$, respectively. Similarly, we find that ${\tilde{\Pi }}_r^{\overline{O}N}<{\tilde{\Pi }}_r^{\overline{O}I}$, which shows that the retailer would share the demand information. Taken together, we find that the retailer always shares her demand information, regardless of the timing of the manufacturer’s channel strategy.

Then, we consider another possible sequence for the manufacturer’s channel strategy; that is, the manufacturer decides whether to open an online channel after the retailer sets the information sharing policy, but before there is a resolution about the demand information. In this setting, the manufacturer’s channel strategy decision depends on the demand information inferred from the retailer’s retail margin rather than the shared demand information. This sequence is similar to the case where the retailer does not share the demand information. Thus, in equilibrium, the manufacturer may not establish an online channel when the retailer does not share information.

6. Conclusions

This paper investigates the strategic interplay between a retailer’s information sharing strategy and a manufacturer’s channel strategy within a retailer-led green supply chain. The analysis encompasses scenarios where the retailer holds private demand information and the manufacturer considers establishing an online channel for direct consumer sales. We evaluate four strategic combinations to discern their effects on the green product level and the economic outcomes for both parties in the supply chain. Furthermore, we also examine the impact of the sequence of the manufacturer’s channel strategy on the retailer’s incentive for sharing information to verify the robustness of our main findings. Our analysis highlights the following main findings:

First, we find that information sharing can improve the green level of products and benefit the manufacturer and the retailer regardless of whether the manufacturer opens an online channel or not. Additionally, opening an online channel can always improve the green level with information sharing but can improve the green level under certain conditions without information sharing. Second, the manufacturer benefits from opening an online channel under the case with information sharing, while this may be harmful under the case with no information sharing. As for the retailer, she can conditionally benefit from opening an online channel under both the case with information sharing and that without. Third, the manufacturer prefers the scenario of opening an online channel with information sharing, and the retailer prefers either the scenario of opening an online channel with information sharing or the scenario of not opening an online channel with information sharing. Finally, our numerical analysis shows that the retailer prefers to share the demand information to inspire the manufacturer to establish an online channel if the direct selling cost is relatively high and to withhold the demand information to prevent the manufacturer from opening an online channel if the direct selling cost is relatively high and the forecast signal precision is relatively high.

Based on the main results, several management insights for supply chain members regarding the incentive for adopting BT can be offered. First, leverage demand information strategically to influence the manufacturer’s decision regarding opening an online channel. The manufacturer benefits from opening an online channel with information sharing but may suffer without it. The retailer can benefit under both scenarios, depending on certain conditions. Second, the retailer may withhold information if the direct selling costs are low and the forecast accuracy is high to prevent the manufacturer from opening an online channel. Finally, collaborate with the retailer to enhance the green product level and achieve mutual benefits. This collaboration is essential for improving the overall sustainability of the supply chain, including reducing environmental impacts and improving operational efficiencies.

Although this work enriches the literature in several aspects, the present paper has some limitations, and possible extensions of our results could be addressed in future research. First, this paper assumes that only the retailer possesses demand forecast information. In practice, with the development of information technology, many manufacturers can also acquire imperfect demand information. Therefore, future research could explore bilateral information sharing between the manufacturer and the retailer under different manufacturer channel strategies. Second, this paper considers only a markup pricing contract. Future studies could investigate the interaction effects of retailer information sharing and manufacturer channel strategies under various contracts, such as two-part tariffs and revenue sharing. Finally, our model focuses on a single retailer–manufacturer relationship in a retailer-led supply chain. It would be valuable to extend the analysis to different market structures, such as a manufacturer-led supply chain, multiple competing retailers, or manufacturers, and to assess how competitive environments influence the strategic interactions and outcomes. For instance, considering oligopolistic competition or a supply chain network with multiple tiers could provide deeper insights into the robustness and applicability of our findings in promoting sustainability.