A Study on the Choice of Online Marketplace Co-Opetition Strategy Considering the Promotional Behavior of a Store on an E-Commerce Platform

Abstract

The huge traffic of e-commerce platforms provides a wide market for stores, but the fierce competition among stores for onsite traffic increases the cost of onsite promotion for stores, and it also reduces the effectiveness of promotion. Therefore, more and more stores on e-commerce platforms are using content platforms for offsite promotion. In the face of stores’ onsite and offsite promotion behaviors, the choice of competition and cooperation strategies among participants in the promotion process is the key to solving the problem of increasing traffic. Accordingly, this paper constructs an onsite and offsite promotion decision model consisting of an e-commerce platform, a store on the e-commerce platform, and a content platform, and it compares the results based on the decentralized decision situation, centralized decision situation, and promotion investment sharing situation. In addition, some results are presented in more detail in the form of numerical analysis. The results show that the promotion investment sharing does not change the level of onsite promotion investment, but the centralized decision makes the onsite promotion investment decrease; promotion investment sharing and centralized decision scenarios increase the promotion investment of the e-commerce platform and the content platform. When the sum of the parts of promotion investments shared by the store and the content platform for the e-commerce platform is relatively small, each participant will be willing to participate in the promotion investment sharing agreement. We believe that this study will provide important implications for the resolution of stores’ traffic dilemmas on e-commerce platforms.

1. Introduction

With the slowdown of the e-commerce market’s development, the traffic dividend of e-commerce platforms is gradually peaking, making the competition for traffic among stores on e-commerce platforms more and more intense [1,2]. As we all know, stores in the e-commerce platform mainly use “onsite promotion” to expand demand—that is, by paying the relevant promotion costs to the e-commerce platform to improve the exposure of the store or products, such as Taobao, to provide stores with paid promotion services [3,4]. With the slowdown of the development of e-commerce, on the one hand, the amount of platform users has stabilized, making it difficult to increase traffic; on the other hand, the increasing number of merchants enrolled on the platforms has led to increasing competition, resulting in a gradual decrease in the relative traffic of individual stores [5]. For example, in the whole Taobao system, 200,000 Tmall head stores account for 80% of the traffic, while 9 million small and medium-sized Taobao stores make up the remaining 20%. From 2018 to 2020, RUMERE Clothing’s promotion expenses on the Taobao platform increased by 147.71% on aggregate, but the payment conversion rate continued to decline, at 2.57%, 2.26%, and 1.9%, respectively. Obviously, although the huge traffic of e-commerce platforms provides a wide market for stores, the fierce competition among stores for onsite traffic not only increases the cost of onsite promotion for merchants, but also reduces the effectiveness of promotion.

In response to the online sales traffic dilemma, many studies have proposed solutions. One solution is that e-commerce platforms should invest more in advertising and assume a greater share of coupons to ease the promotion pressure on stores (Bakos and Brynjolfsson 2000) [6]. There is also a solution that e-commerce platforms should engage in cooperative marketing with stores through commission rates and product price contracts to attract traffic (Yang et al. 2015) [7]. Some scholars believe that the pricing of traffic by e-commerce platforms should be adjusted nonlinearly according to the differences in sales conversion rates between stores, and they suggest promoting interconnection between platforms to form a competitive traffic data market structure and promote the effective allocation of traffic data elements (Ambrus et al. 2016) [8]. In addition, some people believe that e-commerce platforms should match the advertising space according to the increased effect of the stores’ demand, thereby forming the optimal advertising resource allocation mechanism (Bouvard and Levy 2018) [9]. We believe that these measures can solve the stores’ onsite traffic dilemma to a certain extent. However, these measures solve the aforementioned problems from the perspective of the e-commerce platform—not from the perspective of the stores.

In the face of the above situation, more and more stores are gradually shifting their attention to outside of the e-commerce platform, using “offsite promotion” to expand the offsite demand or market [10]. At the same time, the rise of various emerging content platforms, such as TikTok, has provided options for the offsite promotion of stores. The main operation of offsite promotion is that the content platform displays the store’s product information with purchase links in its published content, the audience of the content platform clicks the links to jump to the e-commerce platform for purchase, and the store pays the content platform part of the revenue brought by clicking the links for purchase as the promotion fee. In order to attract more consumers, e-commerce platforms have been increasing their advertising investment, with each e-commerce platform placing their platform’s advertisements in crowded locations such as subways and bus stations. Especially before major e-commerce shopping festivals each year, advertisements for various promotions, offers, and discounts of each e-commerce platform can be seen everywhere offline, and there are also e-commerce platforms placing the above advertisements on various content platforms, such as video websites and browser homepages.

In the abovementioned onsite and offsite promotional activities, the reality is that the onsite promotion (or offsite promotion) activities of stores will have a certain positive impact on the offsite (or onsite) demand, and the academic community refers to this positive impact as the spillover effect of promotion. This gives rise to a number of issues that deserve our attention: (1) In the face of the spillover effect between onsite and offsite promotions and the relationship between the store and the e-commerce platform, or between the store and the content platform, how should each participant make optimal promotion investment decisions? (2) Some e-commerce platforms conduct promotional cooperation programs with stores, i.e., stores share part of the promotional investments of e-commerce platforms. In the face of this cooperation, how does the promotional cooperation between the store, the e-commerce platform, and the content platform affect the promotion investment decisions of each participant? (3) Can centralized decision-making among participants achieve an increase in the total profit of the entire online marketplace?

In order to answer these questions, we consider a framework in which a store on an e-commerce platform not only promotes within the platform to obtain onsite traffic, but also promotes on a content platform to obtain offsite traffic. We show that there is no correlation between the spillover effects of each participant’s promotion investment, while there is a positive correlation between the spillover effect of the store’s onsite promotion and its onsite promotion investment, and between the promotion spillover effect of the e-commerce platform and its promotion investment. In addition, the sharing of part of the promotion investment among the participants does not have a negative impact on the promotion investment of each participant; specifically, the onsite promotion investment of the store is affected by each sharing rate. The promotion investment of the e-commerce platform is not affected by its own sharing rate but increases with the sharing rate of the store and the content platform. The offsite promotion investment of the content platform only increases with the increase in the sharing rate of the e-commerce platform. Finally, the overall revenue of online marketplaces is lower than the overall revenue under centralized decision-making, whether in the scenarios of decentralized decision or promotion investment sharing. However, only when the sum of the investment shares between the store and content platform for the promotion of the e-commerce platform is relatively small can the form of investment sharing provide more overall benefit for the marketplace than decentralized decision-making.

The primary purpose of this study is to address the choice between competition and cooperation among participants in the process of onsite and offsite promotion of stores on e-commerce platforms by focusing on the traffic dilemma in the online sales process. In addition, we hope to achieve the following objectives through this study: (1) improving and perfecting the theoretical system of online promotion selection by depicting the reality of online promotion operations; and (2) further exploring the differences between onsite and offsite demand based on the realistic characteristics of online sales, to segment the onsite and offsite market, and to provide a more detailed depiction of the interaction between participants in the promotion process, so as to build a model that is closer to the actual onsite and offsite promotion decision-making, with a wider application scenario.

2. Related Literature

To address the research content of this paper, we review existing research in two fields: online promotion, and online marketplace choice. The development of online sales has drawn the attention of many scholars to the issue of online promotion [11,12,13]. More and more scholars believe that the rise of online advertising has greatly reduced the effectiveness of regular advertising [14,15,16]. This has further inspired scholars to study online promotion. For example, Chen et al. [17] analyzed how Internet retailers choose between resealing and agency when conducting online promotions. Drawing on the brokerage model adopted by eBay and the advertising model adopted by Taobao, Chen et al. [18] investigated how the revenue model affects the revenue of the platform, the payoff of buyers, the payoff of sellers, and social welfare. Rutz and Bucklin [19] mainly analyzed the overspill effect of generic search on branded search in paid Internet searching advertising. Ghose et al. [20] found that a ranking system based on consumer utility could lead to a significant increase in overall search engine revenue by examining the effects of three different search engines on consumer behavior and search engine revenues. Agarwal et al. [21] studied the impact of ad placement on search site revenue and profits with the help of data on several keywords from an online retailer’s ad campaign. The results found that while click-through rates decreased with position, transform rates increased with position. Goldfarb and Tucker [12] found that the loss of promotional effectiveness was more pronounced for sites with generic content. The loss of promotional effectiveness was also more pronounced for ads with less presence on the page, and for ads without additional interactive, video, or audio features.

Many scholars have studied online promotion through companies’ advertising behavior on search engines. For example, Yang et al. [22] argued that in searching for advertising, the payment of clicking and searching for advertising is private information of the ad placement company, and the asymmetry of this information hinders the optimal ad floor price decision of search platforms. Gopal et al. [23] concluded that there is significant cannibalization between the search and content channels in keyword-based online advertising, and that the profit of impressions for each channel is also decreasing significantly. Sayedi et al. [24] argued that advertising on search engines can reach consumers more closely. They studied how corporate advertising hitchhiking behavior affects the advertising budgets of traditional and search ads, using firms with different advertising budgets. Hu et al. [25] used incentive contracts to investigate how pay-per-click and pay-per-display models affect the risk sharing between advertisers and placeholders.

In addition, this study is related to the literature on online channel construction. Research in this area focuses on a firm’s choice between building its own e-commerce website and entering an e-commerce platform [3,26]. The growth of the e-commerce industry has driven research on online channel selection. Bernstein et al. [27] analyzed the choice of online channels for traditional retailers by constructing a model consisting of a traditional retailer and an e-tailer. Ru and Wang [28] and Ryan et al. [29] focused on Amazon’s operation model and analyzed the choices of traditional retailers to enter Amazon to sell their products or to cooperate with Amazon for sales. Karray and Sigue [4] believe that it is inevitable that traditional retailers will build online sales channels in the face of pressure from the online marketplace. They think that as long as the size of online market is large enough, traditional retailers should open online sales channels and differentiate pricing across channels. Chen et al. [5] focused on the impacts of traditional retailers’ offline discount stores on their online channel-building choices, arguing that retailers’ ultimate choice is determined by cost and by the threat of discount stores. Cao et al. [10] argued that channel choice is more important to consider in the trade-old-for-new context. By comparing retailers’ offline channel strategies, online channel strategies, and dual-channel strategies, they found that consumers’ shopping costs directly influence retailers’ channel choices. This study focuses on the same online sales issues inspired by the operational reality of Amazon and further expands to the promotional issues in the online sales process. We believe that the existing literature can provide a good reference for this study, and that our study, in turn, can enrich the research in related fields to some extent.

Summarizing the existing studies, we found that there has been abundant research on onsite promotion, but most of such studies suggest that the investment or level of promotion is determined by the e-commerce platform, without realizing the reality that the specific investment of onsite promotion is actually determined by the stores on e-commerce platforms. Although research on offsite promotion has also obtained some results, it has mainly focused on the competition between firms and advertising platforms. In addition, the research on advertising platforms focuses on the selection of charging models for advertising on advertising platforms. Finally, few existing relevant studies have explored the traffic distinction between online markets.

According to the promotional reality of the online sales process and the shortage of related research, this research mainly contributes to the innovative exploration and enrichment of research in related fields in the following ways: (1) It builds an online marketing decision model consisting of an e-commerce platform, a store on the platform, and a content platform by drawing on the actual onsite and offsite promotional operations of stores on e-commerce platforms. (2) The entire online market is subdivided into the onsite market and offsite market. In this case, the onsite promotion investment is decided by the store, while the offsite promotion investment is decided by the content platform, so as to study the promotion investment decision and profit of each participant under different decision scenarios. We believe that this study will enrich the results on online sales promotion to a certain extent, and that it will also provide a good reference for the promotion or advertising issues in the supply chain field—especially in the dual-channel field.

The remainder of this paper is structured as follows: In Section 3, we describe the research methodology. The equilibrium results for decentralized and centralized decision scenarios are described in Section 3.1 and Section 3.2, respectively, and the results for the promotion investment sharing scenarios are given in Section 3.3. Section 4 presents the results for empirical analysis. Section 5 concludes the paper.

3. Research Methodology

This paper constructs an online marketplace promotion decision model consisting of an e-commerce platform, a store on the e-commerce platform, and a content platform. In this model, the store accepts the referral rate $\lambda$ of the e-commerce platform and enters the e-commerce platform. Then, the store uses the paid promotion services of the e-commerce platform to conduct onsite promotion to attract onsite customers of the e-commerce platform, but the onsite promotion investment $I_r$ is decided by the store itself. At the same time, the store publishes product information by using offsite promotion with the help of the content platform to attract consumers outside the e-commerce platform. The operation of the promotion mode is as follows: The content platform introduces the store products with purchase links in the content that it publishes (e.g., videos or articles), the audience of the content platform clicks on the purchase links to jump directly to the store on the e-commerce platform to place orders, and the store pays a certain percentage $x$ of the sales revenue generated by clicking on the purchase links as the promotion fee of the content platform, while the content platform decides its own promotion investment $I_n$ based on the promotion profit. In addition, in order to attract more consumers, the e-commerce platform needs to invest a certain fee $I_e$ in advertising or posting information in high-traffic areas, such as subways and bus stations, which to a certain extent have a boosting effect on the demand for stores. The model structure is shown in Figure 1.

According to the promotional reality, when stores implement offsite promotion, the price when the customer enters the store of the e-commerce platform by clicking on the offsite link to place an order is the same as the price when the order is placed directly on the e-commerce platform, so we can assume that the marginal revenue of the store is y. We can also assume that the demand of the whole online market unaffected by advertising is $N$, but some of the customers will not enter the e-commerce platform to buy, so the demand of the online market unaffected by advertising will be divided into two parts: onsite and offsite; the proportion of onsite demand $\alpha$ is $\alpha N$, and the proportion of offsite demand $1-\alpha$ is $(1-\alpha )N$. Many studies have made assumptions about the relationship between advertising investments and demand, so the effect of advertising investments on demand is assumed to be $\sqrt{I_i}$. At the same time, considering that onsite (offsite) promotion will have an impact on offsite (onsite) demand in addition to onsite (offsite) demand, we can assume that the impact of onsite (offsite) advertising on offsite (onsite) demand is $k_r\sqrt{I_r}$ ($k_p\sqrt{k_p}$). However, the promotion placed by the e-commerce platform is not targeted to a particular store, so the impact of the e-commerce platform’s promotion on the store’s onsite demand is assumed to be $k_e\sqrt{I_e}$, and $k_r$, $k_e$, and $k_p$ can be interpreted as the spillover effects of each promotion. According to the above description, it can be seen that the onsite and offsite demand of the store under the influence of promotion is as follows:

$$D_n=\alpha N+\sqrt{I_r}+k_p\sqrt{I_p}+k_e\sqrt{I_e}$$

(1)

$$D_w=(1-\alpha )N+\sqrt{I_p}+k_r\sqrt{I_r}$$

(2)

3.1. Decentralized Decision Scenario

In this scenario, the store, the e-commerce platform, and the content platform compete with one another, and the participants each decide their own optimal promotion investment decisions to achieve optimal profit. To differentiate, the superscript $jz$ is used to denote this scenario. The profit functions for each participant in this scenario are as follows:

$${\prod }_r^{jz}=\left\{\begin{array}{l}(1-\lambda )y(\alpha N+\sqrt{I_r}+k_p\sqrt{I_p}+k_e\sqrt{I_e})+\\ (1-x-\lambda )y[(1-\alpha )N+\sqrt{I_p}+k_r\sqrt{I_r}]-I_r\end{array}\right.$$

(3)

$${\prod }_e^{jz}=\lambda y(\alpha N_n+\sqrt{I_r}+k_p\sqrt{I_p}+k_e\sqrt{I_e})+\lambda y[(1-\alpha )N_w+\sqrt{I_p}+k_r\sqrt{I_r}]-I_e$$

(4)

$${\prod }_p^{jz}=xy[(1-\alpha )N+\sqrt{I_p}+k_r\sqrt{I_r}]-I_p$$

(5)

Proposition 1.

When the store on the e-commerce platform is promoted on- and offsite, if there is no promotional cooperation among the participants, there must be a unique equilibrium decision $(I_r^{jz}{}^{*}, I_e^{jz}{}^{*}, I_p^{jz}{}^{*})$ of promotion investment so that the store, e-commerce platform, and content platform can achieve optimal profit at the same time, where $I_p^{jz}{}^{*}=\frac{x^2{y}^2}{4}$, $I_e^{jz}{}^{*}=\frac{{\lambda }^2{y}^2{k}_e{}^2}{4}$, $I_r^{jz}{}^{*}=\frac{y^2{[1-\lambda +k_r(1-x-\lambda )]}^2}{4}$.

Proof. 

The second-order derivative of Equation (3) with respect to the store’s onsite promotion inputs yields $\frac{{\partial }^2{\prod }_r^{jz}}{\partial I_r{}^2}=\frac{-y[1-\lambda +k_r(1-x-\lambda )]}{4\sqrt{I_r{}^3}}<0$. Due to the presence of a condition $1-x-\lambda >0$, we know that there must be a unique onsite advertising investment decision to make the store optimal and then solve the equation $\frac{\partial {\prod }_r^{jz}}{\partial I_r}=0$ so that we can find $I_r^{jz}{}^{*}$. Substituting the expression of $I_r^{jz}{}^{*}$ into Equation (4) for the second-order derivative of $I_e$ and Equation (5) for the second-order derivative of $I_p$, we have $\frac{{\partial }^2{\prod }_e^{jz}(I_r^{jz}{}^{*})}{\partial I_e{}^2}=-\frac{\lambda y{k}_e}{4\sqrt{I_e{}^3}}<0$ and $\frac{{\partial }^2{\prod }_e^{jz}(I_r^{jz}{}^{*})}{\partial I_e{}^2}=-\frac{\lambda y{k}_e}{4\sqrt{I^3}}<0$, respectively. It follows that there must be unique $I_e$ and $I_p$, such that the e-commerce platform and the content platform simultaneously achieve optimal profit. Finally, solving equations $\frac{\partial {\prod }_e^{jz}(I_r^{jz}{}^{*})}{\partial I_e}=0$ and $\frac{\partial {\prod }_p^{jz}(I_r^{jz}{}^{*})}{\partial I_p}=0$, we can find the expressions of $I_e^{jz}{}^{*}$ and $I_p^{jz}{}^{*}$, respectively. □

It is thus clear that

$${\prod }_r^{jz}{}^{*}=\left\{\begin{array}{l}\frac{y}{4}[y(1+k_r{}^2)+4(1-\lambda -x+\alpha x)N+y(2x{k}_r(1-\lambda )(1+k_r)+\\ k_p+1-x^2(2-k_r{}^3)+\lambda (k(1-\lambda )k_e{}^2-(2-\lambda ){(1+k_r)}^2))]\end{array}\right.$$

(6)

$${\prod }_e^{jz}{}^{*}=\frac{\lambda y}{4}[4N+y(2+2x(1+k_p)-\lambda (2-k_e{}^2)+2k_r((1-x-\lambda )k_r+2-x-2\lambda ))]$$

(7)

$${\prod }_p^{jz}{}^{*}=\frac{xy}{4}[4\alpha N+y(x+2k_r(1-\lambda +k_r(1-x-\lambda )))]$$

(8)

Corollary 1.

In the decentralized decision scenario, there must be $I_r^{jz}{}^{*}>I_p^{jz}{}^{*}$.

The result of Corollary 1 suggests that when the store on an e-commerce platform conducts onsite and offsite promotions, it will invest more in onsite promotion than in offsite promotion on the content platform if they make centralized decisions. This is because we assume that $x+\lambda <1$, i.e., the referral fee rate for the e-commerce platform and the promotion fee rate for the content platform are small. This makes it unlikely that the content platform will invest much in attracting consumers, or its revenue will not cover the investment. If the offsite promotion fee rate is high, the content platform has an incentive to increase the level of offsite promotion investment. However, our assumption is consistent with the reality that the offsite promotion investment of the e-commerce platform cannot be higher than its revenue, or else the store will not conduct offsite promotion.

The variation of each equilibrium result with the parameters in the decentralized decision scenario is shown in

Table 1. The variation of equilibrium results in the decentralized decision scenario.

	$I_r^{jz}{}^{*}$	$I_e^{jz}{}^{*}$	$I_p^{jz}{}^{*}$	${\prod }_r^{jz}{}^{*}$	${\prod }_e^{jz}{}^{*}$	${\prod }_p^{jz}{}^{*}$
$y$	↗	↗	↗	↗	↗	↗
$x$	↘	⊗	↗	↘	If $k_p>\max \{0, \stackrel{\nabla }{k_p}\}$, ↗	↗
					If $k_p<\max \{0, \stackrel{\nabla }{k_p}\}$, ↘
$\lambda$	↘	↗	⊗	↘	If $\lambda >\stackrel{\nabla }{\lambda }$, ↗	↘
					If $\lambda <\stackrel{\nabla }{\lambda }$, ↘
$k_r$	↗	⊗	⊗	↗	↗	↗
$k_e$	⊗	↗	⊗	↗	↗	⊗
$k_p$	⊗	⊗	⊗	↗	↗	⊗

where $\stackrel{\nabla }{\lambda }=\frac{2N+y{(1+k_r)}^2-xy(k_p-\stackrel{\nabla }{k_p})}{y[2{(1+k_r)}^2-k_e{}^2]}$; $\stackrel{\nabla }{k_p}=k_r{}^2+k_r-1$; ↗ and ↘ indicate that the equilibrium result is an increasing and decreasing function of the corresponding parameter, respectively; and ⊗ indicates that the equilibrium result does not change with the corresponding parameter.

.

As can be seen from Table 1, in the decentralized decision scenario, the promotion investment and profit of the e-commerce platform, store, and content platform all increase with the marginal revenue of the store, while the above investment decisions and revenue vary in performance with each parameter. Although the profit of the store and the content platform will decrease and increase with the increase in the offsite promotion fee rate, respectively, the profit of the e-commerce platform will increase with the increase in the offsite promotion fee rate when the promotion spillover effect of the content platform is larger. This is because, when the promotion spillover effect of the content platform is larger, the store’s onsite demand will benefit from it and gain more profit, when the e-commerce platform can receive a greater referral fee. In addition, although the profits of both the store and content platform decrease with the referral fee rate of the e-commerce platform, the e-commerce platform does not necessarily benefit fully from the referral fee rate. Only if the referral fee rate is kept at a high range will the profit of the e-commerce platform increase with the referral fee rate. If the referral rate is low, the e-commerce platform’s profit also decreases with that referral fee rate. This is because the e-commerce platform needs to invest a certain amount of promotion costs to attract consumers; in the case of a certain amount of investment, if the referral fee rate is relatively low, the profit of the store will decrease with the referral fee rate, reducing the revenue of the e-commerce platform. Finally, there is no correlation between the spillover effect of each participant’s promotion and each promotion investment, and there is a positive correlation between the spillover effect of the store’s onsite promotion and its onsite promotion investment, as well as between the promotion spillover effect of the e-commerce platform and its promotion input. Additionally, there is generally a positive correlation between the above spillover effect and the revenue of each participant, but there is no correlation between the revenue of the content platform and the promotion spillover effect of the e-commerce platform and its own promotion spillover effect. This is because the size of the content platform’s promotion spillover effect does not become the basis for the content platform’s promotional charges.

3.2. Centralized Decision Scenario

In this decision scenario, the participants in the whole promotion process are treated as a whole, and they jointly make the promotion investment decisions of different parts to achieve the optimal overall profit. We denote this case by the superscript $zt$. Then, the expression for the whole profit function in this case is as follows:

$${\prod }_{r+e+p}^{zt}=y(\alpha N+\sqrt{I_r}+k_p\sqrt{I_p}+k_e\sqrt{I_e})+y[(1-\alpha )N+\sqrt{I_p}+k_r\sqrt{I_r}]-I_r-I_e-I_p$$

(9)

Proposition 2.

In the case of onsite and offsite promotion by the store in an e-commerce platform, if there is centralized decision-making among the participants, there is a unique equilibrium decision on promotion investment $(I_r^{zt}{}^{*}, I_e^{zt}{}^{*}, I_p^{zt}{}^{*})$, so that the whole profit of all participants in the promotion process is optimized, where $I_r^{zt}{}^{*}=\frac{y^2{(1+k_r)}^2}{4}$, $I_e^{zt}{}^{*}=\frac{y^2{k}_e{}^2}{4}$, and $I_p^{zt}{}^{*}=\frac{y^2{(1+k_p)}^2}{4}$.

Proof.

Finding the Hessian matrix of the return function Equation (9) with respect to $I_r$, $I_e$, and $I_p$ gives the following:

$$H(I_r,I_e,I_p)=\left[\begin{array}{ccc}\frac{{\partial }^2{\prod }_{r+e+p}^{zt}}{\partial I_r{}^2}& \frac{{\partial }^2{\prod }_{r+e+p}^{zt}}{\partial I_r\partial I_e}& \frac{{\partial }^2{\prod }_{r+e+p}^{zt}}{\partial I_r\partial I_p}\\ \frac{{\partial }^2{\prod }_{r+e+p}^{zt}}{\partial I_e\partial I_r}& \frac{{\partial }^2{\prod }_{r+e+p}^{zt}}{\partial I_e{}^2}& \frac{{\partial }^2{\prod }_{r+e+p}^{zt}}{\partial I_e\partial I_p}\\ \frac{{\partial }^2{\prod }_{r+e+p}^{zt}}{\partial I_p\partial I_r}& \frac{{\partial }^2{\prod }_{r+e+p}^{zt}}{\partial I_p\partial I_e}& \frac{{\partial }^2{\prod }_{r+e+p}^{zt}}{\partial I_p{}^2}\end{array}\right]=\left[\begin{array}{ccc}-\frac{y(1+k_r)}{4\sqrt{I_r{}^3}}& 0& 0\\ 0& -\frac{y{k}_e}{4\sqrt{I_e{}^3}}& 0\\ 0& 0& -\frac{y(1+k_p)}{4\sqrt{I_p{}^3}}\end{array}\right]$$

It is clear that the Hessian matrix is definitely negative. Therefore, there must be a unique equilibrium solution that makes the profit function optimal. Finally, solving the system of equations $(\frac{\partial {\prod }_{r+e+p}^{zt}}{\partial I_r}=0, \frac{\partial {\prod }_{r+e+p}^{zt}}{\partial I_e}=0, \frac{\partial {\prod }_{r+e+p}^{zt}}{\partial I_p}=0)$ yields the expressions $I_r^{tz}{}^{*}$, $I_e^{tz}{}^{*}$, and $I_p^{tz}{}^{*}$. □

The whole optimal profit expression for the centralized decision case can be obtained from Proposition 2 as follows:

$${\prod }_{r+e+p}^{zt}{}^{*}=\frac{y[4N+y(2+k_e{}^2+k_p(2+k_p)+k_r(2+k_r))]}{4}$$

(10)

Corollary 2.

In the case of centralized decision-making, the optimal promotion investment decision of each part is the increasing function of the marginal revenue of the store and the increasing function of the spillover effect of each promotion investment on the demand of the other parts; moreover, the whole optimal profit in the case of centralized decision-making increases with the marginal revenue of the store and the spillover effect of each part of the promotion.

Proof. 

Since the equilibrium decision and the optimal payoff expression in this case are relatively simple, the detailed proof process is not shown here. □

From Corollary 2, if the store carries out onsite and offsite promotion, the overall decision should increase the promotion investment of each part according to the increase in the marginal revenue, and the change in the promotion investment of each part should also adjust according to the size of the spillover effect of that part of the promotion, so as to ensure the growth of its profit. The specific performance is an increase in the investment as the spillover effect of each promotion component increases. The above results give us a management insight: under the centralized decision of onsite and offsite store promotion, decision-makers only need to pay attention to the changing trend of the store’s marginal revenue and the spillover effect of each part of the promotion to make the targeted adjustments.

Corollary 3.

$\min \{I_r^{tz}{}^{*}, I_p^{tz}{}^{*}\}>I_e^{tz}{}^{*}$; if $k_r>k_p$, $I_r^{tz}{}^{*}> I_p^{tz}{}^{*}$; if $k_r<k_p$, $I_r^{tz}{}^{*}<I_p^{tz}{}^{*}$.

Proof. 

The results can be obtained by comparing the equilibrium solution expressions for each promotion investment. □

From Corollary 3, it is clear that in the centralized decision situation, the decision-maker should make the e-commerce platform’s promotion the lowest investment among the three kinds of promotions, while the investment in onsite and offsite promotions should be decided according to their respective spillover effects. If the spillover effect of onsite promotion is greater than that of offsite promotion, more budget should be guaranteed for onsite promotion; otherwise, more budget should be guaranteed for offsite promotion. The managerial insight revealed by the above results is that decision-makers should not only see the impact of one type of promotion on the corresponding range of demand when making onsite and offsite promotion investment decisions, but they should also clearly understand the impact of that promotion on other segments of consumers in order to make better promotion investment decisions.

3.3. Promotion Investment Sharing Scenarios

In this case, we combine the decentralized and centralized decision scenarios to a certain extent, so that each participant shares a certain percentage of the other participants’ promotion investment. Since the store needs to pay a promotion fee to the content platform under the offsite promotion scenario, the case of stores sharing the promotion investment of the content platform is not considered. In view of this, we simultaneously consider the case where the store shares part of the promotion investment of the e-commerce platform ${\rho }_r$, the e-commerce platform shares part of the promotion investment of the content platform ${\rho }_e$, and the content platform shares part of the promotion investment of the e-commerce platform ${\rho }_p$. Using the superscript $hz$ to denote this scenario, it can be seen that the profits for each participant are as follows:

$${\prod }_r^{jz}=\left\{\begin{array}{l}(1-\lambda )y(\alpha N+\sqrt{I_r}+k_p\sqrt{I_p}+k_e\sqrt{I_e})+\\ (1-x-\lambda )y[(1-\alpha )N+\sqrt{I_p}+k_r\sqrt{I_r}]-I_r-{\rho }_r{I}_e\end{array}\right.$$

(11)

$${\prod }_e^{jz}=\left\{\begin{array}{l}\lambda y(\alpha N+\sqrt{I_r}+k_p\sqrt{I_p}+k_e\sqrt{I_e})+\\ \lambda y[(1-\alpha )N+\sqrt{I_p}+k_r\sqrt{I_r}]-I_e-{\rho }_e{I}_p+{\rho }_r{I}_e+{\rho }_p{I}_e\end{array}\right.$$

(12)

$${\prod }_p^{jz}=xy[(1-\alpha )N+\sqrt{I_p}+k_r\sqrt{I_r})-I_p-{\rho }_p{I}_e+{\rho }_e{I}_p$$

(13)

Proposition 3.

In the case of onsite and offsite promotion, if the store shares part of the promotion investment of the e-commerce platform, and if the e-commerce platform and the content platform share part of one another’s promotion investment, there must be a unique equilibrium decision $(I_r^{hz}{}^{*}, I_e^{hz}{}^{*}, I_p^{hz}{}^{*})$ of promotion investment so that the profit of each participant is optimal, where $I_r^{hz}{}^{*}=\frac{y^2{[1-\lambda +k_r(1-x-\lambda )]}^2}{4}$, $I_e^{hz}{}^{*}=\frac{{\lambda }^2{y}^2{k}_e{}^2}{4{(1-{\rho }_r-{\rho }_p)}^2}$, and $I_p^{hz}{}^{*}=\frac{x^2{y}^2}{4{(1-{\rho }_e)}^2}$.

Proof. 

The proof is similar to the proof of Proposition 1; we do not go into details here. □

From Proposition 3, the optimal profits for each participant in the promotion investment sharing scenario are as follows:

$${\prod }_r^{hz}{}^{*}=\left\{\begin{array}{l}\frac{y}{4}[4N(1-x-\lambda +\alpha x)+y(1-\frac{2x^2}{1-{\rho }_e}+\frac{2x(1-\lambda )(1+k_p)}{1-{\rho }_e}+\\ 2k_r(1-\lambda )(1-x-\lambda )+k_r{}^2{(1-x-\lambda )}^2+\\ \lambda (\lambda -2+\frac{k_e{}^2(2(1-{\rho }_r-{\rho }_p)-\lambda (2-{\rho }_r-2{\rho }_p))}{{(1-{\rho }_r-{\rho }_p)}^2}))]\end{array}\right.$$

(14)

$${\prod }_e^{hz}{}^{*}=\left\{\begin{array}{l}\frac{y}{4}[4\lambda N-\frac{2y{x}^2{\rho }_e}{{(1-{\rho }_e)}^2}+\frac{2\lambda xy(1+k_p)}{1-{\rho }_e}+\\ \frac{y{\lambda }^2{k}_e{}^2}{1-{\rho }_r-{\rho }_e}+2\lambda y(1+k_r)(1-\lambda +k_r(1-x-\lambda ))]\end{array}\right.$$

(15)

$${\prod }_p^{hz}{}^{*}=\frac{y}{4}[4x(1-\alpha )N+y(\frac{x}{1-{\rho }_e}+2x{k}_r(1-\lambda )-\frac{{\rho }_p{\lambda }^2{k}_e{}^2}{{(1-{\rho }_r-{\rho }_e)}^2}-2x{k}_r{}^2(1-x-\lambda ))]$$

(16)

Corollary 4.

Under the promotion investment sharing scenario, if $\frac{1-x-\lambda }{1-\lambda }<{\rho }_e<\frac{2(1-x-\lambda )}{2-x-2\lambda }\&k_r<\frac{\lambda {\rho }_e+x+\lambda -1}{(1-{\rho }_e)(1-x-\lambda )}$, or $\frac{1-x-\lambda }{1-\lambda }<{\rho }_e<1$, there is $I_r^{hz}{}^{*}<I_p^{hz}{}^{*}$; otherwise, there is $I_r^{hz}{}^{*}>I_p^{hz}{}^{*}$.

Proof. 

The results can be obtained by comparing the equilibrium solution expressions for each promotion investment. □

By comparing the results of Corollary 1 and Corollary 4, it can be seen that promotion investment sharing changes the relationship between onsite and offsite promotion investment. Specifically, when the e-commerce platform’s sharing of offsite promotion investment to the content platform is low, or when the sharing is moderate and the spillover effect of onsite promotion is large, the store’s onsite promotion investment will be higher than the content platform’s offsite promotion investment. This result is different from the decentralized decision situation, in which the onsite promotion investment must be higher than the offsite promotion investment. This is because when the e-commerce platform can share more promotion investment for the content platform, or when the sharing is moderate and the onsite promotion can bring a greater spillover effect for the offsite promotion, the content platform will gain more profit by increasing the investment. The management insight revealed by the above results is that when stores conduct onsite and offsite promotion, they can incentivize content platforms to increase their offsite promotion investment by increasing the spillover effect of onsite promotion, or they can convince e-commerce platforms to share a certain amount of promotion investment for content platforms.

The variation of each equilibrium result with each parameter for the investment sharing scenario is shown in

Table 2. The variation of equilibrium results in the promotion investment sharing scenario.

	${\mathit{I}}_{\mathit{r}}^{\mathit{h}\mathit{z}}{}^{*}$	${\mathit{I}}_{\mathit{e}}^{\mathit{h}\mathit{z}}{}^{*}$	${\mathit{I}}_{\mathit{p}}^{\mathit{h}\mathit{z}}{}^{*}$	${\prod }_{\mathit{r}}^{\mathit{h}\mathit{z}}{}^{*}$	${\prod }_{\mathit{e}}^{\mathit{h}\mathit{z}}{}^{*}$	${\prod }_{\mathit{p}}^{\mathit{h}\mathit{z}}{}^{*}$
$y$	↗	↗	↗	If $y>\stackrel{\angle }{y}$, ↗	If $y>\stackrel{\nabla }{y}$, ↗	If $y>\stackrel{\Delta }{y}$, ↗
				If $y<\stackrel{\angle }{y}$, ↘	If $y<\stackrel{\nabla }{y}$, ↘	If $y<\stackrel{\Delta }{y}$, ↘
$x$	↘	⊗	↗	If $x>\min \{\stackrel{\wedge }{x}, 1\}$, ↗	If $x>\min \{\stackrel{-}{x}, 1\}$, ↗	↗
				If $x<\min \{\stackrel{\wedge }{x}, 1\}$, ↘	If $x<\min \{\stackrel{-}{x}, 1\}$, ↘
$\lambda$	↘	↗	⊗	If $\lambda <\min \{\stackrel{⊳}{\lambda }, 1\}$, ↗	If $\lambda >\min \{\stackrel{\Delta }{\lambda }, 1\}$, ↗	↘
				If $\lambda >\min \{\stackrel{⊳}{\lambda }, 1\}$, ↘	If $\lambda <\min \{\stackrel{\Delta }{\lambda }, 1\}$, ↘
$k_r$	↗	⊗	⊗	↗	↗	↗
$k_e$	⊗	↗	⊗	⊗	⊗	↘
$k_p$	⊗	⊗	⊗	↗	↗	⊗
${\rho }_r$	⊗	↗	⊗	If ${\rho }_r<\min \{{\stackrel{\Delta }{\rho }}_r, 1\}$, ↗	↗	↘
				If ${\rho }_r>\min \{{\stackrel{\Delta }{\rho }}_r, 1\}$, ↘
${\rho }_e$	⊗	⊗	↗	↗	If ${\rho }_e<\stackrel{\Delta }{{\rho }_e}$, ↗	↗
					If ${\rho }_e>\stackrel{\Delta }{{\rho }_e}$, ↘
${\rho }_p$	⊗	↗	⊗	If ${\rho }_p<\stackrel{\Delta }{{\rho }_p}$, ↗	↗	↘
				If ${\rho }_p>\stackrel{\Delta }{{\rho }_p}$, ↘

Note: The expressions for $\stackrel{\angle }{y}$, $\stackrel{\nabla }{y}$, $\stackrel{\Delta }{y}$, $\stackrel{\wedge }{x}$, $\stackrel{-}{x}$, $\stackrel{⊳}{\lambda }$, and $\stackrel{\Delta }{\lambda }$ are not shown here because they are complex; $\stackrel{\angle }{{\rho }_r}=\frac{1+3\lambda {\rho }_p-3\lambda -{\rho }_p}{1-\lambda }$, $\stackrel{\Delta }{{\rho }_e}=1-\frac{2x}{x+2\lambda (1+k_p)}$, and $\stackrel{\Delta }{{\rho }_p}=\frac{1-{\rho }_r-\lambda }{1-\lambda }$.

.

By comparing the results in Table 1 and Table 2, it is clear that the impacts of the factors on the participants’ decisions and profits are more complex in the promotion investment sharing scenario than that in the decentralized decision scenario. First, although each participant in the promotion investment sharing scenario increases their promotion investment due to the increase in the marginal revenue of the store, the profit of each participant will increase with the marginal revenue only if that marginal revenue is high enough. Secondly, although the profit of the content platform increases with the offsite promotion fee rate and decreases with the referral fee rate, the profit of the e-commerce platform and the store increases simultaneously only when the offsite promotion fee rate is high, and the profit of both participants decreases with the promotion fee rate when it is low. This is because when the offsite promotion fee rate is low, the content platform will also reduce its promotion investment, reducing the demand of offsite promotion and directly affecting the profit of the store and the e-commerce platform. The opposite is true for the e-commerce platform and the store in terms of the change in profit with the referral fee rate; the result is more intuitive and will not be explained here. In addition, sharing part of the promotion investment among participants does not have a negative impact on the promotion investment of each participant; specifically, the onsite promotion investment of the store is affected by each sharing rate. The promotion investment of the e-commerce platform is not affected by its own sharing rate but increases with the sharing rate of the store and the content platform. The offsite promotion investment of the content platform only increases with the increase in the sharing rate of the e-commerce platform. Finally, the profit of the e-commerce platform and the content platform varies with the sharing rate of the store or the content platform in the opposite way. When the sharing rate of the store or the sharing rate of the e-commerce platform is kept small, the store’s profit increases with the above sharing rate; otherwise, it decreases with it. When the sharing rate of the e-commerce platform is low, its profit will increase with this sharing rate.

4. Empirical Results

We analyzed the decision problem in different scenarios separately, and in order to present the relevant results more intuitively, we next focus on comparing the equilibrium results of each participant in different scenarios and the whole profit of the online marketplace.

Corollary 5.

(1) $I_r^{jz}{}^{*}=I_r^{hz}{}^{*}>I_r^{zt}{}^{*}$; (2) if $\stackrel{⊲}{\lambda }<\lambda <1$ and $0<{\rho }_e<\stackrel{\nabla }{{\rho }_e}$, there is ${\prod }_r^{jz}{}^{*}>{\prod }_r^{hz}{}^{*}$; if $0<\lambda <\stackrel{⊲}{\lambda }$ or $\stackrel{⊲}{\lambda }<\lambda <1\&\stackrel{\nabla }{{\rho }_e}<{\rho }_e<1$, there is ${\prod }_r^{jz}{}^{*}<{\prod }_r^{hz}{}^{*}$, where $\stackrel{⊲}{\lambda }=\frac{2({\rho }_r+{\rho }_p)({\rho }_p+{\rho }_r-1)}{2{({\rho }_r+{\rho }_p)}^2-2{\rho }_p-3{\rho }_r}$; the expression for $\stackrel{\nabla }{{\rho }_e}$ is more complex and will not be shown here.

Proof. 

Compare the equilibrium results of the store in three decision scenarios. Because the complexity of the calculation requires the use of mathematical software, we will not repeat it here, and specific calculation commands can be provided if needed. □

From Corollary 5, it is clear that in the process of onsite and offsite promotion by a store, the store will not change its level of investment in onsite promotion, regardless of whether the promotion investment costs are shared among the participants, and this investment must be higher than the investment in onsite promotion under the centralized decision scenario. However, the same level of onsite investment in the abovementioned co-opetition situation does not necessarily bring the same profit to the store; when the e-commerce platform charges the store a higher referral fee rate and the content platform’s share of onsite promotion investment is small, for the store, the decentralized decision scenario will bring more profit than the other scenarios. The management insight revealed by the above results is that when promoting onsite and offsite at the same time, whether a store chooses to cooperate with other participants to share the promotion investment should depend not only on the investment sharing among the participants, but also on the size of the referral fee that they should share.

Corollary 6.

(1) $I_e^{jz}{}^{*}<\min \{I_e^{zt}{}^{*}, I_e^{hz}{}^{*}\}$; if ${\rho }_r+{\rho }_p<1-\lambda$, there is $I_e^{zt}{}^{*}>I_e^{hz}{}^{*}$; otherwise, there is $I_e^{zt}{}^{*}<I_e^{hz}{}^{*}$; (2) if $k_p>\min \{\frac{x}{2\lambda (1-{\rho }_e)}-\frac{\lambda k_e{}^2(1-{\rho }_e)({\rho }_r+{\rho }_p)}{2x{\rho }_e(1-{\rho }_r-{\rho }_p)}-1, 1\}$, there is ${\prod }_e^{hz}{}^{*}>{\prod }_e^{jz}{}^{*}$; otherwise, there is ${\prod }_e^{hz}{}^{*}<{\prod }_e^{jz}{}^{*}$.

Proof. 

This is same as the proof of Corollary 5, and it will not be shown here. □

In order to better demonstrate the difference in the profit of the e-commerce platform under the decentralized decision and promotion investment sharing decision scenarios, we analyzed it more specifically in the form of an arithmetic example, where $\Delta {\prod }_e={\prod }_e^{hz}{}^{*}-{\prod }_e^{jz}{}^{*}$. This is shown in Figure 2.

From Corollary 6, it can be seen that the e-commerce platform will change its promotion investment according to the decision scenario, as shown by the fact that under any conditions, the lowest promotion investment is in the decentralized decision scenario, while the e-commerce platform is more willing to make a greater promotion investment in the centralized decision scenario, when the sharing rate between the store and the content platform in the promotion investment sharing scenario is relatively small. This is because under the decentralized decision scenario, there is a perfectly competitive relationship between the players, the e-commerce platform does not profit from the promotion effect, and there is no need to invest too much in the store. Additionally, when the centralized decision scenario is achieved, the e-commerce platform can attract consumers by investing more in promotion to receive a greater referral fee. However, the promotion investment sharing scenario can generate more profit for the e-commerce platform when the promotion spillover effect of the content platform is relatively large. Figure 2 further verifies the results of Corollary 6, i.e., the profit of the e-commerce platform in the decentralized decision case is higher than that in the cooperative decision case only when $k_p$ is small; although the above profit difference and the parameter range of the difference change to some extent as $k_e$ increases, this does not change the relevant results. The management insight revealed by the above results is that when stores conduct onsite and offsite promotion, the e-commerce platform should focus not only on the size of the spillover effect of its own promotion, but also on the spillover effect of the content platform’s promotion, as this is directly related to the e-commerce platform’s profit.

Corollary 7.

(1) $I_p^{jz}{}^{*}<\min \{I_p^{hz}{}^{*}, I_p^{zt}{}^{*}\}$; if $x<(1-{\rho }_e)(1+k_p)$, there is $I_p^{zt}{}^{*}>I_p^{hz}{}^{*}$; (2) if $0<{\rho }_e<\frac{\lambda {\rho }_p{k}_e{}^2}{x^2{(1-{\rho }_p)}^2+{\rho }_r{x}^2({\rho }_r+2{\rho }_p-2)+{\lambda }^2{\rho }_p{k}_e{}^2}$, there is ${\prod }_p^{jz}{}^{*}>{\prod }_p^{hz}{}^{*}$; otherwise, there is ${\prod }_p^{jz}{}^{*}<{\prod }_p^{hz}{}^{*}$.

Proof. 

This is same as the proof of Corollary 5, and it will not be shown here. □

From Corollary 7, it is clear that, as in the case of the e-commerce platform, the content platform will change its offsite promotion investment level according to the decision scenario, and regardless of the change in conditions, the promotion investment under the decentralized decision scenario must be lower than that under other scenarios. Moreover, when the promotion fee rate charged to the store is small, the content platform is more willing to invest more in offsite promotion under the centralized decision scenario than under the promotion investment sharing scenario. This is because, when the promotion fee rate is small, the content platform cannot compensate for its loss of profit, even though it can receive some fees from other participants in the promotion investment sharing scenario. In contrast, when the promotion fee rate is larger, the content platform can gain more revenue, so it is more willing to increase its promotion investment. When the e-commerce platform’s share of promotion investment for the content platform is small (i.e., ${\rho }_e$ is small), the promotion investment sharing scenario will instead result in lower profit for the content platform than the decentralized decision scenario.

As can be seen from Figure 3, the content platform in the promotion investment sharing scenario is more likely to gain more profit when ${\rho }_e$ is larger than in the decentralized decision scenario. Although the magnitude of the above profit differences is influenced to some extent by the promotion spillover effect of the e-commerce platform, this does not change the above results and only has some impact on the feasible range of the relevant results. The management insight revealed by the above results is that in the face of store promotion behavior, content platforms should not only decide the specific level of promotion investment based on the promotion fee received from the store’s provenance, but should also have a clear understanding of the promotion investment sharing level that e-commerce platforms can provide in order to make better decisions.

Corollary 8.

(1) $\max \{{\prod }_{r+e+p}^{jz}{}^{*}, {\prod }_{r+e+p}^{hz}{}^{*}\}<{\prod }_{r+e+p}^{zt}{}^{*}$; (2) there exists a promotion investment sharing combination $({\rho }_r{}^{*}, {\rho }_p{}^{*})$ of the store and the content platform, such that when ${\rho }_r+{\rho }_p>{\rho }_r{}^{*}+{\rho }_p{}^{*}=F(x, \lambda , k_e, k_p,t_e)$, there is ${\prod }_{r+e+p}^{jz}{}^{*}> {\prod }_{r+e+p}^{hz}{}^{*}$; otherwise, there is ${\prod }_{r+e+p}^{jz}{}^{*}<{\prod }_{r+e+p}^{hz}{}^{*}$.

Proof. 

This is same as the proof of Corollary 5, and it will not be shown here. □

From Corollary 8, it is clear that the whole profit of the online marketplace will be lower than that under the centralized decision scenario, whether in the decentralized decision scenario or the promotion investment sharing scenario. However, sharing of the promotion investment does not necessarily improve the whole profit, and promotion investment sharing can only yield more overall profit for the online marketplace than decentralized decision-making if the sum of the rate of the store and the content platform’s sharing of promotion investment with the e-commerce platform is small. The results show that, because both the store and the content platform share a certain amount of promotion investment of the e-commerce platform, the investment sharing scenario makes the burden on the store and the content platform higher if the share of both is too high. The promotion of the e-commerce platform is not for onsite or offsite demand, but simply a general promotion, which has a relatively small impact on the overall profit.

Figure 4 reveals the abovementioned differences in total profit. Although the above differences in the overall profit of the online marketplace increase with the spillover effect of the e-commerce platform’s promotion, the promotion investment sharing scenario is larger than the decentralized decision scenario for the entire online marketplace as long as the sum of the shares of promotion investment (${\rho }_r+{\rho }_p$) shared by the store and the content platform for the e-commerce platform remains relatively low. The management insight revealed by the above results is that stores should be careful in the implementation of cooperation in the process of onsite and offsite promotion in terms of sharing of promotion investment, so that one participant does not receive too high a share of promotional investment from the other participants in total, i.e., the sharing of promotion investment should be kept within a relatively small range.

5. Conclusions

With the gradual maturity of the e-commerce market, the demographic dividend of e-commerce platforms is gradually disappearing, thereby making the competition for onsite traffic more and more fierce among stores on e-commerce platforms. The competition for onsite traffic has caused some stores to shift their focus on traffic expansion to outside the e-commerce platform, while the rise of a large number of new content platforms in recent years has provided the possibility for stores to expand their traffic outside the site. In view of this, this paper constructs an onsite and offsite promotion decision model consisting of a store, an e-commerce platform, and a content platform, taking into account the reality that stores on e-commerce platforms promote onsite while promoting offsite with the help of content platforms. By analyzing the decentralized decision, centralized decision, and promotion investment sharing decision scenarios, the optimal promotion investments and profits of the participants were determined. Then, by examining the investment decisions and profits of each participant in the different decision scenarios, it was found that centralized decision-making is the optimal decision form for the online marketplace as a whole, while promotion investment sharing does not necessarily lead to an increase in profits for each participant at the same time, and only when the sum of the investment sharing rates of the store and the content platform for the e-commerce platform is relatively small does the promotion investment sharing scenario lead to higher total profit than the decentralized decision scenario.

Although this paper provides some inspiration for the co-opetition among participants in the process of onsite and offsite promotion, there are some limitations of the article to be improved in the following respects in the future: (1) To explore the problem of determining the optimal referral fee rate by using the referral fee rate of the e-commerce platform as a decision variable. (2) Introducing the relevant research problem into the horizontal competition environment to explore the promotion decision problem under the competition among multiple e-commerce platforms, multiple stores, and multiple content platforms.