Balancing E-Commerce Platform and Manufacturer Goals in Sustainable Supply Chains: The Impact of Eco-Friendly Private Labels

Abstract

This paper analyzes how consumer preferences for eco-friendly private labels affect platform selling formats and manufacturer channel strategies. We construct a game-theoretic model that encompasses both an e-commerce platform and a manufacturer. In this model, the platform chooses its selling format—either wholesale or agency—while the manufacturer determines whether to launch an online direct channel. Our analysis takes into account consumer preferences for both product brands and sales channels. We compare and analyze the equilibrium outcomes across four different scenarios, leading to the following conclusions: (1) When the platform utilizes the wholesale selling format, the manufacturer is consistently motivated to launch the direct channel if the costs associated with it are low. Conversely, when the platform employs the agency selling format, the manufacturer’s decision to establish an online direct marketing channel depends on the relative strength of consumers’ eco-friendly and channel preferences. (2) The platform’s choice of selling format is affected by the manufacturer’s channel strategy, leading the platform to modify its selling format based on varying manufacturer channel approaches. (3) The choices of the platform’s selling format and the manufacturer’s channel strategy are not always in conflict. When consumer preferences for the manufacturer’s direct channel are strong and the cost of introducing it is moderate, the platform’s decision to adopt the agency selling format can create a win-win outcome for both the platform and the manufacturer. And the establishment of online channels by manufacturers further enhances the sustainable growth of eco-friendly products.

1. Introduction

With the rapid expansion of the internet and transportation services, an increasing number of consumers are electing to acquire items through online channels. E-commerce platforms, including Amazon, JD.com, Taobao, and eBay, have witnessed substantial growth over the past decade. As reported by Statista, the e-commerce market revenues in the United States and China are estimated at USD 1.22 trillion and USD 1.47 trillion, respectively [1]. E-commerce platforms have increasingly sought to capture a significant share of the online market by introducing their own private label (PL) products, in direct competition with offerings from national brand (NB) manufacturers. From 2018 to 2023, over 100 product categories under JD’s PL (Jing Zao) experienced an average annual sales growth of over 300% and Amazon’s PL sales exceeded USD 1 billion in the first quarter of 2024 [2,3]. Additionally, a report from Simon-Kucher indicates that 78% of consumers prioritize sustainability when purchasing products, and 56% show greater loyalty to environmentally certified brands [4]. In alignment with the trend of sustainable development and in response to the global market’s demand for environmentally friendly products, e-commerce platforms have successively launched eco-friendly private label (EFPL) products dedicated to environmental protection and sustainable growth. For instance, Amazon has launched its environmentally sustainable private label, Amazon Aware, which offers a wide range of eco-friendly products, including clothing, home goods, and daily necessities. Similarly, in response to the growing demand for green consumption and sustainable development, JD’s Jing Zao is committed to promoting environmentally friendly, low-carbon production methods by utilizing recyclable materials in its products.

Table 1. E-commerce platforms’ EFPL products.

EFPL	Category	Sustainability and Environmental Contributions
Amazon Aware	Apparel	Made from recycled materials and certified with Organic Content Standard 100, Global Recycled Standard.
	Beauty	Uses natural ingredients and zero plastic packaging.
	Home	Certified by OEKO-TEX for eco-friendly production.
	Packaging	Certified with “Compact by Design” to reduce packaging waste and improve transportation efficiency.
JD Jing Zao	Batteries	Mercury-free, eco-friendly batteries.
	Yoga Mats	Made from eco-safe TPE, no glue used.
	Bedding	Sheets certified by GOTS with eco-friendly dyeing.
	Packaging	Uses biodegradable materials and reduces carbon emissions through new energy vehicle transportation.

The information is sourced from Amazon and JD.com.

summarizes the common EFPL products found on Amazon and JD, along with the sources of materials for these products and their sustainability and environmental contributions.

E-commerce platforms introduce best-selling products that are similar to those of NBs by utilizing their comprehension of consumer preferences. This has undoubtedly led to intense competition between EFPLs and NBs on the e-commerce platforms. Nevertheless, a discrepancy in product quality and brand reputation persists between eco-friendly private labels and national brands, prompting consumers to perceive platform EFPLs as imperfect substitutes for NBs.

In addition to selling their products through e-commerce platforms, an increasing number of manufacturers are opting to establish online direct channels, such as their own websites, which enable them to reach a broader consumer base while reducing their reliance on e-retailers and avoiding the payment of high commissions to e-commerce platforms. For instance, brands like Apple, Xiaomi, Huawei, Midea, and Nike have created their own online sales channels. Through online direct sales channels, manufacturers can interact directly with consumers and sell their products. This allows manufacturers to better control the sales process, provide enhanced customer experiences, and potentially increase their profits. At the same time, this will inevitably lead to channel competition between manufacturers and platforms. Channel competition is a common occurrence in e-commerce supply chains. Manufacturers must balance the benefits and drawbacks of collaborating between online direct sales and e-commerce platforms, aligning their decisions with strategic objectives and market demand. However, consumers often exhibit different preferences for a manufacturer’s online direct channel compared to that of the platform. Consumers can easily search for a large number of products on e-commerce platforms and compare prices and other product features. In addition, e-commerce platforms usually provide consumers with better services in terms of transactions and logistics compared to manufacturers’ direct channels. Therefore, manufacturers’ direct channels are generally not a complete substitute for e-commerce platforms for consumers.

E-commerce platforms commonly utilize two selling formats: agency selling and wholesale selling [5,6]. In the agency selling format, the platform serves as a marketplace, earning a commission from the manufacturer based on a percentage of the sales price, while the manufacturer determines the retail price (e.g., Taobao and Pinduoduo.com). In contrast, the wholesale selling format involves the platform purchasing goods from manufacturers and reselling them to consumers, with the platform setting the retail price (e.g., Amazon.com and JD.com). The key difference between these formats is who controls the retail price [7].

The issues related to the platform selling format and manufacturer encroachment have been widely discussed in the academic literature. However, existing research typically examines manufacturer encroachment or the selling format decisions of e-commerce platforms in isolation [8,9]. Some studies have investigated the strategic interplay between the platform selling format and manufacturer encroachment decisions [10,11,12]. However, they typically ignore the competition between EFPLs and NBs and presume that customer preferences for the manufacturer’s online website and its e-commerce site are horizontally differentiated. In practice, consumers perceive manufacturers’ direct channels as imperfect substitutes for e-commerce platforms [13]. Furthermore, consumers view the PLs as imperfect substitutes for the NBs [14]. Consumers’ purchasing decisions will be influenced by dual factors: their preferences for product brands and sales channels. Building on the analysis above, we propose the following research questions. What should the manufacturer’s direct channel introduction decision be when consumers have different preferences for product brands and distribution channels? How does the e-commerce platform determine its selling format when facing the manufacturer’s channel strategies?

To tackle these questions, our research constructs a theoretical model based on the realistic foundation of consumers’ vertically differentiated preferences for product brands and sales channels. It explores the relationship between a manufacturer’s strategy for launching a direct channel and the platform’s decision on its selling format, considering competition among both brands and sales channels, thereby effectively supplementing previous studies. Specifically, we model four distinct scenarios that consider the e-commerce platform’s selection of a selling format—either wholesale or agency—and the manufacturer’s approach to launching an online direct channel. We derive equilibrium outcomes for these four scenarios, determining the optimal selling format for the e-commerce platform and the market conditions that influence the manufacturer’s decision to introduce a direct channel. Our key research findings are as follows:

(1)

When the platform adopts a wholesale selling format, the manufacturer will consistently be motivated to pursue the direct channel if the associated costs are relatively low. Launching a direct channel can effectively reduce efficiency losses caused by double marginalization, thereby enhancing the manufacturer’s profitability. However, when the platform employs an agency selling format, the manufacturer’s decision to launch an online direct channel is contingent upon consumers’ relative preferences for EFPLs and direct channels. If consumers have a strong preference for EFPLs, the manufacturer will not launch a direct channel, as this would serve only to exacerbate market competition and diminish profit margins. Conversely, when consumers have a stronger preference for direct channels, the manufacturer will decide to initiate the direct channel if the associated costs are low.

(2)

The platform’s choice of selling format is affected by the manufacturer’s channel strategy. The platform’s decision regarding its selling format may vary based on the manufacturer’s different channel strategies and consumer preferences for brands and channels.

(3)

The platform’s selling format and the manufacturer’s channel strategies are not always in opposition. When consumers demonstrate a greater preference for the manufacturer’s direct channel over the platform’s EFPL, and the costs for the manufacturer to introduce the direct channel are moderate, opting for an agency selling format can create a mutually beneficial outcome for both the platform and the manufacturer.

The paper is organized as follows: Section 2 reviews the literature. Section 3 establishes the model and derives the inverse demand functions for different scenarios. Section 4 presents the equilibrium solutions for each scenario. Section 5 explores the equilibrium strategies. Section 6 concludes the paper and suggests future research directions.

2. Literature Review

This paper contributes to three areas of the literature. The first area focuses on the private labels. The second focuses on studies comparing selling formats on platforms. The third area focuses on manufacturer encroachment. In the subsequent sections, we will compare our study with the existing literature and outline our contributions.

2.1. Private Label

The introduction of PLs has altered the competitive and collaborative relationships between manufacturers and retailers. Research suggests that PLs can effectively segment the market and enhance market efficiency [15]. The introduction of a PL by the retailer benefits both consumers and the retailer, yet it presents a challenge to the established dominance of incumbent NB manufacturers [16]. In response to the threat of the PL, the manufacturer may choose to reposition products within existing product lines. The manufacturer can opt to become the supplier for the PL. Hara and Matsubayashi examine the motivations for retailers and manufacturers to jointly create premium PLs that align with consumer preferences [17]. The manufacturer can also introduce new products to compete with the PL. Li et al. argue that launching differentiated products is the best coping strategy for the manufacturer when a PL threatens the market share of an NB [18]. They consider how factors such as new product expansion effects, quality differences, and commission rates influence the strategic decisions of the manufacturer. Li et al. explore strategies for launching a PL on e-retail platforms and analyze the procurement approaches associated with a PL [19]. Additionally, some studies suggest that the launch of a PL by a retailer can generate a win-win situation for all members of the supply chain. Ru et al. find that a PL can be advantageous for manufacturers. A PL can alleviate the double marginalization issue within the NB supply chain [20]. However, this occurs in an unconventional manner: the efficiency improvement may result from a decrease in retail price rather than a reduction in wholesale price. Maesen examines the impact of launching a specialist organic private label (OPL) on retailer performance. The research highlights that retailers may be better off maintaining national brand assortments, as reducing them can lead to a “larger share of a smaller pie” [21].

However, these studies are usually conducted in the context of traditional retailers and do not consider the launch of an online direct channel by the manufacturer or the various selling formats that the e-commerce platform may adopt. In practice, manufacturers often establish online direct channels to compete with e-commerce platforms, which allows them to avoid high commission fees and reduce dependence on these platforms. Zhang et al. examine the strategic dynamics between the introduction of a PL and a manufacturer’s channel decision [22]. They find that introducing low-quality PLs can deter manufacturers’ entry, while introducing high-quality private brands does not have the same effect. Wang et al. examine the effect of launching an EFPL on an e-commerce platform [23]. By using game theory and big data from JD.com, Niu et al. reveal complex interactions between platform advantages, logistics, and private label competition in shaping manufacturers’ channel strategies [24]. However, they do not consider the selling format of the platform.

Additionally, e-commerce platforms can choose to sell NBs through either wholesale or agency models, leading to significant differences in pricing strategies between EFPLs and NBs under different selling formats. Shopova examines the scenario of a large platform that adopts an agency selling format, such as Amazon, introducing its own PL to compete directly with sellers [25]. Liu et al. investigate the selling format options available to original equipment manufacturers (OEMs) in the context of PL outsourcing strategies [14]. Hemmati et al. investigate the influence of advertising and product quality on the selling format selection of an e-retailer offering a PL [26]. The results indicate that when faced with intense competition between a PL and an NB, the platform is inclined to utilize an agency model. Agency sales may result in higher prices for both the NB and PL.

In contrast to the previously mentioned studies, we simultaneously examine the strategic interplay between manufacturers introducing online direct channels and platforms’ selling format decisions. We also consider the presence of an EFPL, an aspect that has not been addressed in the existing research.

2.2. Manufacturer Encroachment

Our study is also intimately linked to the body of literature on manufacturer encroachment. Chiang et al. indicate that manufacturers introducing online channels with high consumer acceptance can alleviate double marginalization [27]. Arya et al. indicate that both manufacturers and retailers may benefit from the increase in manufacturer sales channels [28]. Cattani et al. provide strategies for manufacturers to mitigate competition and discuss how both direct and wholesale channels can benefit simultaneously [29]. Cai identifies that a win-win outcome is achieved when wholesale channels have sufficient advantages over direct channels in terms of demand or operational costs [30]. Li et al. consider the scenario where retailers possess private information about demand [31]. Yoo and Lee analyze a supply chain in which a manufacturer distributes products via either a single channel or multiple wholesale channels, discussing the impact of introducing online channels on businesses [32]. Hu et al. consider a wholesale channel in which resellers may transfer some products to their gray market channels [33]. Guan et al. examine the relationship between supplier encroachment and inventory in two competing supply chains [34]. Zhang et al. investigate the effects of advertising by suppliers or retailers on supplier encroachment decisions [35].

However, these studies consider manufacturer encroachment in traditional supply chains. A few studies have explored dual-channel competition within e-commerce platform supply chains. Shen et al. investigate manufacturers’ channel decisions regarding selling products through agency channels on online retail platforms and/or physical wholesale channels operated by other retailers [36]. Dong et al. investigate how cannibalization effects influence the pricing and quality strategies of competing manufacturers [37]. Ye et al. explore the decisions of hotels regarding partnerships with online travel agencies for room sales, and they analyze which business model—wholesale or agency—best fits hotels utilizing a dual-channel strategy [38]. Shi et al. explore manufacturer encroachment in e-commerce, comparing integrated and decentralized structures. They find decentralization expands encroachment potential but increases competition with retailers, potentially reducing profits. An integrated structure may lead to better outcomes when direct selling costs are low [39].

Limited research has investigated how private labels influence manufacturers’ encroachment decisions. Zennyo proposes a platform encroachment model in which the platform serves as a marketplace between consumers and sellers while also selling its first-party products [40,41]. However, these works do not take into account the effect of different platform selling formats on supplier pricing and online direct channel decisions. Unlike his study, we explore the interaction between manufacturers’ encroachment decisions and the selling format choices made by the platform in the context of private labels.

2.3. Platform Selling Formats

Many studies have examined whether online retail platforms adopt a wholesale or agency platform selling format. Hagiu and Wright investigate the optimal business model decision for platforms in the context of manufacturers and platforms providing services, concluding that the decision depends on who possesses more information: the platform or the manufacturer [42]. Abhishek et al. examine platform selling format decisions in a duopolistic competition context, suggesting that the platform selling format is influenced by the level of competition among platforms and the spillover effects across different channels [5]. Kwark et al. investigate the influence of user reviews on the selection of selling formats by platforms [43]. Zhang and Zhang examined platform selling format decisions and manufacturers’ direct channel establishment decisions under different information-sharing strategies [12]. Liu et al. investigate the impact of data-driven marketing and market size on online retail platform selling format decisions [44]. Yan et al. evaluate manufacturers’ multi-channel selling issues under factors such as online spillover effects, platform fees, and manufacturer retail efficiency [45]. Zhang et al. focus on quality decision-making by manufacturers entering the platform under different contracts and selling formats [46]. Other studies have considered the impact of return policies on business models [47].

Tian et al. further consider that platforms can adopt a mixed platform selling format, analyzing the conditions under which platforms employ different formats for different manufacturers and demonstrating the impact of varying manufacturer competition intensity and order fulfillment costs on platform selling format selection [48]. Wei and Dong explore the impact of order fulfillment costs on platform selling format decisions when platforms sell vertically differentiated products [49]. Ha et al. examine the equilibrium conditions for platforms adopting wholesale, agency, or hybrid platform selling formats when a monopolistic retailer enhances its sales through service efforts [50]. Xi and Zhang investigate platform selling format decision-making under uniform and differentiated pricing models, also considering the simultaneous use of agency and wholesale mixed formats by platforms [51]. Etro examines mixed platform commission decisions and corresponding welfare effects under conditions of free entry for manufacturers [52]. Gautier et al. analyze the endogenous decision-making of platform selling formats when numerous homogeneous merchants compete in a market with consumer and merchant search [53].

Similar to our research, Hagiu et al. study whether platforms should adopt a mixed platform selling format that simultaneously sells manufacturer brands and private labels, examining the impact of platform self-preference and private label imitation of manufacturer products on welfare [54]. But they did not consider manufacturer encroachment. Zhang and Zhang examine manufacturers’ direct channel introduction strategies and platform business model decisions under different information-sharing strategies employed by online retail platforms [12]. Matsui considers the choice of distribution strategies among two competing suppliers and studied the business model selection of e-commerce platforms for competing suppliers in a typical dual-channel supply chain [11]. But they did not consider eco-friendly private labels.

Unlike previous studies, we not only study the interaction between the manufacturer’s encroachment and the platform’s selling format decisions, but also consider the competition between EFPLs and NBs.

Table 2. Comparison between relevant literature and our study’s focus.

Related Study	Selling Format	Manufacturer Encroachment	Private Label	Eco-Friendly Product	Consumer Preferences
Zhang and Zhang [12]	✓	✓
Li et al. [13]		✓	✓		✓
Zennyo [41]		✓	✓
Liu et al. [14]	✓		✓		✓
Etro [52]	✓		✓
Matsui [11]	✓	✓
Wang et al. [23]			✓	✓	✓
This study	✓	✓	✓	✓	✓

✓ indicates the research topic involved in the literature.

provides a summary of the relevant literature, highlighting both the similarities and distinctions in comparison to our research.

3. The Model

3.1. When the Manufacturer Does Not Introduce an Online Direct Channel

Consider an e-commerce supply chain with a manufacturer and a platform. The platform not only sells its own eco-friendly private label (EFPL) product 1, but also offers a national brand (NB) product 2, which is produced by the manufacturer. The platform determines the retail price of its EFPL product, $p_1^Z$. As for the NB, while the manufacturer handles the production, the sales must go through the platform to reach the end customer. Platforms have two selling formats for selling NBs, namely the wholesale selling format and agency selling format. In the wholesale selling format, the platform buys the NB from the manufacturer at a wholesale price $w$ and then determines the retail price $p_2^R$. In the agency selling format, the manufacturer sets the retail price $p_2^A$ for the NB, selling products to consumers via the platform, and subsequently pays a commission $\alpha$, which is a percentage of the sales price, to the platform. $Z=R,A$ represents the platform’s wholesale and agency selling format, respectively.

Since most consumers perceive the EFPL to be less valuable than the NB [6,13,15], we regard the EFPL and NB as products that are vertically differentiated. Based on the fact that consumers are influenced by both price and brand preference, and referring to the utility model of [13,55], it is assumed that consumers’ valuation of the NB is $v$. $v$ is uniformly distributed on $\left[0,1\right]$. When a consumer makes a NB purchase, the resulting utility is $v-p_2$. $\delta$ represents the extent of consumer preference for the EFPL, where $\delta \in \left(0,1\right)$. When a consumer makes an EFPL purchase, the resulting utility is $\delta v-p_1$. Our emphasis is on identifying the optimal contract decision between the platform and the manufacturer, considering the competition between the EFPL and NB. We do not consider the situation where no consumer makes an EFPL purchase, i.e., assuming that $\delta >{\displaystyle \frac{p_1}{p_2}}$. We can obtain the consumer demand functions for the EFPL and NB, respectively:

$$d_1={\displaystyle \frac{p_2-p_1}{1-\delta }}-{\displaystyle \frac{p_1}{\delta }},$$

(1)

$$d_2=1-{\displaystyle \frac{p_2-p_1}{1-\delta }}.$$

(2)

3.2. When the Manufacturer Introduces an Online Direct Channel

In addition to selling the NB through the platform, the manufacturer also offers it via an online direct channel, pricing the NB in that channel at $p_3^Z$. Consumers have a higher preference for the platform channel because platforms tend to outperform direct manufacturer channels in terms of transactions, logistics, and other services. In line with the study by Li et al., we view the manufacturer’s direct channel as vertically differentiated from the platform [13].

As assumed in Section 3.1, consumers have a preference for NBs over EFPLs. Let $\delta \in \left(0,1\right)$ represent the level of preference for EFPLs sold on platforms, and let $\theta \in (0,1)$ represent the level of preference for NBs sold via the manufacturer’s online direct channel. In the existing research, $\theta$ is also defined as the acceptance of online direct channels by the consumer [13]. The level of preference for NBs sold through e-commerce platforms is normalized to 1.

Therefore, the utility for consumers when purchasing NBs through e-commerce platforms is $v-p_2$. That for consumers when purchasing the NB via the manufacturer’s direct channel is $\theta v-p_3$. The utility when purchasing the EFPL through the e-commerce platform is $\delta v-p_1$.

We need to evaluate two scenarios: one where the consumers’ preference for the NB available through the online direct sales channel exceeds that for the EFPL sold on e-commerce platforms, and another where it is lower.

(1) Consider the case where $\theta <\delta <1$, where consumer preference for the NB bought through the manufacturer’s online direct channel is less than that for the EFPL acquired via platforms. The market can be divided into four segments: the consumers with the highest willingness to purchase the NB through the platform; the consumers with the second-highest willingness to purchase the EFPL through the platform; the consumers with the third-highest willingness to purchase the NB through the manufacturer’s online direct channel; and finally, the consumers with the lowest willingness, who do not make any purchases.

Consumers who are indifferent when it comes to purchasing the NB or the EFPL through the platform are located at ${\displaystyle \frac{p_2-p_1}{1-\delta }}$; consumers who are indifferent when it comes to purchasing the NB through the online direct channel or the EFPL through the platform are located at ${\displaystyle \frac{p_1-p_3}{\delta -\theta }}$; and consumers who are indifferent when it comes to purchasing the NB through the online direct sales channel or not making a purchase are located at ${\displaystyle \frac{p_3}{\theta }}$. The consumer demand functions can be derived as follows:

$$d_1={\displaystyle \frac{p_2-p_1}{1-\delta }}-{\displaystyle \frac{p_1-p_3}{\delta -\theta }},$$

(3)

$$d_2=1-{\displaystyle \frac{p_2-p_1}{1-\delta }},$$

(4)

$$d_3={\displaystyle \frac{p_1-p_3}{\delta -\theta }}-{\displaystyle \frac{p_3}{\theta }}.$$

(5)

(2) Similarly, in the case where $\delta <\theta <1$, the market is divided into four segments. Consumers who are indifferent when it comes to purchasing the NB through the platform or through the online direct channel are located at ${\displaystyle \frac{p_2-p_3}{1-\theta }}$; consumers who are indifferent when it comes to purchasing the NB through the online direct channel or purchasing the EFPL through the platform are located at ${\displaystyle \frac{p_3-p_1}{\theta -\delta }}$; and consumers who are indifferent when it comes to purchasing EFPL through platform or not making a purchase are located at ${\displaystyle \frac{p_1}{\delta }}$. The consumer demand functions can be derived as follows:

$$d_1^{\prime }={\displaystyle \frac{p_3-p_1}{\theta -\delta }}-{\displaystyle \frac{p_1}{\delta }},$$

(6)

$$d_2^{\prime }=1-{\displaystyle \frac{p_2-p_3}{1-\theta }},$$

(7)

$$d_3^{\prime }={\displaystyle \frac{p_2-p_3}{1-\theta }}-{\displaystyle \frac{p_3-p_1}{\theta -\delta }}.$$

(8)

Due to the economies of scale and decreasing marginal cost effects associated with e-commerce platforms, the entry of a new manufacturer has a minimal effect on the platform’s costs. In contrast, the manufacturer needs to invest in establishing an online direct channel, which incurs costs related to website development, warehouse construction, leasing, and employee wages. This paper adopts a similar assumption to Zhang and Zhang by setting the fixed costs incurred by the manufacturer in establishing direct channels as $F$ [12]. For analytical convenience and without losing generality, this study assumes that the production costs for both the NB and EFPL are zero [50,56].

When an e-commerce platform signs an agency contract with a manufacturer, it charges a commission based on a certain percentage of sales revenue. Generally, the commission rate is not easily altered. Therefore, let $\alpha$ be treated as an exogenous variable that satisfies the following condition: $0<\alpha <1/2$ [57].

Table 3. Notations.

Parameters and Variables	Description
$v$	Consumer valuation of the NB.
$d_i^Z$	Demand for $i$ under strategy $Z$.
$p_i^Z$	Price for $i$ under strategy $Z$.
$w$	Wholesale price of the NB.
$\delta$	Consumer eco-friendly preference for the EFPL.
$\theta$	Consumer preference for the online direct channel.
$\alpha$	Commission rate.
$F$	Fixed costs for the manufacturer to establish an online direct channel.
${\prod }_k^Z$	Profit of Firm $k$ under Strategy $Z$.
Subscript $i=1,2,3$	Denote the EFPL, the NB sold through the platform channel, and the NB sold through the online direct channel, respectively.
Subscript $\mathrm{k}=P,M$	Denote the platform and the manufacturer, respectively.
Superscript $Z=R,A$	Denote the wholesale selling and agency selling.
Superscript $“*”,“**”$	Denote the equilibrium results under the conditions of $\theta <\delta <1$ and $\delta <\theta <1$.

summarizes notations used in the model for easy reference.

3.3. Sequence of Events

The sequence of events is as follows. First, the platform chooses between adopting a wholesale or agency selling format. Next, the manufacturer determines whether to launch an online direct channel. Depending on the platform’s selling format strategy (either wholesale or agency selling) and the manufacturer’s entry strategy (to introduce an online direct channel or not), we examine the following four cases, as illustrated in Figure 1. The pricing sequences under these four scenarios are as follows.

In Scenario RN, the platform adopts the wholesale selling format and the manufacturer does not introduce the online direct channel. The wholesale price of the NB is first determined by the manufacturer, followed by the platform setting the prices for both the NB and the EFPL. In Scenario RE, the platform chooses the wholesale selling format and the manufacturer introduces the online direct channel. The manufacturer first determines the wholesale price of the NB, then the platform sets the prices for the NB and the EFPL, and finally, the manufacturer sets the price of the NB for the online direct channel. In Scenario AN, the platform adopts the agency selling format and the manufacturer does not introduce the online direct channel. The manufacturer first determines the price of the NB in the platform, then the platform sets the price for the EFPL. In Scenario AE, the platform adopts the agency selling format and the manufacturer introduces the online direct channel. The prices of the NB and EFPL in the platform channel are determined sequentially. Finally, the manufacturer sets the price of the NB in the direct channel. The game sequences related to the above cases have been extensively used in related studies [11,12,13].

4. Equilibrium Results

This section presents the equilibrium results under four different scenarios. In the equilibrium analysis, we employ the concept of the subgame-perfect Nash equilibrium and solve for the game equilibrium using backward induction. The proofs of the results in this paper are provided in the Supplementary Materials.

4.1. The Platform Adopts a Wholesale Selling Format

4.1.1. Scenario RN: Wholesale Selling Without Online Direct Channel

When the platform adopts a wholesale selling format, the manufacturer sells the NB to the platform at a wholesale price $w^{RN}$. The platform sells the EFPL and NB to consumers at prices $p_1^{RN}$ and $p_2^{RN}$, respectively. At this point, the profit functions for the platform and the manufacturer are as follows:

$${\prod }_P^{RN}=d_1^{RN}{p}_1^{RN}+d_2^{RN}(p_2^{RN}-w^{RN}),$$

(9)

$${\prod }_M^R=w{d}_2^M.$$

(10)

Using the backward induction method to solve for the equilibrium, the equilibrium results for Scenario RN are summarized in

Table 4. Equilibrium results for Scenario RN.

$p_1^{RN}={\displaystyle \frac{\delta }{2}}$	${\prod }_P^{RN}={\displaystyle \frac{1+3\delta }{16}}$
$p_2^{RN}={\displaystyle \frac{3-\delta }{4}}$	${\prod }_M^{RN}={\displaystyle \frac{1-\delta }{8}}$
$w^{RN}={\displaystyle \frac{1-\delta }{2}}$

.

Lemma 1.

${\displaystyle \frac{\partial p_1^{RN}}{\partial \delta }}>0$, ${\displaystyle \frac{\partial p_2^{RN}}{\partial \delta }}<0$, ${\displaystyle \frac{\partial w^{RN}}{\partial \delta }}<0$, ${\displaystyle \frac{\partial {\prod }_P^{RN}}{\partial \delta }}>0$, ${\displaystyle \frac{\partial {\prod }_M^{RN}}{\partial \delta }}<0$.

Lemma 1 indicates that when the platform adopts a wholesale selling format, an increase in consumer preference for the EFPL leads to higher retail prices and equilibrium profits for the platform. In contrast, the wholesale price, retail price, and equilibrium profits of the manufacturer decrease as consumer preference for the EFPL rises.

From Lemma 1, we can deduce that $d_1^{RN}=d_2^{RN}={\displaystyle \frac{1}{4}}$. This result indicates that for the platform, the increase in consumer preference for the EFPL does not influence its demand but leads to a higher retail price for the EFPL, thereby increasing the platform’s profits. In contrast, for the NB, the increase in consumer preference for the EFPL does not impact demand either, but it lowers the wholesale price, resulting in decreased profits for the manufacturer. Lemma 1 reflects the relationship between the platform and the manufacturer. On one hand, the manufacturer relies on the platform to sell products to end consumers, and it does not have pricing power. As consumer preference for the EFPL increases, the manufacturer is likely to lower wholesale prices to enhance product competitiveness. On the other hand, the platform holds pricing power over both products. As consumer preference for the EFPL increases, the platform will raise the price of the EFPL and decrease the price of the NB to maximize profits.

4.1.2. Scenario RE: Wholesale Selling with Online Direct Channel

When the platform adopts the wholesale selling format, it produces and sells the EFPL, whereas the manufacturer supplies the NB product to the platform at a wholesale price $w^{RE}$. The platform sells the EFPL and NB to consumers at prices $p_1^{RE}$ and $p_2^{RE}$, respectively. The manufacturer sells the NB through the online direct channel at price $p_3^{RE}$ and incurs a fixed cost $F$ for establishing the online direct channel. At this point, the profit functions for the platform and the manufacturer are as follows:

$${\prod }_P^{RE}=d_1^{RE}{p}_1^{RE}+d_2^{RE}(p_2^{RE}-w_2^{RE}),$$

(11)

$${\prod }_M^{RE}=d_3^{RE}{p}_3^{RE}+w^{RE}{d}_2^{RE}-F.$$

(12)

The equilibrium results for Scenario RE are summarized in

Table 5. Equilibrium results for Scenario RE.

$\theta <\delta <1$	$\delta <\theta <1$
$p_1^{RE*}={\displaystyle \frac{\delta (\delta -\theta )}{2\delta -\theta }}$	$p_1^{RE**}={\displaystyle \frac{\delta (\delta -\theta )A_2}{2\left({\delta }^3(1+\theta )+\delta \theta \left(8+3\theta -{\theta }^2\right)-A_1\right)}}$
$p_2^{RE*}=-{\displaystyle \frac{2{\delta }^2+\delta (-6+\theta )+3\theta }{8\delta -4\theta }}$	$p_2^{RE**}={\displaystyle \frac{(\delta -2\theta +\delta \theta )A_2}{2\left({\delta }^3(1+\theta )+\delta \theta \left(8+3\theta -{\theta }^2\right)-A_1\right)}}$
$p_3^{RE*}={\displaystyle \frac{(\delta -\theta )\theta }{2(2\delta -\theta )}}$	$p_3^{RE**}={\displaystyle \frac{(\delta -\theta )\left({\delta }^2{(1+\theta )}^2+A_4\right)}{2\left({\delta }^3(1+\theta )+\delta \theta \left(8+3\theta -{\theta }^2\right)-A_1\right)}}$
$w^{RE*}={\displaystyle \frac{1-\delta }{2}}$	$w^{RE**}={\displaystyle \frac{(-1+\delta )\left(\delta \theta \left(-4+\theta -{\theta }^2\right)+A_3\right)}{{\delta }^3(1+\theta )+\delta \theta \left(8+3\theta -{\theta }^2\right)-A_1}}$
${\prod }_P^{RE*}={\displaystyle \frac{2\delta +6{\delta }^2-\theta -7\delta \theta }{32\delta -16\theta }}$	${\prod }_P^{RE**}={\displaystyle \frac{A_5}{4\left({\delta }^3(1+\theta )+\delta \theta \left(8+3\theta -{\theta }^2\right)-A_1\right)}}$
${\prod }_M^{RE*}={\displaystyle \frac{-4{\delta }^3+{\theta }^2-\delta \theta (4+3\theta )+{\delta }^2(4+6\theta )}{8{(-2\delta +\theta )}^2}}-F$	${\prod }_M^{RE**}={\displaystyle \frac{(-1+\delta )\left(4{\theta }^2+{\delta }^2\left(1+3{\theta }^2\right)+A_6\right)}{4\left({\delta }^3(1+\theta )+\delta \theta \left(8+3\theta -{\theta }^2\right)-A_1\right)}}-F$

The expressions from $A_1$ to $A_6$ are provided in the Supplementary Materials, for detailed reference.

.

4.1.3. The Impact of Introducing an Online Direct Channel

Lemma 2.

(1) 

When the platform adopts a wholesale selling format and $\theta <\delta <1$, $w^{RE*}=w^{RN}$$p_1^{RE*}<p_1^{RN}$, $p_2^{RE*}<p_2^{RN}$, ${\prod }_P^{RE*}<{\prod }_P^{RN}$.

(2) 

When the platform adopts a wholesale selling format and $\delta <\theta <1$, $p_1^{RE**}<p_1^{RN}$, $p_2^{RE**}<p_2^{RN}$, ${\prod }_P^{RE**}<{\prod }_P^{RN}$; when $\delta >{\delta }_1$, $w^{RE**}>w^{RN}$ and when $0<\delta <{\delta }_1$, $w^{RE**}<w^{RN}$.

Lemma 2 (1) indicates that when the platform adopts a wholesale selling format, the manufacturer’s choice to establish an online direct channel does not have a direct impact on the wholesale price within the platform channel. In comparison to the RN scenario, the equilibrium wholesale price stays the same in the RE scenario, while the equilibrium prices for the EFPL and NB in the platform channel decrease, resulting in a decline in the platform’s profits. The launching of an online direct channel allows the manufacturer to attract consumers, generating additional profits. But the decline in retail prices will reduce the platform’s profits. It suggests that the manufacturer’s introduction of an online direct channel intensifies market competition. Consequently, the introduction of an online direct channel by the manufacturer negatively impacts the platform’s profits. Overall, compared to Scenario RN, and ignoring the costs associated with it, the manufacturer’s profits increase, while the platform’s profits decline. This demonstrates that the introduction of an online direct channel not only alters the structure of sales channels but also affects the profit distribution among market participants.

Lemma 2 (2) indicates that, compared to the RN scenario, in the RE scenario, when $\delta >{\delta }_1$, the platform’s wholesale price increases; when $0<\delta <{\delta }_1$, the wholesale price decreases. The prices of the NB and EFPL decline, leading to a reduction in the platform’s profits. When the consumer preference for the manufacturer’s direct channel is stronger, the introduction of the online direct channel may cause changes in the wholesale price. If the consumer preference for the EFPL is weak ($0<\delta <{\delta }_1$), the EFPL is not a good substitute for the NB, resulting in weaker competition between brands. In this case, the manufacturer is motivated to reduce the wholesale price to boost sales via the platform channel, thereby increasing profits. Conversely, when consumer preference for the EFPL is strong, since the EFPL and NB are close substitutes for one another, the manufacturer is incentivized to draw consumers to the online direct channel by raising the wholesale price, thereby mitigating intense competition in the platform channel. Similarly, as in Lemma 1, compared to the RN scenario, and disregarding the costs associated with it, the manufacturer’s profitability is enhanced, while the platform’s profits decline.

4.2. The Platform Adopts an Agency Selling Format

4.2.1. Scenario AN: Agency Selling Without Online Direct Channel

When the platform operates under an agency selling format, it produces and sells the EFPL, while the manufacturer sells the NB through the platform, which charges a commission rate of $\alpha$ for the manufacturer. The platform sells the EFPL to consumers at price $p_1^{AN}$, and the manufacturer sells the NB to consumers at price $p_2^{AN}$. At this point, the profit functions for the platform and the manufacturer are as follows:

$${\prod }_P^{AN}=d_1^{AN}{p}_1^{AN}+\alpha d_2^{AN}{p}_2^{AN},$$

(13)

$${\prod }_M^{AN}=(1-\alpha )d_2^{AN}{p}_2^{AN}.$$

(14)

The equilibrium results for Scenario AN are summarized in

Table 6. Equilibrium results for Scenario AN.

$p_1^{AN}={\displaystyle \frac{(1+\alpha )(-1+\delta )\delta }{-4+\delta +\alpha \delta }}$	${\prod }_P^{AN}={\displaystyle \frac{(1-\delta )\left(\delta +4\alpha -{\alpha }^2\delta \right)}{{(4-\delta -\alpha \delta )}^2}}$
$p_2^{AN}={\displaystyle \frac{2(-1+\delta )}{-4+\delta +\alpha \delta }}$	${\prod }_M^{AN}={\displaystyle \frac{4(1-\delta )(1-\alpha )}{{(4-\delta -\alpha \delta )}^2}}$

.

Lemma 3.

When $\delta \in (0,{\delta }_2)$ , ${\displaystyle \frac{\partial p_1^{AN}}{\partial \delta }}>0$; when $\delta \in ({\delta }_2,1)$, ${\displaystyle \frac{\partial p_1^{AN}}{\partial \delta }}<0$.${\displaystyle \frac{\partial p_2^{AN}}{\partial \delta }}<0,{\displaystyle \frac{\partial {\prod }_P^{AN}}{\partial \alpha }}>0,{\displaystyle \frac{\partial {\prod }_M^{AN}}{\partial \alpha }}<0,{\displaystyle \frac{\partial {\prod }_M^{AN}}{\partial \delta }}<0$.  When $\delta \in (0,{\delta }_3)$, ${\displaystyle \frac{\partial {\prod }_P^{AN}}{\partial \delta }}>0$; when $\delta \in ({\delta }_3,1)$, ${\displaystyle \frac{\partial {\prod }_P^{AN}}{\partial \delta }}<0$.

When the platform utilizes an agency selling format, an increase in the commission rate leads to higher profits for the platform, while the manufacturer’s profits decline. Additionally, as consumer preference for the EFPL increases, the manufacturer’s equilibrium profits will decline, while the platform’s equilibrium profits will exhibit a non-monotonic change, initially increasing and then decreasing.

Lemma 3 indicates that, first, the commission rate essentially represents the revenue distribution between manufacturer and platform regarding the NB. As the commission rate rises, the platform’s revenue similarly increases, whereas the manufacturer’s revenue declines. Second, when consumer preference for the EFPL rises, this leads to a decrease in the price of the NB, while the price of the EFPL exhibits a non-monotonic change, initially increasing and then decreasing. Simultaneously, the demand for both products increases, as the two kinds of products become increasingly homogeneous with the rise in brand preference, intensifying competition. To capture more demand, the platform is compelled to lower the price.

In summary, as consumer preference for the EFPL increases, the manufacturer’s equilibrium profits exhibit a declining trend, while the platform’s equilibrium profits initially rise before declining. This indicates that, for the platform’s EFPL strategy, a high-quality EFPL may not always be advantageous. Therefore, it is essential to comprehensively consider both quality and price factors.

4.2.2. Scenario AE: Agency Selling with Online Direct Channel

When the platform adopts an agency selling format and the manufacturer launches an online direct channel, the platform charges the manufacturer a commission rate of $\alpha$. The manufacturer sells the NB to consumers at prices $p_2^{AE}$ and $p_3^{AE}$ through the platform channel and the online direct channel, respectively, and incurs a fixed cost $F$ for selling products through the direct channel. The platform sells the EFPL at price $p_1^{AE}$ to consumers. The profit functions for the platform and the manufacturer are as follows:

$${\prod }_P^{AE}=\alpha d_2^{AE}{p}_2^{AE}+d_1^{AE}{p}_1^{AE},$$

(15)

$${\prod }_M^{AE}=(1-\alpha )d_2^{AE}{p}_2^{AE}+d_3^{AE}{p}_3^{AE}-F.$$

(16)

The equilibrium results for Scenario AE are summarized in

Table 7. Equilibrium results for Scenario AE.

$\theta <\delta <1$	$\delta <\theta <1$
$p_1^{AE*}=-{\displaystyle \frac{\alpha (1+\alpha )(-1+\delta )\delta (\delta -\theta )}{\delta (\delta -\theta )+{\alpha }^2\delta (\delta -\theta )+B_1}}$	$p_1^{AE**}=-{\displaystyle \frac{\alpha (-1+\delta )\delta (\delta -\theta )(-1+\theta )}{-\delta \left(\delta -\theta \right)\left(-1+\theta \right)+B_3}}$
$p_2^{AE*}=-{\displaystyle \frac{(-1+\delta )(-2\alpha \delta +\alpha \theta +\alpha \delta \theta )}{\delta (\delta -\theta )+{\alpha }^2\delta (\delta -\theta )+B_1}}$	$p_2^{AE**}={\displaystyle \frac{\alpha (-1+\delta )(-1+\theta )(\delta -2\theta +\delta \theta )}{-\delta \left(\delta -\theta \right)\left(-1+\theta \right)+B_3}}$
$p_3^{AE*}={\displaystyle \frac{\alpha (1+\alpha )(-1+\delta )(\delta -\theta )\theta }{2\left(\delta (\delta -\theta )+{\alpha }^2\delta (\delta -\theta )+B_1\right)}}$	$p_3^{AE**}={\displaystyle \frac{\delta (-1+\alpha -3\theta +\alpha \theta )+B_4}{2\left(-\delta \left(\delta -\theta \right)\left(-1+\theta \right)+B_3\right)}}$
${\prod }_P^{AE*}=-{\displaystyle \frac{{\alpha }^2(-1+\delta )(\delta (-2+\theta )+\theta )}{2\left(\delta (\delta -\theta )+{\alpha }^2\delta (\delta -\theta )+B_1\right)}}$	${\prod }_P^{AE**}-{\displaystyle \frac{{\alpha }^2(-1+\delta )(-1+\theta )(\delta +(-2+\delta )\theta )}{2\left(-\delta \left(\delta -\theta \right)\left(-1+\theta \right)+B_3\right)}}$
${\prod }_M^{AE*}={\displaystyle \frac{\alpha (-1+\delta )B_2}{4{\left(\delta (\delta -\theta )+{\alpha }^2\delta (\delta -\theta )+B_1\right)}^2}}-F$	${\prod }_M^{AE**}={\displaystyle \frac{\alpha (-1+\delta )(-1+\theta )B_5}{4{\left(-\delta \left(\delta -\theta \right)\left(-1+\theta \right)+B_3\right)}^2}}-F$

The expressions from $B_1$ to $B_5$ are provided in the Supplementary Materials, for detailed reference.

.

4.2.3. The Impact of Introducing an Online Direct Channel

Lemma 4.

(1) 

When $\theta <\delta <1$ and $0<\delta <{\delta }_4$ or ${\delta }_5<\delta <1$, $p_2^{AE*}<p_2^{AN}$; when ${\delta }_4<\delta <{\delta }_5$, $p_2^{AE*}>p_2^{AN}$.

(2) 

When $\delta <\theta <1$ and $\alpha <{\alpha }_1$ or $\alpha >{\alpha }_2$, $p_1^{AE**}<p_1^{AN}$, $p_2^{AE**}<p_2^{AN}$; when ${\alpha }_1<\alpha <{\alpha }_2$, $p_1^{AE**}>p_1^{AN},p_2^{AE**}>p_2^{AN}$.

Lemma 4 indicates that when the platform employs an agency selling format, the manufacturer’s introduction of an online direct channel does not always lead to a reduction in the prices of the EFPL and NB on the platform channel. When $\theta <\delta <1$, in the AE scenario, if consumer preference for the EFPL is either high or low, the NB price in the platform channel is lower compared to the AN scenario. Conversely, when consumer preference for the EFPL is in an intermediate range, the NB price is higher than in the AN scenario. When consumer preference for the EFPL is at an intermediate level, the sales price of the EFPL is higher than in the AN scenario, whereas when consumer preference for the EFPL is high or low, the sales price of the EFPL is lower than in the AN scenario. When $\delta <\theta <1$, in the AE scenario, the prices of the EFPL and NB are lower than in the AN scenario when the commission is either low or high; however, when the commission level is moderate, the prices of the EFPL and NB are higher compared to the AN scenario.

In other words, under the agency selling format, the online direct channel does not always intensify competition; rather, it can promote collusion by a manufacturer across different channels, thereby alleviating conflicts between them.

5. Equilibrium Analysis

By comparing the equilibrium profits of the manufacturer and the platform under the four scenarios, we can derive the manufacturer’s decision to introduce an online direct channel and the platform’s choice of selling format at equilibrium. First, we solve for the optimal direct channel introduction decision of the manufacturer under either the wholesale or agency selling format employed by the platform. Following this, we determine the platform’s optimal selling format based on the anticipated optimal decisions made by the manufacturer.

5.1. Manufacturer’s Channel Decision

5.1.1. Channel Decision Under the Wholesale Selling Format

Proposition 1.

Under the wholesale selling format, the manufacturer’s optimal channel decisions are as follows:

(1) 

Under the wholesale selling format and $\theta <\delta < 1$. If and only if $0<F<F_1$, ${\prod }_M^{RE*}>{\prod }_M^{RN}$, the manufacturer launches the online direct channel, where $\partial F_1/\partial \delta >0$, and when $0<\theta <2/3$ and $3\theta /2<\delta <1$, $\partial F_1/\partial \theta >0$.

(2) 

Under the wholesale selling format and $\delta <\theta < 1$. If and only if $0<F<F_2$, ${\prod }_M^{RE**}>{\prod }_M^{RN}$, the manufacturer launches the online direct channel, where $\partial F_2/\partial \theta >0$, and when $0<\delta <\tilde{\delta }$ and ${\theta }_1<\theta <{\theta }_2$, $\partial F_2/\partial \delta >0$.

Proposition 1 (1) indicates that under the wholesale format, and $\theta <\delta < 1$, if the fixed cost of the manufacturer introducing an online direct channel is below a certain threshold, the manufacturer introduces the online direct channel. Moreover, this threshold increases as consumer preference for the EFPL grows. In other words, as consumer preference for the EFPL increases, manufacturers are more inclined to establish new distribution channels to capture more profits. Additionally, when the consumers’ preference for the manufacturer’s direct channel is low and their preference for the EFPL is relatively high, the motivation for the manufacturer to introduce an online direct channel becomes stronger as the consumers’ preference for the online direct channel increases. This occurs because, when consumers’ preference for the EFPL is high and their initial preference for the direct channel is low, an increase in preference for the direct channel allows the manufacturer to draw consumers to purchase from it, thus avoiding intense competition with the standardized EFPL. Conversely, when consumer preference for the direct channel further increases, and given that $\theta <\delta < 1$, consumer preference for the EFPL also remains high. According to Lemma 1, the wholesale price set by the manufacturer for the platform, $\partial w/\partial \delta <0$, will be driven down to a very low level. Furthermore, at this point, the competition between the manufacturer’s online direct channel and the platform channel becomes increasingly homogenized, resulting in a gradual decline in profits from the manufacturer’s online direct channel, which subsequently weakens the motivation to establish it.

Proposition 1 (2) shows that under the wholesale format and $\delta <\theta < 1$, if the fixed cost for the manufacturer to introduce an online direct channel is below the threshold $F_2$, the manufacturer introduces the online direct channel. Furthermore, the threshold $F_2$ increases as $\theta$ rises. In other words, as consumer preference for the online direct channel increases, manufacturers are more inclined to establish it. At this point, a manufacturer can effectively attract consumers to shift their purchases to its direct channel, alleviating the efficiency losses caused by double marginalization. Additionally, when consumer preference for the EFPL is low, and consumer preference for the online direct channel is at an intermediate level, the motivation for the manufacturer to introduce an online direct channel becomes stronger as consumer preference for the EFPL increases. This allows manufacturers to counter the impact of the EFPL on their profits by establishing direct channels. Conversely, when consumer preference is high for both the EFPL and the manufacturer’s direct channel, the intense competition between brands and channels can harm manufacturer profits, thereby reducing the motivation to establish direct channels.

5.1.2. Channel Decision Under the Agency Selling Format

Proposition 2.

Under the agency selling format, the manufacturer’s optimal channel decisions are as follows:

(1) 

Under the agency selling format and $\theta <\delta < 1$,  ${\prod }_M^{AN}>{\prod }_M^{AE*}$  always holds; the manufacturer does not launch an online direct channel.

(2) 

Under the agency selling and $\delta <\theta < 1$. If and only if $0<F<F_3$, ${\prod }_M^{AE**}>{\prod }_M^{AN}$ holds and the manufacturer launches an online direct channel, where $\partial F_3/\partial \alpha >0$.

Proposition 2 (1) indicates that under the agency selling format, the manufacturer cannot improve its profits by introducing an online direct channel. This contrasts with Li et al., whose study is based on the context of the platform adopting a wholesale selling format [13]. When the platform adopts a wholesale selling format, the manufacturer can always reduce the efficiency loss caused by double marginalization by establishing an online direct sales channel, thereby increasing profits. However, in the case of the agency selling format, the manufacturer determines the price of the NB. If the manufacturer launches a new distribution channel, it will trigger intense competition among the manufacturer’s brands across different distribution channels. This leads to a decrease in the manufacturer’s profits. More importantly, because consumers prefer the EFPL to the manufacturers’ direct online channels, the loss of profits from the competitive effects of establishing a new channel outweighs the increase in profits from the demand generated through that channel.

Proposition 2 (2) indicates that under the agency selling format, the manufacturer will establish an online direct channel if the fixed cost of this channel is less than the threshold $F_3$. Moreover, as the commission fees increase, the manufacturer’s motivation to establish an online direct channel also strengthens. This result differs from that of Proposition 2 (1). This is because when consumers favor the manufacturer’s online direct channel over the EFPL, the manufacturer can successfully draw more demand through this channel, thereby boosting profits. As the platform’s commission increases, the manufacturer’s motivation to introduce an online direct channel also grows. This is intuitive, as high commissions erode the manufacturer’s profits. Therefore, with the increase in commissions, the motivation for the manufacturer to establish an online direct channel also intensifies. This allows the manufacturer to guide consumers towards purchasing through its own channel, thereby avoiding the platform’s high commission fees.

5.2. Decision on the Platform’s Selling Format

Anticipating the manufacturer’s channel decision, as illustrated in Propositions 1 and 2, the platform’s selling format decision is summarized in Propositions 3 and 4.

Proposition 3.

When $\theta <\delta < 1$, the platform’s selling format decisions are as follows:

(1) 

When $F>F_1$, the platform anticipates that the manufacturer will not introduce the online direct channel. When $0<\delta <\widehat{\delta }$, the platform adopts the agency selling format (${\prod }_P^{AN}>{\prod }_P^{RN}$); when $\widehat{\delta }<\delta <1$, the platform anticipates that the manufacturer will not introduce the online direct channel (${\prod }_P^{AN}<{\prod }_P^{RN}$); the platform adopts the wholesale selling format. Moreover, $\partial \widehat{\delta }/\partial \alpha >0$.

(2) 

When $F<F_1$, the platform anticipates that the manufacturer will introduce the online direct channel under the wholesale selling format, whereas under the agency selling format, the manufacturer will not introduce the online direct channel. When $0<\theta <\widehat{\theta }$, the platform adopts the wholesale selling format (${\prod }_P^{AN}<{\prod }_P^{RE*}$); when $\widehat{\theta }<\theta <\delta$, the platform anticipates that the manufacturer will not introduce the online direct channel (${\prod }_P^{AN}>{\prod }_P^{RE*}$); the platform adopts the wholesale selling format. Moreover, $\partial \widehat{\theta }/\partial \alpha <0$, $\partial \widehat{\theta }/\partial \delta >0$.

Proposition 3 (1) indicates that when the cost of the manufacturer’s online direct channel is high, the manufacturer will not establish such a channel, leading the platform to anticipate that the manufacturer will refrain from entering the direct channel. When consumer preference for the EFPL is low, the platform adopts the agency selling format; conversely, when this preference is high, the platform opts for the wholesale selling format. This shift occurs because a low preference results in weak substitutability between the EFPL and the NB, leading to less competition. The agency selling format effectively avoids the efficiency losses associated with double marginalization, thus enhancing the platform’s profits. However, as consumer preference for the EFPL increases, substitutability grows, intensifying competition. Under these conditions, the wholesale selling format allows the platform to set prices for both brands independently, effectively colluding to mitigate profit reductions caused by fierce competition. Additionally, the threshold for adopting this model rises with the platform’s commission rate, as higher commissions increase the platform’s revenue share from the manufacturer, partially offsetting profit losses due to intense brand competition. Consequently, a higher commission rate leads the platform to tolerate greater competition between the EFPL and NB, resulting in a higher threshold for collusion via the wholesale model.

Proposition 3 (2) indicates that when the cost of the manufacturer’s direct channel is low, the platform anticipates that the manufacturer will introduce the online direct channel when the platform adopts the wholesale selling format. When $0<\theta <\widehat{\theta }$, the platform expects the manufacturer to introduce the online direct channel and choose the wholesale selling format; when $\widehat{\theta }<\theta <\delta$, the platform anticipates that the manufacturer will not introduce the online direct channel and will select the agency selling format.

The reasoning is that when consumer preference for the manufacturer’s direct channel is low, the platform anticipates that the manufacturer’s establishment of an online direct channel will not significantly attract consumers away from the platform. Additionally, by adopting the wholesale selling format, the platform can benefit from brand collusion effects that mitigate intense competition. Conversely, when consumer preference for the online direct channel is high, the platform expects that the manufacturer will not establish an online direct channel when adopting the agency selling format. This approach not only enhances efficiency by reducing the double marginalization effect but also avoids the intense competition that may arise from the manufacturer introducing an online direct channel. Thus, in this case, the platform opts for the agency selling format.

Moreover, as consumer preference for the EFPL increases, the threshold for adopting the agency selling format rises. This is because the platform becomes more resilient to demand losses resulting from the manufacturer establishing an online direct channel. Consequently, as consumer preference for the EFPL continues to grow, the collusion effects associated with the wholesale selling format contribute more significantly to the platform’s profits than the efficiency gains and profit enhancements achieved through the agency selling format that mitigates competition from the manufacturer’s online direct channel. Due to the complexity of the analytical solution, we employ a numerical study to examine the impacts of consumer preferences for brands and sales channels on the platform’s selling format decisions. We present the platform’s equilibrium selling format in Figure 2. Region A indicates that the platform adopts an agency selling format, while Region R signifies that the platform employs a wholesale selling format.

Proposition 4.

When $\delta <\theta < 1$, the platform’s selling format decisions are as follows:

(1) 

When $F>F_2$, the platform anticipates that the manufacturer will not introduce the online direct channel. When $0<\delta <\widehat{\delta }$, the platform adopts the agency selling format (${\prod }_P^{AN}>{\prod }_P^{RN}$); when $\widehat{\delta }<\delta <1$, the platform anticipates that the manufacturer will not introduce the online direct channel (${\prod }_P^{AN}>{\prod }_P^{RN}$); the platform adopts the wholesale selling format. Moreover, $\partial \widehat{\delta }/\partial \alpha >0$.

(2) 

When  $F_3<F<F_2$, the platform anticipates that the manufacturer will introduce the online direct channel under the wholesale selling format, whereas under the agency selling format, the manufacturer will not introduce the online direct channel. Since  ${\prod }_P^{AN}>{\prod }_P^{RE**}$ always holds, the platform will adopt the agency selling format.

(3) 

When  $F<F_3$, the platform anticipates that the manufacturer will always introduce an online direct channel, regardless of the selling format it adopts. Since  ${\prod }_P^{AE**}>{\prod }_P^{RE**}$ always holds, the platform will adopt the agency selling format.

Proposition 4 (1) indicates that when the manufacturer’s online direct channel costs are high, the platform expects the manufacturer not to establish a direct sales channel. In this case, when consumer preference for the EFPL is low, the platform adopts the agency selling format; when consumer preference for the EFPL is high, it adopts the wholesale selling format. This is because, with low consumer preference for the EFPL, the substitutability between the EFPL and NB is low, resulting in weaker competition and higher wholesale prices supplied by the manufacturer. Compared to the wholesale selling format, the agency selling format effectively mitigates the efficiency loss from double marginalization and can enhance platform profits. As consumer preference for the EFPL increases, the substitutability between the EFPL and NB also increases, intensifying competition and significantly lowering the wholesale prices from the manufacturer. In this scenario, adopting the wholesale selling format allows the platform to set prices independently for both brands, effectively forming a collusion between the EFPL and NB, which can help avoid profit reductions due to intense competition. Additionally, the threshold increases with the rise in the platform’s commission rate. As the commission rate rises, the platform’s share of revenue from the manufacturer grows, partially offsetting profit losses from intense brand competition. Thus, when the commission rate is high, the threshold for profit loss due to competition between the EFPL and NB in the agency selling format becomes larger.

When $F<F_2$, the platform will always adopt the agency selling format. This conclusion differs from Proposition 3 (2). Through numerical simulations, we find that when the platform adopts the wholesale selling format, the fixed cost threshold for the manufacturer to introduce the online direct channel is higher than under the agency selling format ($F_2>F_3$), leading us to discuss the platform’s selling format decision in two intervals.

In the first interval ($F_3<F<F_2$), the platform expects the manufacturer to introduce the online direct channel when adopting the wholesale selling format, while it anticipates that the manufacturer will not introduce the online direct channel under the agency selling format. This expectation arises because, on one hand, consumer preference for the manufacturer’s direct channel is consistently greater than their preference for the platform’s EFPL. The introduction of an online direct channel by the manufacturer can lead to a demand diversion effect, significantly reducing the demand for the EFPL. On the other hand, the manufacturer’s entry into the online direct channel can also create channel competition, resulting in a decrease in the price of the NB. Both of these factors contribute to a decline in the platform’s revenue. Therefore, by adopting the agency model, the platform can prevent the manufacturer from introducing an online direct channel, thereby avoiding profit reductions due to channel competition and decreases in demand. In this scenario, the manufacturer also does not wish to introduce an online direct channel, indicating that the platform’s decision aligns with the manufacturer’s interests, achieving a win-win outcome.

In the second interval ($F<F_3$), the platform expects the manufacturer to introduce an online direct channel regardless of the selling format it adopted, and the platform can effectively manage pricing through the agency selling format. In this case, the manufacturer sets prices for both channels, internalizing any conflicts and thereby facilitating price collusion between the channels, which prevents intense competition that could erode platform profits. In contrast, under the wholesale selling format, three effects can lead to profit loss for the platform. First, the wholesale selling format incurs a significant demand diversion effect. Here, the manufacturer sets the price for the NB in the online direct channel, while the platform sets the price for the NB. The cost advantage of the manufacturer’s online direct channel results in greater demand loss for the platform compared to the agency selling format, adversely affecting platform profits. Second, consumer preference for the manufacturer’s online direct channel is stronger than for the EFPL. This preference exacerbates the competitive disadvantage of the EFPL under the wholesale selling format, further diminishing profits from the EFPL. Lastly, the wholesale selling format introduces the classic double marginalization effect, leading to a reduction in overall efficiency, which the agency selling format effectively mitigates. In summary, adopting the agency selling format is the optimal choice for the platform. Thus, when $F<F_2$, the platform will always adopt the agency selling format. However, the underlying mechanisms for adopting the agency selling format differ between the ranges $F_3<F<F_2$ and $F<F_3$.

To facilitate understanding, we employ a numerical study to examine the impacts of consumer preferences for brands and sales channels on the platform’s selling format decisions, shown in Figure 3. Region A indicates that the platform adopts an agency selling format, while Region R signifies that the platform employs a wholesale selling format.

6. Conclusions

In practice, decisions regarding the manufacturer’s introduction of online direct sales channels and platform selling format are common and non-negligible important issues. However, the existing literature often ignores the situation where the platform introduces an EFPL to compete with the NB in sustainable e-commerce supply chains.

In this paper, we develop a game-theoretic model with four scenarios that account for the manufacturer’s channel strategy and the platform’s selling format to derive equilibrium results. The model emphasizes two key features relevant to consumer utility: the variation in the perceived value of the NB across sales channels and the consumer’s eco-friendly preferences for the EFPL. Our main conclusions are as follows.

6.1. Main Findings

(1)

Under the platform’s wholesale selling format, the manufacturer is consistently motivated to establish a direct channel if the associated costs are low. The reason is that by selling online directly, the manufacturer can effectively reduce the efficiency loss brought by double marginalization under the wholesale selling format. Conversely, under the agency selling format and with consumers’ preference for the platform’s EFPL over the manufacturer’s online direct channel, the manufacturer will not be able to improve its profitability by selling online directly in any case. This differs from the findings of Li et al., as their study was based on the context of platforms entering into wholesale contracts with manufacturers [13]. However, when the platform adopts the agency selling format, the introduction of an online direct channel by the manufacturer can lead to fierce competition for NBs across different distribution channels, ultimately diminishing the manufacturer’s profitability. Moreover, since consumers prefer EFPLs over the manufacturer’s online direct channel, the introduction of the online direct channel leads to a lower NB price. Facing EFPL competition, the profit loss from the competition effect outweighs the gain from the increased demand brought by the new channel. At this point, the manufacturer’s direct channel always leads to a reduction in profits. When consumers favor the manufacturer’s direct channel over the platform’s EFPL and the fixed costs of establishing the direct channel are low, the manufacturer is likely to introduce an online direct channel. This is because, when consumers prefer the manufacturer’s direct channel over EFPL, the manufacturer can capture greater demand through the online direct channel, resulting in a profit increase that exceeds the losses from channel competition.

(2)

When consumers prefer the EFPL over the manufacturer’s online direct channel, the platform expects that the manufacturer will not implement a direct channel with a high cost. The platform chooses the agency selling format when consumer preference for the EFPL is low, and switches to the wholesale selling format when preference for the EFPL is high. This is because, with low consumer preference for EFPL, the substitutability between the EFPL and NB is reduced, leading to weaker competition. Compared to wholesale selling, agency selling can effectively avoid the efficiency reduction caused by double marginalization. When the cost of the manufacturer’s online direct channel is low, the platform anticipates that the manufacturer will launch the channel under the wholesale selling format, but will refrain from doing so under the agency selling format. When consumer preference for the manufacturer’s online direct channel is low, the platform expects the manufacturer to introduce the online direct channel and choose the wholesale selling format. In this case, the platform adopts the wholesale selling format, which can avoid too much competition through the brand collusion effect of the platform channel. When consumer preference for the manufacturer’s online direct channel is high, the platform expects the manufacturer not to introduce the online direct channel and to choose the agency selling format. Through adopting the agency selling format, the platform can improve efficiency by attenuating the double marginalization effect, and avoid a fierce channel competition brought about by the manufacturer’s online direct channel.

(3)

When consumers prefer the manufacturer’s online direct channel over the EFPL, the platform expects that the manufacturer will not introduce an online direct channel if it is costly for the manufacturer. The platform adopts the agency selling format when consumer preference for the EFPL is low, while it switches to the wholesale selling format when preference for the EFPL is high. When it is less costly for the manufacturer to introduce an online direct channel, it can be discussed in the context of two cases. The results indicate that the agency selling format is always the optimal choice for the platform, but the intrinsic mechanisms by which the platform determines its selling format are different.

6.2. Managerial Insights

From a practical perspective, the research underscores the significance of platforms selling eco-friendly private label products in influencing both the platform’s selling format decisions and the manufacturer’s channel strategies. Based on these findings, we offer the following management recommendations:

(1)

For e-commerce platforms: When a platform introduces eco-friendly private labels, adopting a wholesale selling format helps safeguard its profitability by avoiding direct competition with manufacturers, particularly when consumer preferences for environmentally friendly products are strong. This rationale aligns with the practices of platforms like Amazon.com and JD.com, which prioritize wholesale selling formats while maintaining high-quality private labels. However, when consumer preferences for eco-friendly products are weak, adopting an agency selling format becomes the optimal choice. This approach effectively mitigates efficiency losses caused by the double marginalization effect. To counter the manufacturer’s introduction of an online direct channel, the platform is better off adopting an agency selling format, particularly when consumer preferences for the manufacturer’s direct channel are strong. Additionally, to address the threat of manufacturer entry and to promote eco-friendly private labels, the platform can consider reducing its commission rates appropriately.

(2)

For manufacturers: When an e-commerce platform sells eco-friendly private labels alongside the manufacturer’s branded products, the manufacturer must carefully evaluate the decision to establish a direct channel, as this strategy does not always guarantee profitability. In addition to the costs of introducing an online direct channel, manufacturers must also consider the platform’s selling format and consumer preferences for environmentally friendly products. If the platform adopts a wholesale selling format and the cost of establishing a direct channel is low, it is optimal for the manufacturer to introduce the channel. However, when consumer preferences for eco-friendly products are strong, the manufacturer should avoid establishing a direct channel if the platform adopts an agency selling format. In such cases, the manufacturer and platform can achieve a win-win outcome.

(3)

For society: On one hand, the potential threat of manufacturers introducing direct channels can incentivize e-commerce platforms to adopt the more efficient agency selling format. If manufacturers can influence platforms to adopt this efficient selling format without actually establishing direct channels, it further avoids the sunk costs associated with direct channel investments, contributing to overall efficiency improvements. On the other hand, the less efficient wholesale selling format facilitates the development of eco-friendly private labels. Thus, when guiding the growth of eco-friendly private labels, governments should balance the efficiency contributions of these two effects. What is more, when consumer acceptance of eco-friendly private labels is initially low, governments can encourage their development through subsidies or other support mechanisms. Simultaneously, attention must be paid to the impact of platform selling formats on overall efficiency.

6.3. Contributions and Future Studies

The main contributions of this study are as follows: First, we investigate the decision-making process of an e-commerce platform regarding its choice of selling format and the manufacturer’s decision to introduce a direct channel when selling an eco-friendly private label. Second, we focus on analyzing the impact of consumers’ preferences for eco-friendly private label products and their varying preferences for sales channels on the decisions of platforms and manufacturers. Finally, based on the research findings, we provide relevant managerial insights.

This study can be expanded in several ways. Our model currently assumes deterministic market demand. However, in reality, platforms possess extensive user data due to their large user bases. Future research could integrate the e-commerce platform’s information-sharing strategies with manufacturers into the modeling framework. In addition, platforms’ private labels customized based on consumer information and platforms’ non-neutral recommendation strategies are noteworthy.