Agency or Reselling? Multi-Product Sales Mode Selection on E-Commerce Platform

Abstract

As environmental issues become increasingly prominent, the sustainable practices of enterprises, especially measures at the product level, have garnered widespread attention from scholars. Although numerous studies have explored suppliers’ sales strategies for green products, they often overlook the scenario where suppliers simultaneously sell both green and non-green products.This study focuses on the sales mode selection strategies of suppliers when providing green and non-green products through e-commerce platforms. Utilizing a game model, we analyze the equilibrium strategies between suppliers and e-commerce platforms, and conduct sensitivity analyses to evaluate the impact of key parameters on decision-making. The results reveal that there are significant differences in the strategic preferences of suppliers and e-commerce platforms. However, when commission rates are moderate and green products incur high production costs, these preferences tend to align, leading to Pareto optimal outcomes. Additionally, our findings demonstrate that adopting differentiated sales modes for the two product types can effectively mitigate the problem of double marginalization, thereby enhancing the efficiencyof supply chains. These insights provide valuable guidance for e-commerce platform managers and suppliers in making decisions on sales models for managing multiple types of products.

1. Introduction

The intensification of the global environmental crisis has accelerated the widespread acceptance of low-carbon and green sustainable development concepts, establishing them as an unstoppable mainstream trend. This paradigm shift has prompted an increasing number of companies to proactively adopt green products and integrate environmental strategies into their core operations. For example, Unilever has enhanced the utilization of renewable and recyclable materials in product packaging while committing to reducing single-use plastics. Similarly, Procter & Gamble, another large consumer packaged goods (CPG) company, has implemented plant-based packaging, effectively mitigating its environmental impact. Coca-Cola and PepsiCo are also pursuing comparable initiatives, progressively transitioning from traditional corrugated materials to recyclable plastic containers to minimize their environmental footprint. Beyond these CPG companies, numerous large manufacturing enterprises are actively exploring green and sustainable technologies to reduce environmental impacts. For instance, the well-known automaker BYD has launched a range of electric vehicles, while Mercedes-Benz introduced its all-electric EQ series, covering models from standard to luxury vehicles. However, these companies demonstrate significant variations in the cost structure associated with green products development. For large suppliers such as BYD and BMW, developing green products typically necessitates significant financial investment in technological research, hence the designation as “development-intensive green (DIG) products”. Compared to conventional products, the cost disparity for these green products is considerable and follows a convex function [1,2,3]. In contrast, for CPG companies like Unilever, the cost of green products primarily manifests as increased marginal costs, leading to their classification as “marginal-cost-intensive green (MIG) products (Marginal cost-intensive green products are our research object, because our research mainly considers the mode strategy when green products are distributed through e-commerce platforms. Development-intensive green products such as new energy vehicles are rarely distributed through e-commerce platforms.)”. The fundamental distinction between these green products and conventional ones lies in the utilization of more environmentally friendly, albeit more expensive, materials or packaging. Consequently, the cost difference between green and conventional products is consequently driven by the higher marginal cost of eco-friendly materials. Companies such as Unilever and Procter & Gamble have successfully established significant market advantages by implementing a comprehensive strategy incorporating both MIG products and non-green (NG) products. By simultaneously addressing the needs of environmentally-conscious and cost-sensitive customers, these companies can expand their market presence and progressively enhance their economic returns. Nonetheless, this approach introduces considerable operational complexities, particularly regarding sales formats, necessitating more sophisticated business considerations.

When implementing product distribution through e-commerce platforms, suppliers typically operate under two contractual arrangements: the reselling mode and agency mode. These frameworks demonstrate fundamental operational divergences. Under the reselling mode, suppliers transfer goods in bulk to e-commerce platforms, which then distribute items to end consumers after applying commercial markups. In the agency mode, suppliers retain full autonomy over retail pricing and consumer transactions, whereas the platform merely obtains a predetermined sales percentage as service remuneration, entirely removed from price determination processes. This structural dichotomy critically impacts suppliers’ strategic control over market positioning and revenue segmentation within digital distribution ecosystems [4,5]. Several questions arise: For suppliers, should suppliers employ different sales modes for different products? If suppliers decide to use the same sales mode for all products, What kind of sales mode should they adopt? And if they choose to implement different sales modes for different products, how should they make decisions based on product characteristic, market positioning, and target consumer segments?This paper investigates these questions by examining a supplier that sells both MIG and NG products via an e-commerce platforms.

Although substantial scholarly investigations have addressed suppliers’ strategic selection of sales modes through e- commerce platforms, the predominant focus remains confined to individual product scenarios. The individual product perspective is difficult to explain the complex trade-offs of suppliers in managing multiple types of products, and this assumption also fails to account for the reality that many suppliers sell both MIG and NG products simultaneously. Specifically, ref. [6] individual product model cannot explain how the differences between multi-product in terms of cost structure, consumer preferences, and environmental attributes affect the selection of the optimal model. As noted previously, to cater to the preferences of both environmentally conscious and cost-sensitive consumers, suppliers frequently implement multi-product sales strategy. Moreover, studies focusing on green operation decisions, for example, ref. [7] emphasizing the core role of green costs and consumer preferences, but generally ignore the key downstream decision variable of the sales mode selection of e-commerce platforms. Consequently, the aforementioned studies provide insufficient guidance for mode selection in a dual product environment. To bridge this academic void, our study constructs a strategic decision-making model grounded in game theory, analyzing how suppliers optimize distribution approaches for MIG and NG product within e-commerce platforms. Our analysis considers various factors that affect decision-making, including commission rate, cost of MIG products, and consumer acceptance of NG products. Furthermore, we examine the disparities in mode preferences between these enterprises. The research reveals significant differences in mode selection preferences between suppliers and e-commerce platforms. The wholesale price invasion effect remains pronounced even when both product types are sold simultaneously. Notably, we found that selecting distinct sales modes for MIG and NG products can effectively mitigate the double marginalization issue. This study not only adds new insights to the fields of platform economy and mode strategy but also provides valuable practical guidance for suppliers and policymakers.

In the following sections, we review the relevant literature on the mode selection problem for suppliers (Section 2), propose a theoretical model and outline the hypotheses (Section 3), solve the model and obtain analytical solutions under different sales modes (Section 4), analyze results and discuss the mode preferences of suppliers and e-commerce platforms (Section 5), and conclude by summarizing research findings and limitations, and propose potential directions for future research (Section 6).

2. Literature Review

There is substantial literature that touches upon some of the issues related to this paper here. We can classify the relevant literature into two categories: (i) works that examine mode selection, (ii) works that product differentiation.

2.1. Mode Selection

Current research on sales modes can be broadly categorized into three types: direct sales mode, traditional retail mode, and agency mode. In direct sales modes, suppliers set the retail price and sell directly to consumers. In traditional retail modes, also known as reselling modes, suppliers sell products to retailers at wholesale prices, and the retailers then determine the retail price and sell to consumers. In agency modes, suppliers set the retail price but sell through retailers, paying them a commission.

Our research is highly associated with the in-depth exploration of mode strategies, with many existing works concentrating on the issue of dual mode selection. For example, ref. [8] and other scholars have deeply analyzed the choice of suppliers between direct sales and reselling modes, exploring whether to adopt a single mode or dual modes to distribute their products. On this basis, subsequent studies have further integrated multiple factors, such as free-riding of consumers, the logistics sharing mechanism, and consumers’ preferences in the modes, enriching the analytical framework [9,10,11]. With the booming economy, e-commerce platforms have emerged as an indispensable link in the supply chain, prompting many suppliers to turn to agency modes to expand the market. In this context, some researchers have begun to explore mode selection issues closely associated with agency modes. For example, ref. [12] explored how retailers should make strategic choices between agency modes and reselling modes in a complex supply chain environment involving a single supplier and two competing retailers. Many scholars have subsequently analyzed the profound impact of factors such as marketing efforts and information sharing on mode selection strategies [13,14,15]. In addition to the widely discussed dual-mode selection, some scholars have expanded their research scope to a more complex three-mode environment, that is, the coexistence of direct modes, agency modes, and reselling modes. These studies have analyzed from multiple angles and levels how supply chain members should formulate efficient product distribution strategies to cope with the increasingly fierce market competition when there are three sales modes in the market at the same time [16,17,18]. While previous research on mode selection explores this issue for suppliers, it typically assumes that suppliers sell only one type of product. Our study, however, examines mode selection when suppliers sell two types of products, thereby relaxing this assumption.

2.2. Sales of Differentiated Products

Because our study considers the situation where suppliers sell two products at the same time, the existing research on the sale of multiple products is highly relevant to this study. Among these literatures, some literatures consider the sales of differentiated products, but only consider one sales mode, and ignore the in-depth analysis of the supplier mode selection problem. For example, ref. [1] compares the product differentiation strategies of suppliers in decentralized supply chains and centralized supply chains in detail, focusing on the sales of marginal cost-intensive and development cost-intensive products, and deeply evaluates the specific impact of product type on the effectiveness of the strategy. In addition, there are many comprehensive and in-depth analyses of supplier sales strategies from multiple dimensions such as product cost and quality differences and consumer preferences [19,20,21,22,23], but these studies only consider one sales mode and ignore the situation where suppliers can adopt different sales modes. Some other studies consider the sales of different products using different modes. For example, ref. [24] conducted an in-depth study on the two-stage pricing problem and product differentiation strategy when suppliers sell two products. They considered the reselling and direct sales modes in their study, but this study mainly focused on the impact of decentralized and centralized decision-making rather than discussing the choice of sales mode. Ref. [2] considered the same market structure, but their research focused on the impact of changes in government policies on the pricing strategy of suppliers when selling different products. In addition, some studies have discussed how changes in factors such as market size, market power structure, and consumer preferences affect suppliers’ pricing strategies [25,26,27,28,29].

Although these studies are closely related to our research in terms of topic, they are significantly different from our research. Specifically, ref. [29] also involves the reselling mode and the agency mode, but its research is based on the assumption that the supplier only sells one product; while our research considers the actual situation that the supplier sells two interchangeable products at the same time. On the other hand, although ref. [27] also examined the situation where suppliers sell two products at the same time, their research focused on the combination strategy of direct sales mode and reselling mode, and assumed that the two products must be sold through different modes; our research focuses on the combination of agency mode and reselling mode. At the same time, we also explored the possibility of selling two products through the same mode and deeply analyzed the impact of consumer product preferences on strategy.

3. Model Framework

This research analyzes an exclusive market framework involving bilateral interactions between a single supplier and an e-commerce platform, a configuration frequently adopted in industrial organization studies. The supplier supplies both NG products and modified substitute MIG goods, utilizing the platform’s dual operational configurations—agency mode or reselling mode—for market expansion. Under agency mode arrangements, the e-commerce platform operates as a transactional intermediary linking production entities with end-users, while full pricing autonomy resides with the supplier. Platform compensation is derived from predetermined commission rates r ($0<r<1$), which this study treats as exogenous parameters following established academic conventions [15]. In reselling mode operations, the platform functions as an independent distributor, purchasing goods through wholesale agreements before implementing autonomous pricing strategies for consumer sales. We use $c_{MIG}$ and $c_{NG}$ to denote the marginal production costs of MIG products and NG products, respectively. Considering that MIG products tend to have stricter raw material requirements, we assume $c_{NG}<c_{MIG}$. For convenience in calculation and without loss of generality, we set $c_{NG}=0,c_{MIG}=c>0$. We assume that the operational costs are all fixed and equal to zero for the sake of convenience and without loss of generality [15,30]. In the calculation, to ensure that the profit function is strictly convex and the associated prices and demands are positive, we derive a range of values for c$(c\in (0,(1-r)·(1-\theta ))$. Let v represent the perceived value of MIG products among consumers, assumed to uniformly distribute within the $(0,1)$. For NG products, the perceived value is given as $\theta ·v$, where $0<\theta <1$, an increase in $\theta$ signifies a higher level of consumer acceptance for NG products. Utilizing relevant literature, we formulate the consumer demand function based on their utility. Consumers derive utility $U_{MIG}=v-p_{MIG}$ from purchasing MIG products, and $U_{NG}=\theta ·v-p_{NG}$ from NG products. Purchase behavior only occurs when the consumer utility is positive. The point where the utility of purchasing versus not purchasing MIG products intersects is at $v_{MIG}=p_{MIG}$. Similarly, for NG products, this intersection point is $v_{NG}={\displaystyle \frac{p_{NG}}{\theta }}$. Lastly, the equilibrium point between the utilities of buying MIG and NG products is found at $v={\displaystyle \frac{p_{MIG}-p_{NG}}{1-\theta }}-{\displaystyle \frac{p_{NG}}{\theta }}$. Referring to the relevant literature [31,32,33], we can infer the demand for both MIG and NG products as follows:

$$\begin{array}{cc}\hfill D_{MIG}^{XY}& =1-{\displaystyle \frac{p_{MIG}-p_{NG}}{1-\theta }}\hfill \end{array}$$

(1)

$$\begin{array}{cc}\hfill D_{NG}^{XY}& ={\displaystyle \frac{p_{MIG}-p_{NG}}{1-\theta }}-{\displaystyle \frac{p_{NG}}{\theta }}\hfill \end{array}$$

(2)

where the subscript MIG and NG, respectively represent MIG products and NG products. Superscripts X and Y indicate the type of mode utilized when selling products-X for NG products and Y for MIG products—both operating between the supplier and the platform. $X\&Y=A,R$. Specifically, A signifies that the product sale is conducted through an agency mode, while R indicates a reselling mode (Parameters and variables see

Table 1. Notations.

Notation	Definition
Parameters
v	Consumers’ perceived value of the MIG product
$\theta$	Consumer acceptance of NG products
c	Marginal cost of production of MIG products
r	The commission rate
Decision variables
$w_{MIG}$	Wholesale price of MIG products
$w_{NG}$	Wholesale price of NG products
$p_{MIG}$	Market retail prices for MIG products
$p_{NG}$	Market retail prices for NG products
Objective
${\Pi }_M$	Profit of the supplier
${\Pi }_P$	Profits of the platform

).

Since the supplier offers both MIG and NG products through two distinct modes on platforms, we explore four scenarios. AA scenario: NG and MIG products are offered via the agency mode. AR scenario: NG products use the agency mode, while MIG products adopt the reselling mode. RA scenario: NG products are assigned to the reselling mode, and MIG products utilize the agency mode. RR scenario: NG and MIG products are distributed via the reselling mode. A Stackelberg game model is employed, positioning the supplier as the leader and the online platform as the follower. Initially, the supplier sets the retail prices in the agency mode and wholesale prices in the reselling mode, which is followed by the platform setting retail prices in the reselling mode.

4. Model Analysis

In Section 4, we present the objective functions of suppliers and e-commerce platforms when they adopt different strategies, as well as the corresponding optimal decisions.

AR: Distributing NG products via the agency mode, while MIG products are offered via the reselling mode. The objective functions of suppliers and e-commerce platforms are as follows:

$$\begin{array}{cc}\hfill {\Pi }_M& =(w_{MIG}-c)·D_{MIG}+(1-r)·p_{NG}·D_{NG}\hfill \end{array}$$

(3)

$$\begin{array}{cc}\hfill {\Pi }_P& =(p_{MIG}-w_{MIG})·D_{MIG}+r·p_{NG}·D_{NG}\hfill \end{array}$$

(4)

The best solutions are detailed below:

$$\begin{array}{cc}\hfill w_{MIG}^{AR}& ={\displaystyle \frac{-r\theta +c+1}{2}}\hfill \end{array}$$

(5)

$$\begin{array}{cc}\hfill p_{MIG}^{AR}& ={\displaystyle \frac{-\theta +c+3}{4}}\hfill \end{array}$$

(6)

$$\begin{array}{cc}\hfill p_{NG}^{AR}& ={\displaystyle \frac{\theta }{2}}\hfill \end{array}$$

(7)

The profits of of suppliers and e-commerce platforms are detailed below:

$$\begin{array}{cc}\hfill {\Pi }_M^{AR}& ={\displaystyle \frac{\left(-2r+1\right){\theta }^2+\left(-2c+2r\right)\theta -{\left(-1+c\right)}^2}{-8+8\theta }}\hfill \end{array}$$

(8)

$$\begin{array}{cc}\hfill {\Pi }_P^{AR}& ={\displaystyle \frac{\left(4r-1\right){\theta }^2+\left(-2c-4r+2\right)\theta -{\left(-1+c\right)}^2}{-16+16\theta }}\hfill \end{array}$$

(9)

RA: Distributing NG products via the reselling mode, while MIG products are offered via the agency mode. The objective functions of suppliers and e-commerce platforms are as follows:

$$\begin{array}{cc}\hfill {\Pi }_M& =w_{NG}·D_{NG}+(1-r)·p_{MIG}·D_{MIG}-c·D_{MIG}\hfill \end{array}$$

(10)

$$\begin{array}{cc}\hfill {\Pi }_P& =(p_{NG}-w_{NG})·D_{NG}+r·p_{MIG}·D_{MIG}\hfill \end{array}$$

(11)

The best solutions are detailed below:

$$\begin{array}{cc}\hfill w_{NG}^{RA}& =-{\displaystyle \frac{\theta \left(-1+r\right)}{2}}\hfill \end{array}$$

(12)

$$\begin{array}{cc}\hfill p_{NG}^{RA}& =-{\displaystyle \frac{\left(\left(c-2\right)r+c+2\right)\theta }{-4+4r}}\hfill \end{array}$$

(13)

$$\begin{array}{cc}\hfill p_{MIG}^{RA}& =-{\displaystyle \frac{c-r+1}{-2+2r}}\hfill \end{array}$$

(14)

The profits of of suppliers and e-commerce platforms are detailed below:

$$\begin{array}{cc}\hfill {\Pi }_M^{RA}& ={\displaystyle \frac{\left(-2r^2+\left(-c^2-4c+4\right)r-c^2+4c-2\right)\theta +2{\left(c+r-1\right)}^2}{8\left(-1+r\right)\left(-1+\theta \right)}}\hfill \end{array}$$

(15)

$$\begin{array}{cc}\hfill {\Pi }_P^{RA}& ={\displaystyle \frac{\left(-4+4\theta \right)r^3+\left(-c^2\theta -8\theta +8\right)r^2+\left(-2c^2\theta +4c^2+4\theta -4\right)r-c^2\theta }{16{\left(-1+r\right)}^2\left(-1+\theta \right)}}\hfill \end{array}$$

(16)

AA: NG and MIG products are offered via the agency mode. The objective functions of suppliers and e-commerce platforms are as follows:

$$\begin{array}{cc}\hfill {\Pi }_M& =(1-r)·(p_{MIG}·D_{MIG}+p_{NG}·D_{NG})-c·D_{MIG}\hfill \end{array}$$

(17)

$$\begin{array}{cc}\hfill {\Pi }_P& =r·(p_{MIG}·D_{MIG}+p_{NG}·D_{NG})\hfill \end{array}$$

(18)

The best solutions are detailed below:

$$\begin{array}{cc}\hfill p_{MIG}^{AA}& =-{\displaystyle \frac{c-r+1}{-2+2r}}\hfill \end{array}$$

(19)

$$\begin{array}{cc}\hfill p_{NG}^{AA}& ={\displaystyle \frac{\theta }{2}}\hfill \end{array}$$

(20)

The profits of of suppliers and e-commerce platforms are detailed below:

$$\begin{array}{cc}\hfill {\Pi }_M^{AA}& ={\displaystyle \frac{\left(1-\theta \right)r^2-2\left(-1+\theta \right)\left(-1+c\right)r+\left(2c-1\right)\theta +{\left(-1+c\right)}^2}{4\left(-1+r\right)\left(-1+\theta \right)}}\hfill \end{array}$$

(21)

$$\begin{array}{cc}\hfill {\Pi }_P^{AA}& ={\displaystyle \frac{r\left(\left(-1+\theta \right)r^2+\left(-2\theta +2\right)r+c^2+\theta -1\right)}{4{\left(-1+r\right)}^2\left(-1+\theta \right)}}\hfill \end{array}$$

(22)

RR: NG and MIG products are offered via the reselling mode. The objective functions of suppliers and e-commerce platforms are as follows:

$$\begin{array}{cc}\hfill {\Pi }_M& =(w_{MIG}-c)·D_{MIG}+w_{NG}·D_{NG}\hfill \end{array}$$

(23)

$$\begin{array}{cc}\hfill {\Pi }_P& =(p_{MIG}-w_{MIG})·D_{MIG}+(p_{NG}-w_{NG})·D_{NG}\hfill \end{array}$$

(24)

The best solutions are detailed below:

$$\begin{array}{cc}\hfill w_{MIG}^{RR}& ={\displaystyle \frac{c+1}{2}}\hfill \end{array}$$

(25)

$$\begin{array}{cc}\hfill w_{NG}^{RR}& ={\displaystyle \frac{\theta }{2}}\hfill \end{array}$$

(26)

$$\begin{array}{cc}\hfill p_{MIG}^{RR}& ={\displaystyle \frac{c+3}{4}}\hfill \end{array}$$

(27)

$$\begin{array}{cc}\hfill p_{NG}^{RR}& ={\displaystyle \frac{3\theta }{4}}\hfill \end{array}$$

(28)

The profits of of suppliers and e-commerce platforms are detailed below:

$$\begin{array}{c}\hfill {\Pi }_M^{RR}={\displaystyle \frac{\left(-2c+1\right)\theta -{\left(c-1\right)}^2}{-8+8\theta }}\end{array}$$

(29)

$$\begin{array}{c}\hfill {\Pi }_P^{RR}={\displaystyle \frac{\left(-2c+1\right)\theta -{\left(c-1\right)}^2}{-16+16\theta }}\end{array}$$

(30)

5. Analysis and Result

5.1. Comparison of Price

Lemma 1.

(a) $w_{MIG}^{AR}<w_{MIG}^{RR}$, $w_{NG}^{RA}<w_{NG}^{RR}$; (b) ${\displaystyle \frac{\partial w_{MIG}^{AR}}{\partial r}}<0$, ${\displaystyle \frac{\partial w_{NG}^{RA}}{\partial r}}<0$.

Lemma 1 suggests that when MIG and NG products are originally distributed via the reselling mode, transitioning one of them to the agency mode results in a decline in the wholesale price of the other product. This decline becomes more pronounced as the commission rates imposed by the e-commerce platform increase. The underlying reason for this effect is that selling through the same mode entails purely vertical competition, whereas employing distinct modes introduces both vertical and horizontal competitive dynamics. In such cases, suppliers are likely to adopt strategic pricing adjustments to encourage e-commerce platform partners to uphold their procurement agreements.

Lemma 2.

(a) $p_{MIG}^{AA}=p_{MIG}^{RA}$, $p_{NG}^{AA}=p_{NG}^{AR}$; (b) $p_{MIG}^{AR}<p_{MIG}^{RR}$; (c) If (1) $\theta >{\displaystyle \frac{r}{1+r}}$ or (2) $\theta <{\displaystyle \frac{r}{1+r}}$ and $c\in (0,{\displaystyle \frac{1-r}{1+r}})$, then $p_{NG}^{RA}<p_{NG}^{RR}$, otherwise $p_{NG}^{RA}>p_{NG}^{RR}$.

Part (a) of Lemma 2 indicates that the pricing for MIG products remains stable when distributed through the agency mode, regardless of the strategy utilized for NG products; conversely, this principle is also applicable in reverse situations. In Parts (b) and (c) of Lemma 2, it is demonstrated that if MIG and NG products are initially marketed through the reselling mode, transferring NG products to the agency mode will lead to a decrease in the price of MIG products. Furthermore, shifting MIG products to the agency mode may reduce the price of NG products, particularly when consumer acceptance of NG products is high, or when both acceptance levels and the associated costs of MIG products are relatively low. These results imply that utilizing different modes for MIG and NG products may alleviate the challenges presented by double marginalization.

The strategy evolution of supplier L on the Jingdong platform (JD) clearly shows the sales mode differentiation practice of multi-product suppliers. Supplier L is one of the leading sports brand enterprises in China. In 2017, JD established “L JD Self-operated Flagship Store”, which adopts the reselling mode, in which JD directly purchases products from supplier L and sells them to consumers at a markup. In 2018, Supplier L launched the “L2” series of products. Although this series is similar to traditional products of supplier L in function, such as sportswear and sports shoes, there is a difference in product positioning. The “L2” series adopts more exquisite packaging, a more novel style, and pattern design. Initially, the “L2” series was also sold through JD self-operated flagship stores. In 2019, supplier L separately established the “L2 JD Official Flagship Store”, which adopts the agency mode, operated by supplier L for the “L2” series. Since then, the selling price of the same type of products in the L JD self-operated flagship store has decreased. This phenomenon verifies the theoretical expectation in this study that contractual arrangements differentiation can alleviate the problem of double marginalization.

5.2. Comparison of Supplier’s Profit

Corollary 1.

${\Pi }_M^{\mathit{AA}}>{\Pi }_M^{\mathit{RA}}$.

Figure 1 shows the preferences of suppliers for AA strategy and RA strategy under different conditions.

Corollary 1 demonstrates that suppliers consistently achieve superior profitability through dual agency mode deployment (AA strategy) for NG and MIG products, compared to hybrid mode adoption combining reselling mode for NG with agency mode for MIG (RA strategy). This outcome may seem counterintuitive, as one might expect the RA strategy to become more advantageous under high commission rates due to lower fee expenses. However, the opposite holds true. A key explanation lies in the AA strategy’s ability to generate greater sales volumes for NG products, which enhances overall supplier profitability. These insights conclusively imply that suppliers would never choose the RA strategy, as a more profitable alternative always exists.

Corollary 2.

If (1) $r>{\displaystyle \frac{1}{2}}$ or (2) $r<{\displaystyle \frac{1}{2}}$ and $c\in ({\displaystyle \frac{\left(1-r-r\sqrt{2(1-r)}\right)\left(1-\theta \right)}{1+r}},(1-r)·(1-\theta ))$, then ${\Pi }_M^{\mathit{AA}}<{\Pi }_M^{\mathit{AR}}$, otherwise ${\Pi }_M^{\mathit{AA}}>{\Pi }_M^{\mathit{AR}}$.

Figure 2 shows the preferences of suppliers for AA strategy and AR strategy under different conditions.

Corollary 2 shows that that when NG products are allocated to the agency mode, suppliers attain higher profitability by distributing MIG products through the reselling mode under two conditions: high commission rates or low commission rates combined with relatively high production costs for MIG products. Conversely, when these parametric thresholds are unmet, assigning MIG products to the agency mode generates superior returns. Notably, high production costs (c) correlate with diminished sales volumes of MIG products under the AA strategy, which directly contributes to reduced supplier profitability in such scenarios. This interplay between cost parameters and sales performance highlights the AA strategy’s limitations in high-cost of MIG products environments.

Corollary 3.

There is $c_{1,2}={\displaystyle \frac{r\theta \pm \sqrt{\left(-1+\theta \right)\left(-1+r\right)\left(2r^2+r\theta -\theta \right)}-r-\theta +1}{r+1}}$ such that ${\Pi }_M^{AA}={\Pi }_M^{RR}$. If (1) $r>{\displaystyle \frac{1}{2}}$ or (2) $-{\displaystyle \frac{\theta }{4}}+{\displaystyle \frac{\sqrt{{\theta }^2+8\theta }}{4}}<r<-{\displaystyle \frac{-\theta +\sqrt{\theta }}{\theta -1}}$ and $c\in (c_1,c_2)$ or (3) ${\displaystyle \frac{-\theta +\sqrt{\theta }}{\theta -1}}<r<{\displaystyle \frac{1}{2}}$ and $c\in (c_1,(1-\theta )·(1-r))$, then ${\Pi }_M^{AA}<{\Pi }_M^{RR}$; and ${\Pi }_M^{AA}>{\Pi }_M^{RR}$ otherwise.

Figure 3 shows the preferences of suppliers for AA strategy and RR strategy under different conditions.

Corollary 3 suggests: (1) Higher commission rates prompt suppliers to adopt the reselling mode for both NG and MIG products, achieving maximum profitability; (2) Lower commission rates enhance profitability when both products utilize the agency mode; (3) Under exceptionally high commission rates combined with moderate MIG production costs, employing the agency mode for both products increases supplier profits, whereas the reselling mode generally yields superior profitability across other conditions.

Corollary 4.

If $r>{\displaystyle \frac{1}{2}}$, then ${\Pi }_M^{AR}<{\Pi }_M^{RR}$, and ${\Pi }_M^{AR}>{\Pi }_M^{RR}$ otherwise.

Figure 4 shows the preferences of suppliers for AR strategy and RR strategy under different conditions.

Corollary 4 establishes that high commission rates drive suppliers to make more profits by distributing NG and MIG products through the reselling mode. When commission rates fall below specific thresholds, adopting a hybrid strategy—NG products allocated to the agency mode and MIG products to the reselling mode—yields higher profitability. Notably, when NG products are marketed through the agency mode with high commission rates, suppliers incur substantially higher platform fees, which may account for the observed profit relationship ${\Pi }_M^{AR}<{\Pi }_M^{RR}$.

Proposition 1.

The supplier’s preferences on sale modes: (1) When $r>{\displaystyle \frac{1}{2}}$, the supplier will sell NG and MIG products with the reselling mode; (2) when $r<{\displaystyle \frac{1}{2}}$ and $c\in (0,{\displaystyle \frac{\left(1-r-r\sqrt{2(1-r)}\right)\left(1-\theta \right)}{1+r}})$, the supplier will opt to distribute both NG and MIG products with the agency mode; (3) when $r<{\displaystyle \frac{1}{2}}$ and $c\in ({\displaystyle \frac{\left(1-r-r\sqrt{2(1-r)}\right)\left(1-\theta \right)}{1+r}},(1-r)·(1-\theta ))$, the supplier opts to distribute NG products via the agency mode, while promoting MIG products with the reselling mode.

Figure 5 shows the final sales mode preferences of suppliers under different conditions.

Based on the comparative analysis presented in Corollary 1–4, we determined that the supplier’s choice of sales modes for NG and MIG products is limited to either AA, AR, or RR. This is because the RA strategy consistently proves suboptimal for the supplier. We further introduced the supplier’s strategic preferences, as delineated in Proposition 1. Our findings suggest that a high commission rate encourages the supplier to sell NG and MIG products via the reselling mode preferentially. Conversely, at a low commission rate, if the cost of MIG products is minimal, selling both NG and MIG products via the agency mode yields greater profitability for the supplier. However, suppose the costs of MIG products are significant. In that case, it is more profitable for the supplier to sell NG products via the agency mode and MIG products via the reselling mode.

5.3. Comparison of Platform’s Profit

Corollary 5.

${\Pi }_P^{RA}>{\Pi }_P^{AA}$.

Figure 6 shows the preferences of platforms for RA strategy and AA strategy under different conditions.

Corollary 5 reveals a counterintuitive phenomenon: e-Commerce platforms systematically achieve greater profitability by allocating NG products to the reselling mode and MIG products to the agency mode, rather than adopting dual agency mode strategies. While elevated commission rates might suggest that assigning both products to the agency mode (AA) would amplify platform earnings through heightened fee revenues, analytical findings demonstrate that AA strategy allocations result in diminished sales volumes and reduced retail prices for NG products relative to RA strategy configurations. This constrained performance under the AA strategy limits commission gains from NG products, thus validating the consistent profitability superiority of the RA strategy. Consequently, the commissions earned from NG products under AA are not substantial, which could explain this counterintuitive result. Thus, platforms will never opt for selling both product types through the agency mode, as there is always a more profitable strategy available.

Corollary 6.

There is $c_{1,2}={\displaystyle \frac{\left(-1+r\pm \sqrt{-4r^3\theta +4r^3-7r^2\theta +8r^2-2r\theta +\theta }\right)\left(-1+r\right)}{r^2+2r+1}}$ such that ${\Pi }_P^{AR}={\Pi }_P^{RA}$. (a). In Region A ($0<r<{\displaystyle \frac{1}{4}}$, $0<\theta <{\displaystyle \frac{1}{25}}$). If (1) $r<{\displaystyle \frac{\theta +\sqrt{\theta }}{1-\theta }}$ and $c\in (c_1,(1-r)·(1-\theta ))$, (2) $r>{\displaystyle \frac{\theta +\sqrt{\theta }}{1-\theta }}$ and $c\in (c_1,c_2)$, then ${\Pi }_P^{RA}>{\Pi }_P^{AR}$, otherwise ${\Pi }_P^{RA}<{\Pi }_P^{AR}$. (b). In Region B ($0<r<{\displaystyle \frac{1}{4}}$, ${\displaystyle \frac{1}{25}}<\theta <1$). If $c\in (c_1,(1-r)·(1-\theta ))$, then ${\Pi }_P^{RA}>{\Pi }_P^{AR}$, otherwise ${\Pi }_P^{RA}<{\Pi }_P^{AR}$. (c). In Region C (${\displaystyle \frac{1}{4}}<r<1$, $0<\theta <{\displaystyle \frac{1}{25}}$). If $c\in (0,c_2)$, then ${\Pi }_P^{RA}>{\Pi }_P^{AR}$, otherwise ${\Pi }_P^{RA}<{\Pi }_P^{AR}$. (d). In Region D (${\displaystyle \frac{1}{4}}<r<1$, ${\displaystyle \frac{1}{25}}<\theta <{\displaystyle \frac{1}{4}}$). If (1) $r<{\displaystyle \frac{\theta +\sqrt{\theta }}{1-\theta }}$, (2) $r>{\displaystyle \frac{\theta +\sqrt{\theta }}{1-\theta }}$ and $c\in (0,c_2)$, then ${\Pi }_P^{RA}>{\Pi }_P^{AR}$, otherwise ${\Pi }_P^{RA}<{\Pi }_P^{AR}$. (e). In Region E (${\displaystyle \frac{1}{4}}<r<1$, ${\displaystyle \frac{1}{4}}<\theta <1$), ${\Pi }_P^{RA}>{\Pi }_P^{AR}$ (see Figure 7) (In Figure 7, the values of r and θ are chosen because the results of Corollary 6 are very complicated. We hope to show the results of each part to you through diagrams.Values f r and θ in Figure 8 are the same reason).

Figure 7 shows the preferences of platforms for RA strategy and AR strategy under different conditions.

Corollary 6 delineates platform profitability dynamics under varying parametric conditions: (1) When both commission rates and consumer acceptance of NG products are low, platforms maximize gains by allocating NG products to the agency mode and MIG products to the reselling mode, provided MIG production costs remain comparatively low; conversely, elevated MIG costs necessitate reversed allocations (NG to reselling mode, MIG to agency mode). (2) Under high commission rates with low NG acceptance, platforms favor agency mode allocations for NG products and reselling mode distributions for MIG products when MIG costs are extremely high, while moderate MIG costs justify inverted allocations. (3) When both commission rates and NG acceptance are high, platforms consistently attain superior profitability by assigning NG products to the reselling mode and MIG products to the agency mode, regardless of MIG cost levels.

Corollary 7.

There is $c_{1,2}={\displaystyle \frac{\left(r-1\pm 2\sqrt{r^2\left(r+2\right)}\right)\left(-1+r\right)}{{\left(r+1\right)}^2}}$ such that ${\Pi }_P^{RA}={\Pi }_P^{RR}$. (a). In Region A ($0<r<{\displaystyle \frac{1}{4}}$, $0<\theta <{\displaystyle \frac{1}{25}}$). If (1) ${\displaystyle \frac{\theta ·(3-\theta -2\sqrt{2-\theta })}{{(1-\theta )}^2}}<r<{\displaystyle \frac{\theta ·(3-\theta +2\sqrt{2-\theta })}{{(1-\theta )}^2}}$ and $c\in (c_1,(1-r)·(1-\theta ))$, (2) $r>{\displaystyle \frac{\theta ·(3-\theta +2\sqrt{2-\theta })}{{(1-\theta )}^2}}$ and $c\in (c_1,c_2)$, then ${\Pi }_P^{RA}>{\Pi }_P^{RR}$, otherwise ${\Pi }_P^{RA}<{\Pi }_P^{RR}$. (b). In Region B ($0<r<{\displaystyle \frac{1}{4}}$, ${\displaystyle \frac{1}{25}}<\theta <1$). If $r>{\displaystyle \frac{\theta ·(3-\theta -2\sqrt{2-\theta })}{{(1-\theta )}^2}}$ and $c\in (c_1,(1-r)·(1-\theta ))$, then ${\Pi }_P^{RA}>{\Pi }_P^{RR}$, otherwise ${\Pi }_P^{RA}<{\Pi }_P^{RR}$. (c). In Region C (${\displaystyle \frac{1}{4}}<r<1$, $0<\theta <{\displaystyle \frac{1}{25}}$). If $c\in (0,c_2)$, then ${\Pi }_P^{RA}>{\Pi }_P^{RR}$, otherwise ${\Pi }_P^{RA}<{\Pi }_P^{RR}$. (d). In Region D (${\displaystyle \frac{1}{4}}<r<1$, ${\displaystyle \frac{1}{25}}<\theta <1-{\displaystyle \frac{\sqrt{3}}{2}}$). If (1) $r<{\displaystyle \frac{\theta ·(3-\theta +2\sqrt{2-\theta })}{{(1-\theta )}^2}}$, (2) $r>{\displaystyle \frac{\theta ·(3-\theta +2\sqrt{2-\theta })}{{(1-\theta )}^2}}$ and $c\in (0,c_2)$, then ${\Pi }_P^{RA}>{\Pi }_P^{RR}$, otherwise ${\Pi }_P^{RA}<{\Pi }_P^{RR}$. (e). In Region E (${\displaystyle \frac{1}{4}}<r<1$, $1-{\displaystyle \frac{\sqrt{3}}{2}}<\theta <1$), ${\Pi }_P^{RA}>{\Pi }_P^{RR}$ (see Figure 8).

Figure 8 shows the preferences of platforms for RA strategy and RR strategy under different conditions.

Corollary 7 indicates that platform strategies to maximize profits under varying conditions of commission rates, consumer acceptance of NG products, and production costs of MIG products. (1) When commission rates are low and consumer acceptance of NG products is limited, platforms achieve optimal returns by allocating NG products to the reselling mode and MIG products to the agency mode if production costs of MIG products are high; conversely, lower production costs of MIG products favor assigning both NG products and MIG products to the reselling mode. (2) Under low commission rates paired with high consumer acceptance of NG products, platforms universally prioritize reselling mode allocations for both NG products and MIG products, regardless of production costs of MIG products. (3) For moderate commission rates and high consumer acceptance of NG products, NG products remain consistently assigned to the reselling mode, while MIG products adopt the reselling mode unless their production costs exceed critical thresholds, necessitating a shift to the agency mode. (4) When commission rates are high but consumer acceptance of NG products is relatively low, exceptionally high production costs of MIG products justify dual reselling mode allocations for both NG products and MIG products, whereas moderately bounded production costs of MIG products favor a hybrid approach with NG products in the reselling mode and MIG products in the agency mode. (5) Finally, under high commission rates and strong consumer acceptance of NG products, platforms consistently maximize profits by allocating both NG products and MIG products to the reselling mode, demonstrating the cost-agnostic dominance of this strategy.

Corollary 8.

If $r<{\displaystyle \frac{1}{4}}$, then ${\Pi }_P^{AR}<{\Pi }_P^{RR}$, otherwise ${\Pi }_P^{AR}>{\Pi }_P^{RR}$.

Figure 9 shows the preferences of platforms for AR strategy and RR strategy under different conditions.

Corollary 8 demonstrates that platforms maximize profits through distinct distribution strategies depending on commission rate levels. Under low commission rates, allocating both NG and MIG products to the reselling mode yields higher profitability. Conversely, when commission rates are high, platforms achieve optimal returns by assigning NG products to the agency mode while retaining MIG products in the reselling mode. This strategic shift balances cost efficiency and revenue optimization under varying fee structures.

Proposition 2.

The platform’s preferences on sales mode: (a). When $0<r<{\displaystyle \frac{1}{4}}$ and $0<\theta <{\displaystyle \frac{1}{25}}$, if (1) ${\displaystyle \frac{\theta ·(3-\theta -2\sqrt{2-\theta })}{{(1-\theta )}^2}}<r<{\displaystyle \frac{\theta ·(3-\theta +2\sqrt{2-\theta })}{{(1-\theta )}^2}}$ and $c\in ({\displaystyle \frac{\left(-1+r+2\sqrt{r^3+2r^2}\right)\left(-1+r\right)}{r^2+2r+1}},(1-r)·(1-\theta ))$, (2) $r>{\displaystyle \frac{\theta ·(3-\theta +2\sqrt{2-\theta })}{{(1-\theta )}^2}}$ and $c\in ({\displaystyle \frac{\left(-1+r+2\sqrt{r^3+2r^2}\right)\left(-1+r\right)}{r^2+2r+1}},-{\displaystyle \frac{\left(1-r+2\sqrt{r^3+2r^2}\right)\left(-1+r\right)}{r^2+2r+1}})$, the platform prioritizes allocating NG products to the reselling mode and MIG products to the agency mode. In alternative scenarios, it opts to assign both products to the reselling mode. (b). When $0<r<{\displaystyle \frac{1}{4}}$ and ${\displaystyle \frac{1}{25}}<\theta <1$, if $r>{\displaystyle \frac{\theta ·(3-\theta -2\sqrt{2-\theta })}{{(1-\theta )}^2}}$ and $c\in ({\displaystyle \frac{\left(-1+r+2\sqrt{r^3+2r^2}\right)\left(-1+r\right)}{r^2+2r+1}},(1-r)·(1-\theta ))$, the platform prioritizes allocating NG products to the reselling mode and MIG products to the agency mode. In alternative scenarios, it opts to assign both products to the reselling mode. (c). When ${\displaystyle \frac{1}{4}}<r<1$ and $0<\theta <{\displaystyle \frac{1}{25}}$, if $c\in (0,{\displaystyle \frac{\left(-1+r-\sqrt{\left(-4r^3-7r^2-2r+1\right)\theta +4r^3+8r^2}\right)\left(-1+r\right)}{{\left(r+1\right)}^2}})$, the platform prioritizes allocating NG products to the reselling mode and MIG products to the agency mode; otherwise, the modes are reversed. (d). When ${\displaystyle \frac{1}{4}}<r<1$, ${\displaystyle \frac{1}{25}}<\theta <{\displaystyle \frac{1}{4}}$, if (1) $r<{\displaystyle \frac{\theta +\sqrt{\theta }}{1-\theta }}$, (2) $r>{\displaystyle \frac{\theta +\sqrt{\theta }}{1-\theta }}$ and $c\in (0,{\displaystyle \frac{\left(-1+r-\sqrt{\left(-4r^3-7r^2-2r+1\right)\theta +4r^3+8r^2}\right)\left(-1+r\right)}{{\left(r+1\right)}^2}})$, the platform prioritizes allocating NG products to the reselling mode and MIG products to the agency mode; otherwise, the modes are reversed. (e). When ${\displaystyle \frac{1}{4}}<r<1$ and ${\displaystyle \frac{1}{4}}<\theta <1$, the platform always prioritizes allocating NG products to the reselling mode and MIG products to the agency mode.

Our comparative analysis, as delineated in Corollary 5–8, reveals that the platform’s sales strategies for NG and MIG products are constrained to either RA, AR, or RR. This is due to the consistently suboptimal nature of the AA strategy for the platform. We have subsequently proposed the platform’s strategic preferences, as articulated in Proposition 2. The findings suggest that the platform exhibits a preference for either the RA or RR strategy when r is relatively low. Conversely, when r is high, the platform demonstrates a preference for either the RA or AR strategy. Despite notable differences in mode preferences between suppliers and platforms, a comparison with Proposition 1 reveals a convergence of their preferences when r lies within a medium range (${\displaystyle \frac{1}{4}}<r<{\displaystyle \frac{1}{2}}$) and the cost of MIG products is elevated. In such scenarios, both parties prefer to distribute NG products through the agency mode and MIG products via the reselling mode.

6. Conclusions

This study develops a game mode where the supplier leads and the e-commerce platforms follows. It investigates the mode strategies between them when the supplier sells two types of products via the platform. We analyze which sales modes are optimal for either the supplier or the e-commerce platforms across various market conditions. Our research enriches the platform economics literature and offers theoretical insights valuable for policymakers involved in platform operations. Moreover, by examining how suppliers’ mode preferences evolve when they concurrently sell two types of products, our study challenges the conventional assumption that suppliers deal with only one product type. Consequently, our findings deepen understanding of suppliers’ decisions in selecting optimal modes.

Through our analysis of the game mode, we obtained several noteworthy findings. First, we found that when one product is sold under the reselling mode and the other under the agency mode, the wholesale price of the product is lower than when both products are sold under the reselling mode. This indicates that the intrusion effect of wholesale prices persists when selling two different products. Additionally, using two different sales modes for different products can mitigate double marginalization to some extent. For suppliers, they sell two substitutable products in the agency mode and the reselling mode respectively, use the pricing power of the agency mode to set the benchmark price, and suppress the price increase space; at the same time, by adjusting the wholesale price, they force retailers to converge the price increase range, thereby systematically weakening the double marginal effect in the environment of substitute competition. And strengthen channel diversion to transform price-sensitive demand into service premium demand. For the platform, the platform needs to strengthen the price anchoring effect in the agency mode through the traffic distribution mechanism, and at the same time mark the differentiated service value for the self-operated business, amplify the differentiated positioning of the product; promote mode complementarity rather than mutual exclusion, and achieve overall supply chain efficiency improvement.

Second, we examined the mode preferences of suppliers and e-commerce platforms. Our results show that suppliers’ mode preferences are mainly affected by the r and the cost of MIG products, whereas e-commerce platforms’ preferences are also affected by consumers’ acceptance of NG products. We also identified significant differences in mode preferences between suppliers and e-commerce platforms. Specifically, the RA strategy is the least favorable for suppliers, while the AA strategy is the least favorable for e-commerce platforms. Third, we observed that in most cases, the profits of both suppliers and e-commerce platforms increase as consumers’ acceptance of NG products grows. From an environmental standpoint, consumer acceptance of NG products can be seen as a reflection of their awareness of environmental protection. Enhancing this acceptance can potentially increase corporate profits. However, to maximize these profits, companies may choose not to highlight the differences between MIG and NG products to consumers, which is detrimental to environmental improvement. The government could implement preferential policies for MIG products (such as subsidies for suppliers) to encourage increased sales of these items. Simultaneously, promoting the unique attributes of MIG products (e.g., health benefits) can raise consumer environmental consciousness and stimulate demand for MIG products, thereby achieving environmental improvement objectives. Considering the research findings, we propose two recommendations. Firstly, relying solely on enterprises to promote MIG products is challenging. This is because selling a significant volume of MIG products can potentially decrease corporate profits. Consequently, governmental oversight and incentives are essential. Secondly, bolstering consumer environmental awareness is an effective strategy for fostering the MIG development of enterprises.

Of course, our study also has some limitations. First, we assume that each product can only be sold in one mode. However, in reality, in order to maximize the coverage of consumers, suppliers often adopt multiple modes to sell the same product at the same time. Therefore, future research can examine the supplier’s multi-product sales mode selection strategy when using multiple modes to sell the same product at the same time. Secondly, our research mainly focuses on the marginal cost differences between substitutable products, but the differences between some products are mainly reflected in the development costs. For example, electronic products such as mobile phones and computers usually require a lot of development investment to upgrade from the Nth generation to the N+1th generation. Subsequent research can explore how the technical development costs of products affect the optimal strategy of suppliers. Third, in our study, we assume that the operating costs and efficiency of suppliers and platforms in the agency mode and reselling mode are the same. However, in reality, due to various factors such as sales methods, after-sales service, logistics and transportation, and warehousing, there are certain differences in the operating costs and efficiency of the agency mode and the reselling mode. Therefore, in future research, it is possible to explore the preferences of suppliers and platforms for sales modes when different modes have different operating costs and efficiencies.