Live-Streaming Commerce in the Supply Chain with Equity Cooperation: Independent or Cooperative?

Abstract

Live-streaming commerce (LSC) has been adopted by an increasing number of supply-chain enterprises to enhance their market competitiveness. However, the question of who will lead live-streaming e-commerce in the supply chain (SC-LSC) is a key issue, especially when there is equity cooperation between upstream and downstream enterprises. Three main SC-LSC models are examined: independent SC-LSC run by manufacturers, independent SC-LSC run by retailers, and cooperatively run SC-LSC. Then, a novel LSC demand function composed of online popularity, price discount and sales conversion rate is proposed. Furthermore, four scenarios have been comprehensively investigated considering whether there is an online-to-offline drainage effect and whether there is equity cooperation. Regardless of the scenario, having both parties reach an agreement on a given SC-LSC model is difficult, and even equity cooperation cannot promote SC-LSC cooperation. In most cases, manufacturers tend to offset the losses caused by the drainage effect by adopting high wholesale prices, which will in turn exacerbate retailers’ resistance to SC-LSC. These findings provide insight into how LSC is modeled and how LSC can be better implemented in various types of supply chains such as that of Gree Electric.

1. Introduction

Compared with conventional e-commerce, live-streaming commerce (LSC) is becoming a very popular online channel due to its ability to improve consumer experience, enhance consumer trust, and increase sales with relatively low operating costs and technical barriers ([1,2]). In China, a large number of anchors and consumers are active every day on platforms such as Taobao, Douyin, Kuaishou, and JD.com. Especially in recent years, COVID-19 has further increased the popularity of LSC by restricting people’s travel and the normal operation of offline stores ([3]). Therefore, in addition to well-known third-party anchors such as Li Jiaqi, many supply-chain enterprises, including Gree Electric, Midea, TCL, VIVO, and New Oriental, have joined the LSC market. Among these enterprises are not only brand manufacturers (such as “Gree Electric, Zhuhai, China”) but also a large number of distributors and retailers (such as Gree Electric retailers and “VIVO, Dongguan, China” retailers). They no longer pay exorbitant fees to cooperate with third-party anchors but choose to conduct LSC independently and achieve success. Furthermore, as with Gree Electric and Gree Electric retailers, one of them is responsible for live streaming and the other is responsible for order distribution, cooperating to complete the SC-LSC. Obviously, the establishment of independent or cooperative SC-LSC is changing the way the supply chain operates, but it has not received enough attention.

Intuitively, if SC-LSC is profitable, both the manufacturer and retailer will tend to independently develop SC-LSC to maximize their profits. However, in the supply chain, the retailer’s live-streaming products need to be obtained wholesale from the manufacturer, and the wholesale price depends on the manufacturer. Due to differences in profit functions, their attitudes toward operational investments and product discounts in SC-LSC are also different. In addition, if retailers hold shares in manufacturers such as Gree Electric retailers, they can share in part of the SC-LSC profits from manufacturers. Under these conditions, it is uncertain whether independent or cooperative SC-LSC is more suitable for the supply chain. Moreover, SC-LSC often uses high price discounts to attract consumers, which may lead to the loss of some consumers from offline channels and cause resistance from retailers ([4,5]). Thus, how should supply-chain enterprises such as Gree Electric maintain cooperation with downstream retailers while developing SC-LSC? Which SC-LSC model is better for consumers and the entire supply chain? How should these enterprises make decisions on price, sales volume and operating costs of SC-LSC?

Unfortunately, these questions cannot be answered based on the existing literature. These questions are undoubtedly crucial for supply chains trying to transition from offline to online, as well as for the development of the LSC market. In fact, most of the current research on LSC explores consumer behavior and the influencing factors in LSC ([6,7]). However, from the perspective of conventional supply-chain modeling, it is not clear how LSC is priced and how its demand function should be expressed. The number of fans, online popularity, price discounts, and sales conversion rate, which are new important variables introduced by LSC, will undoubtedly change the modeling of and decision-making in the supply chain ([3,8]).

The main contributions of this study are as follows: (1) the pricing differences between LSC and conventional e-commerce are addressed, and the demand function of LSC is designed considering online popularity, price discount and sales conversion rate; (2) three SC-LSC models, including independent SC-LSC run by manufacturers, cooperative SC-LSC shared between manufacturers and retailers, and independent SC-LSC run by retailers, are proposed and investigated; (3) four common real-world scenarios are comprehensively discussed and compared, including a scenario without a drainage effect and without equity cooperation; a scenario without a drainage effect and with equity cooperation; a scenario with a drainage effect and without equity cooperation; and a scenario with a drainage effect and with equity cooperation. These comparisons expand new knowledge about LSC pricing and operational decisions and will help supply-chain enterprises gain a competitive advantage in the LSC market.

This study mainly draws on real cases of live-streaming supply chains in China, such as those of Gree Electric and Xiaomi, including channel conflicts caused by LSC and conflicts of interest caused by equity cooperation. Therefore, the modeling methods and conclusions of this study are applicable to the vast majority of traditional supply chains that are trying LSC, such as the manufacturing and apparel industries. The purpose of this study is to find the model that most effectively supports the supply chain in implementing LSC.

The remainder of this paper is organized as follows: after relevant studies are reviewed in Section 2, the model notations and assumptions are presented in Section 3. Section 4 and Section 5 examine the optimal SC-LSC model under scenarios that differ with regard to drainage effect and equity cooperation. Some important propositions are obtained through comprehensive discussion and numerical analysis. Finally, we draw conclusions and present managerial insights in Section 6. All of the proofs are presented in Appendix A.

2. Literature Review

Unlike conventional e-commerce, LSC is regarded as a special type of e-commerce with green advertising and social networking ([9]). Online sellers can increase sales conversions by performing, explaining, and persuading customers to buy their products through live chat ([10,11]). Therefore, consumers can obtain more detailed product information from LSC than from conventional e-commerce ([2]). In fact, there are differences in each consumer’s motivation to view live-stream broadcasts, and this difference even covers gender differences ([12,13]). Luo et al. (2024) [4] adopted the elaboration likelihood model of persuasion to examine how live-streaming influences customers’ engagement and impulse-buying behavior, as moderated by their deal-proneness. Xu et al. (2024) [7] explored the influence of anchor reputation on product sales and innovatively considered both historical and real-time reputation signals. Xie et al. (2022) [14] identified a value-based marketing framework of tourism e-commerce live streaming using a mixed-method approach following value-based marketing theory. Zheng et al. (2022) [15] explored how customer engagement behaviors are related to both purchase intention and customer acquisition. Chen et al. (2022) [6] proposed a research model with an information-signals mechanism and an automatic-habit mechanism. Park and Lin (2020) [16] developed and tested an integrative model of internet celebrity endorsement by investigating congruence effects on live-streaming viewers. Ma et al. (2022) [17] constructed a conceptual model of purchase hesitation based on the theory of telepresence and trust from the perspective of participants’ interaction in order to study the factors influencing consumers’ purchase hesitation in the context of live streaming. These technical factors, social factors, personal factors, and platform factors that affect live e-commerce can help us to better understand consumers’ cognition, behavior, and motivation ([18,19]).

However, no matter how many factors affect them, enhancing consumer engagement and creating consumer loyalty are critical to the success of LSC ([20,21]). Lu and Yun (2024) [22] examined live-streaming strategies for vertically differentiated firms that sell products of high or low quality. Zhang et al. (2023) [23] investigated the influence patterns of environmental stimuli on consumers’ intention to purchase offline/in-store after watching an in-store live-stream session. Hao and Yang (2023) [24] established live streaming, selling models with two sales formats (resale and agency sale) and three pricing strategies (same high, same low and differential pricing). Chen and Lu (2021) [25] proposed that broadcasters’ physical characteristics conveyed through vicarious product trials and values shared via instant interaction as two signals that can help reduce product uncertainty and cultivate trust for consumers with similar physical traits and values. Guo et al. (2021) [20] examined the factors that impact the success of LSC by exploring the relationships among factors influencing customer trust. Lin et al. (2021) [1] estimated a panel vector autoregression model on data at the minute level from 1450 live streams to account for the possibility that broadcaster emotion, viewer emotion, and viewer activities influence each other. Some scholars have designed recommendation models to capture the preference matching between anchors and viewers to enhance user experience ([26]). Obviously, these studies focus mainly on the influencing factors and consumer motivation in LSC. The operation management and pricing decisions made in LSC, especially how they are implemented and cooperated in the supply chain, are rarely covered.

Actually, how LSC is priced and modeled is unclear. Generally, we are accustomed to using a linear demand function to represent the consumer’s demand function; that is, in a fixed market, a clearing price is used to achieve product sales ([27,28]). Such demand functions are also widely used in product selection and dual-channel competition ([29,30]). However, for LSC, the number of online consumers in one or several live streams is limited and cannot represent the entire potential market at all. This means that the pricing strategy in LSC is not aimed at product clearance but seeks to increase the popularity of the live-streaming room ([31,32]). Moreover, the conventional demand function considers only the price factor and does not consider other factors, such as the sales conversion rate. In other words, as long as the price is right, then by default all consumers will buy the product. Obviously, this is completely inconsistent with reality, especially in LSC ([11,33]). In a difference from conventional e-commerce and offline channels, the key to LSC success is to increase the sales conversion rate by improving the consumer experience and creating trust ([5,19]). In addition, the main purpose of SC-LSC is to acquire new consumers from external competitors rather than just to engage in online and offline stock competition, as assumed by conventional competitive demand functions ([34,35]). Therefore, it is necessary to redesign the demand function to create one that is closer to modeling the real business according to the characteristics of LSC, considering factors such as online consumers, price discounts, and sales conversion rates.

Finally, this study is also related to equity cooperation in supply chains. Equity cooperation refers to deepening the relationship between upstream and downstream enterprises in the supply chain through mutual shareholding ([36]). For example, Gree Electric retailers obtain preferential sales rights by taking shares in Gree Electric. Therefore, equity cooperation can essentially be regarded as a special revenue-sharing contract to achieve supply=chain coordination ([37]). The impact of equity cooperation on SC-LSC is undoubtedly not negligible but has not been addressed.

3. Notation and Assumptions

This study considers a supply chain consisting of a manufacturer and a retailer. If this supply chain mainly sells products or services through LSC rather than through traditional offline stores or e-commerce platforms, then it can be called SC-LSC; examples include “Gree Electric, Zhuhai, China”, “Vancl, Beijing, China”, and “East Buy Holding Limited, Beijing China”. This means that sales volume and pricing issues in traditional channels do not need to be considered in our models. In general, there are three main SC-LSC models: independent SC-LSC run by manufacturers (such as Midea, TCL), independent SC-LSC run by retailers (such as “OPPO, Dongguan, China” and “VIVO, Dongguan, China” sellers on “Taobao, Hangzhou, China”, “JD, Beijing, China”, etc.), and cooperative SC-LSC run by collaboration between the manufacturer and the retailer (for example, Gree Electric is responsible for SC-LSC, while the retailers complete the order delivery), as shown in Figure 1.

Then, how will these supply-chain enterprises choose the SC-LSC model? How are these live-streaming products priced and sold? What is the essential difference between LSC and conventional e-commerce? As shown in Figure 2, many studies assume that both manufacturers and retailers in the supply chain can make discretionary decisions about retail prices ([38,39]). However, in reality, this is often not the case. Especially in the LSC market, the vast majority of products, such as food, home appliances, and cosmetics, are perfectly competitive markets rather than monopolistic markets. Thus, both retailers and competitors outside the supply chain face the same general price. At this general price, the supply chain and competitors’ sales depend on their brand, product quality and marketing services ([40]). Wholesale prices are used by manufacturers to coordinate their distribution of profits with retailers.

Although a large number of consumers are now switching from offline channels to online channels, conventional e-commerce often suffers from unsatisfactory sales conversion rates due to lack of product experience and consumer trust ([25,41]). SC-LSC can effectively improve the sales conversion rate through live-streaming interaction. Assuming that the SC-LSC sales conversion rate is $t$ (lower than the sales conversion rate of offline channels but higher than that of conventional e-commerce), the corresponding SC-LSC cost function is ${\displaystyle \frac{1}{2}}k{t}^2$ ([8,42]), where $k$ is the SC-LSC cost coefficient related to labor cost, publicity cost, and operating cost ([9]). The production cost of the SC-LSC product is $c$.

At the general price $p$, the number of online consumers in the live-streaming room is $m$ and depends on the number of fans and the brand appeal of the enterprise. Price discount rate $r$ is adopted to lure price-sensitive consumers into the live-streaming room (so, the price of the LSC product is $(1-r)p$). The greater the price discount, the more consumers will be attracted into the live-streaming room (some anchors even claim the lowest price in the entire market). This factor is represented by $br$, where $b$ is the LSC consumer’s sensitivity to price discounts ([3]). In addition, the greater the initial number of online consumers $m$, the more new consumers the same price discount can attract due to the publicity effect of word of mouth ([43]). Therefore, we can obtain the actual number of online consumers in the live-streaming room, that is, $(1+br)m$; among them, $t$ proportion of consumers will place orders. For simplicity, we assume that each consumer places only one order. Then, the LSC sales are $q=(1+br)mt$. In addition, we can use $V=rq$ to measure the total benefits that consumers obtain in the LSC, that is, the price concession of manufacturers.

Generally, the consumers attracted by LSC through price discounts come mainly from competitors, but there are also some brand advocates who will switch from offline channels to live-streaming rooms for preferential prices. This phenomenon can be named the “drainage effect” ([44]). If the drainage coefficient is $\mu$, $r\mu$ consumers will switch from offline channels to live-streaming rooms. Obviously, the drainage effect will lead to the loss of consumers and profits from offline channels, so offline retailers tend to boycott independent SC-LSC run by manufacturers. Conventional models use pricing to counter each retailer’s drainage effect; that is, they use competitive demand functions. In addition to efforts in pricing (retail price, discount rate, and wholesale price) and sales conversion rate, this study also considered the choice of SC-LSC models to deal with the drainage effect.

However, with the rapid development of the financial industry, an increasing number of enterprises now hold shares in each other ([45]). For example, Gree Electric retailers held an 8.91% stake in the upstream manufacturer Gree Electric in 2022. Equity cooperation can share the benefits of LSC and achieve supply-chain coordination to a certain extent. This study assumes that the ratio of the retailer’s shareholding in the manufacturer is $\lambda$, $0\le \lambda <1$. If $\lambda$ = 0, there is no equity cooperation between the manufacturer and the retailer; if $\lambda$ = 1, the two parties are wholly owned subsidiaries, which can be regarded as a special independent SC-LSC model run by the manufacturer or retailer. The $m$, $r$, and $s$ indices represent the manufacturer, the retailer, and the entire supply chain, respectively.

4. SC-LSC without a Drainage Effect

This section first considers the scenario wherein SC-LSC has no drainage effect on offline channels, that is, SC-LSC consumers mainly come from other competitors outside of the supply chain ([46]). In other words, the number of brand loyalists who switch from offline to live-streaming rooms is negligible. This scenario is very consistent with the current offline consumption situation. For example, due to the impact of COVID-19, much offline consumption was restricted and many offline stores had to close, so only online channels could be considered. Then, the profit functions of the manufacturer and retailer in SC-LSC are as follows.

For the independent SC-LSC run by the manufacturer without a drainage effect (Model 1),

$${\pi }_{m1}^N=\left(1-\lambda \right)\left\{(1+br)mt\left[(1-r)p-c\right]-{\displaystyle \frac{1}{2}}k{t}^2\right\},$$

(1)

$${\pi }_{r1}^N=\lambda \left\{(1+br)mt\left[(1-r)p-c\right]-{\displaystyle \frac{1}{2}}k{t}^2\right\}.$$

(2)

For the cooperative SC-LSC without a drainage effect (Model 2),

$${\pi }_{m2}^N=\left(1-\lambda \right)\left[(1+br)mt(w-c)-{\displaystyle \frac{1}{2}}k{t}^2\right],$$

(3)

$${\pi }_{r2}^N=\left(1+br\right)mt\left[\left(1-r\right)p-w\right]+\lambda \left[(1+br)mt(w-c)-{\displaystyle \frac{1}{2}}k{t}^2\right].$$

(4)

For the independent SC-LSC run by the retailer without a drainage effect (Model 3),

$${\pi }_{m3}^N=\left(1-\lambda \right)(1+br)mt(w-c),$$

(5)

$${\pi }_{r3}^N=\left(1+br\right)mt\left[\left(1-r\right)p-w\right]-{\displaystyle \frac{1}{2}}k{t}^2+\lambda (1+br)mt(w-c).$$

(6)

In Equations (1)–(6), the profit functions of both parties include two parts: the SC-LSC profit and the profit of equity cooperation. Regarding Model 2, this is one of Gree Electric’s current practices; that is, the manufacturer conducts LSC sales but does not ship these LSC orders directly to consumers. Instead, they are still shipped by retailers in various regions. In this scenario, the manufacturer uses its own live-streaming advantages to help the entire supply chain increase sales but does not abandon downstream retailers. For Gree Electric, this is mainly because they still need these retailers to complete product delivery, installation, and after-sales service. Therefore, in Model 2, manufacturers need to pay the conversion effort cost, while retailers obtain retail profits and equity income.

In Model 1, the manufacturer is responsible for making decisions on price discount rates $r$ and sales conversion rates $t$. Obviously, this is a problem concerning the maximum value of the $(r,t)$ binary function. The only stationary point (${\displaystyle \frac{1}{2}}(\theta -{\displaystyle \frac{1}{b}}),{\displaystyle \frac{{(1+b\theta )}^{2}pm}{4bk}})$ can be obtained from the first-order derivative ${\displaystyle \frac{\partial {\pi }_{m1}^N}{\partial r}}=0$, ${\displaystyle \frac{\partial {\pi }_{m1}^N}{\partial t}}=0$. Furthermore, according to the maximum judgment condition of the binary function ${\displaystyle \frac{{\partial }^2{\pi }_{m1}^N}{{\partial }^{2}r}}{\displaystyle \frac{{\partial }^2{\pi }_{m1}^N}{{\partial }^{2}t}}-{\left({\displaystyle \frac{{\partial }^2{\pi }_{m1}^N}{\partial r\partial t}}\right)}^2={\displaystyle \frac{1}{2}}{m}^2{p}^2{(1-\lambda )}^2{(1+b\theta )}^2>0$ and ${\displaystyle \frac{{\partial }^2{\pi }_{m1}^N}{{\partial }^{2}r}}=-2bmpt(1-\lambda )<0$, it can be known that this stationary point is the maximum point. Similarly, in Model 3, it is still a binary function maximum problem, but the decision maker changes from the manufacturer to the retailer. However, in Model 2, the manufacturer uses its own advantages to make decisions on sales conversion rates but the orders need to be completed by the retailer, so the price discount rate is decided jointly by both parties. Since ${\displaystyle \frac{{\partial }^2{(\pi }_{m2}^N+{\pi }_{r2}^N)}{{\partial }^{2}r}}=-2bmpt<0$, ${\displaystyle \frac{{\partial }^2{\pi }_{m2}^N}{{\partial }^{2}t}}=-k(1-\lambda )<0$, the corresponding maximum point can be obtained according to the first-order derivative ${\displaystyle \frac{\partial ({\pi }_{m2}^N+{\pi }_{r2}^N)}{\partial r}}=0, {\displaystyle \frac{\partial {\pi }_{m2}^N}{\partial t}}=0$.

It should be noted that this study does not define who occupies the leading position in the market, so the decisions of manufacturers and retailers are made simultaneously. The Nash equilibrium solutions for each model are shown in

Table 1. Optimal solutions for the three models without the drainage effect.

Variable	Model 1	Model 2	Model 3
$r^N$	${\displaystyle \frac{1}{2}}(\theta -{\displaystyle \frac{1}{b}})$	${\displaystyle \frac{1}{2}}(\theta -{\displaystyle \frac{1}{b}})$	${\displaystyle \frac{1}{8}}\left(3\theta -{\displaystyle \frac{5}{b}}\right)$
$t^N$	${\displaystyle \frac{{(1+b\theta )}^{2}pm}{4bk}}$	${\displaystyle \frac{(w-c)(1+b\theta )m}{2k}}$	${\displaystyle \frac{9{(1+b\theta )}^{2}pm}{64bk}}$
$q^N$	${\displaystyle \frac{p{(1+b\theta )}^3{m}^2}{8bk}}$	${\displaystyle \frac{(w-c){(1+b\theta )}^2{m}^2}{4k}}$	${\displaystyle \frac{27p{(1+b\theta )}^3{m}^2}{512bk}}$
${\pi }_m^N$	${\displaystyle \frac{{(1-\lambda )(1+b\theta )}^4{p}^2{m}^2}{32b^{2}k}}$	${\displaystyle \frac{{(1-\lambda )(w-c)}^2{(1+b\theta )}^2{m}^2}{8k}}$	${\displaystyle \frac{27{(1+b\theta )}^4{p}^2{m}^2}{2048b^{2}k}}$
${\pi }_r^N$	${\displaystyle \frac{{\lambda (1+b\theta )}^4{p}^2{m}^2}{32b^{2}k}}$	${\displaystyle \frac{(w-c)\left\{\begin{array}{c}\left(1+b\right)p-\\ b\left[\left(2-\lambda \right)w-(1-\lambda )c\right]\end{array}\right\}{(1+b\theta )}^2{m}^2}{8bk}}$	${\displaystyle \frac{81{(1+b\theta )}^4{p}^2{m}^2}{8192b^{2}k}}$

Note: $\theta ={\displaystyle \frac{p-c}{p}}$.

.

In Model 3, there is a unique wholesale price that maximizes the manufacturer’s profit; that is, $w_3^N={\displaystyle \frac{(1+b)p+(3-4\lambda )bc}{4(1-\lambda )b}}$ ($\mathrm{a}\mathrm{c}\mathrm{c}\mathrm{o}\mathrm{r}\mathrm{d}\mathrm{i}\mathrm{n}\mathrm{g} \mathrm{t}\mathrm{o} {\displaystyle \frac{\partial {\pi }_{m3}^N}{\partial w}}=0$), while in Model 2, such a wholesale price does not exist and needs to be negotiated by both parties. Model 1 does not involve wholesale prices since orders are shipped directly to consumers by the manufacturer.

Table 1 shows that without the drainage effect, no matter which SC-LSC model is adopted, equity cooperation has no substantial impact on the price discount, sales conversion rate, or sales. However, only in Model 1 and Model 2 is there a certain impact on the profit distribution of both parties. By comparing the optimal solutions of the three models, some important conclusions can be drawn. We first examine the simplest scenario, that is, the one in which there is no drainage effect and no equity cooperation in the supply chain ($\lambda$ = 0).

Proposition 1. 

For the three models without drainage effects or equity cooperation,

(1)

$r_1^{NN}=r_2^{NN}>r_3^{NN}$.

(2)

When $c<w\le w_1^{NN}$, $t_1^{NN}>t_3^{NN}\ge t_2^{NN}$; when $w_1^{NN}<w\le w_0^{NN}$, $t_1^{NN}\ge t_2^{NN}>t_3^{NN}$; when $w_0^{NN}<w<p$, $t_1^{NN}>t_3^{NN}$.

(3)

When $c<w\le w_2^{NN}$, $q_1^{NN}>q_3^{NN}\ge q_2^{NN}$; when $w_2^{NN}<w\le w_0^{NN}$, $q_1^{NN}\ge q_2^{NN}>q_3^{NN}$; when $w_0^{NN}<w<p$, $q_1^{NN}>q_3^{NN}$.

(4)

When $c<w\le w_3^{NN}$, $V_1^{NN}>V_3^{NN}\ge V_2^{NN}$; when $w_3^{NN}<w\le w_0^{NN}$, $V_1^{NN}\ge V_2^{NN}>V_3^{NN}$; when $w_0^{NN}<w<p$, $V_1^{NN}>V_3^{NN}$.

(5)

When $c<w\le w_4^{NN}$, ${\pi }_{s1}^{NN}>{\pi }_{s3}^{NN}\ge {\pi }_{s2}^{NN}$; when $w_4^{NN}<w\le w_0^{NN}$, ${\pi }_{s1}^{NN}\ge {\pi }_{s2}^{NN}>{\pi }_{s3}^{NN}$; when $w_0^{NN}<w<p$, ${\pi }_{s1}^{NN}>{\pi }_{s3}^{NN}$.

Proposition 1 implies that when there is no drainage effect and no equity cooperation in the supply chain, the price discount, sales conversion rate, sales volume, consumer benefits, and SC-LSC profit of the entire supply chain under Model 1 are always the largest of those among the three models. In other words, in such a scenario, SC-LSC independently run by the manufacturer is the best choice. However, this may be resisted by retailers, making it difficult to make Model 1 enforceable in the supply chain, as shown in the following proposition.

Proposition 2. 

For the three models without drainage effects or equity cooperation:

(1)

When $c<w\le w_5^{NN}$, ${\pi }_{m1}^{NN}>{\pi }_{m3}^{NN}>{\pi }_{m2}^{NN}$, ${\pi }_{r3}^{NN}{\ge \pi }_{r2}^{NN}>{\pi }_{r1}^{NN}$.

(2)

When $w_5^{NN}<w\le w_6^{NN}$(Interval II), ${\pi }_{m1}^{NN}>{\pi }_{m3}^{NN}\ge {\pi }_{m2}^{NN}$, ${\pi }_{r2}^{NN}>{\pi }_{r3}^{NN}>{\pi }_{r1}^{NN}$.

(3)

When $w_6^{NN}<w\le w_7^{NN}$, ${\pi }_{m1}^{NN}>{\pi }_{m2}^{NN}>{\pi }_{m3}^{NN}$, ${\pi }_{r2}^{NN}\ge {\pi }_{r3}^{NN}>{\pi }_{r1}^{NN}$.

(4)

When $w_7^{NN}<w\le w_0^{NN}$(Interval IV), ${\pi }_{m1}^{NN}>{\pi }_{m2}^{NN}>{\pi }_{m3}^{NN}$, ${\pi }_{r3}^{NN}>{\pi }_{r2}^{NN}>{\pi }_{r1}^{NN}$.

(5)

When $w_0^{NN}<w<p$, ${\pi }_{m1}^{NN}>{\pi }_{m3}^{NN}$, ${\pi }_{r3}^{NN}>{\pi }_{r1}^{NN}$.

Proposition 2 (refer to Figure 3) suggests that the SC-LSC profits obtained by the manufacturer and retailer in Model 1 and Model 3 are fixed, but in the cooperative SC-LSC (Model 2), their profits will be subject to the manufacturer’s wholesale price. However, no matter how the wholesale price is adjusted, neither side can find a model that simultaneously maximizes their profits; that is, there is no equilibrium SC-LSC model in this scenario. For the manufacturer, independent SC-LSC (Model 1) is the best choice in any interval of wholesale price, while it is always the worst option for the retailer. However, for the retailer, when the wholesale price (Model 2) is in interval II or interval III, cooperative SC-LSC will be the best strategy. Otherwise, independent SC-LSC run by the retailer would be a wise choice.

In addition, Table 1 shows that the price discount, conversion rate, sales volume, and SC-LSC profit under the three models are all positively correlated with the product profit rate and the initial number of online consumers but negatively correlated with the SC-LSC cost coefficient. This means that manufacturers or streamers with high-margin products and large numbers of fans are more likely to succeed in SC-LSC. Otherwise, it would be wiser to seek cooperation with external anchors with more fans and lower operating costs.

Proposition 1 indicates that in a supply chain without a drainage effect or equity cooperation, Model 1 is the SC-LSC model that is most beneficial to consumer benefits and to profits throughout the entire supply chain. However, proposition 2 also proves that this model will suffer the fiercest resistance from downstream retailers, making it difficult to implement in the supply chain. Then, is there a suboptimal solution rather than an optimal solution that can be accepted by both parties and allow them to implement LSC in such a supply chain? Actually, when the wholesale price is in interval I or interval III, the two parties can reach a compromise in the form of Model 3 or Model 2, respectively, to reach an acceptable equilibrium. Although such compromises would be detrimental to the profits of manufacturers and consumers, it is entirely possible for forward-looking manufacturers to follow such a game equilibrium, considering the long-term benefits of LSC to supply-chain development. However, in other intervals, the conflicting interests of the two parties are large enough that such a suboptimal solution cannot be found.

The retailer cannot benefit from independent SC-LSC run by the manufacturer (Model 1) and will naturally resist. As Table 1 shows, equity cooperation can change both parties’ SC-LSC profits in Model 1 and Model 2. Then, does this mean that equity cooperation can help them find an equilibrium SC-LSC model that simultaneously maximizes their respective profits? As shown in Figure 4, equity cooperation greatly changes their game decision interval (compared to Figure 3). According to the critical point of the wholesale price (Model 2), we further divide equity cooperation into three subscenarios: “low-shareholding,” “medium-shareholding”, and “high-shareholding.” In reality, retailers tend to have a low stake in a manufacturer, typically no more than 50%. For example, at the beginning of 2022, Gree Electric distributors held only 8.34% of Gree Electric. Therefore, the focus should be on the low-shareholding scenario.

Table 2. The optimal model without the drainage effect but with equity cooperation.

Low Shareholding I	Medium Shareholding II	High Shareholding III
${\pi }_{m2}^{NE}>{\pi }_{m1}^{NE}>{\pi }_{m3}^{NE}$, ${\pi }_{r3}^{NE}{>\pi }_{r1}^{NE}>{\pi }_{r2}^{NE}$.	${\pi }_{m2}^{NE}>{\mathit{\pi }}_{\mathit{m}1}^{\mathit{N}\mathit{E}}>{\pi }_{m3}^{NE}$, ${\mathit{\pi }}_{\mathit{r}1}^{\mathit{N}\mathit{E}}{>\pi }_{r3}^{NE}>{\pi }_{r2}^{NE}$.	${\pi }_{m2}^{NE}>{\pi }_{m3}^{NE}>{\pi }_{m1}^{NE}$, ${\pi }_{r1}^{NE}{>\pi }_{r3}^{NE}>{\pi }_{r2}^{NE}$.
${\pi }_{m1}^{NE}>{\pi }_{m2}^{NE}>{\pi }_{m3}^{NE}$, ${\pi }_{r3}^{NE}{>\pi }_{r2}^{NE}>{\pi }_{r1}^{NE}$.	${\mathit{\pi }}_{\mathit{m}2}^{\mathit{N}\mathit{E}}>{\mathit{\pi }}_{\mathit{m}1}^{\mathit{N}\mathit{E}}>{\pi }_{m3}^{NE}$, ${\mathit{\pi }}_{\mathit{r}1}^{\mathit{N}\mathit{E}}{>\mathit{\pi }}_{\mathit{r}2}^{\mathit{N}\mathit{E}}>{\pi }_{r3}^{NE}$.	${\mathit{\pi }}_{\mathit{m}2}^{\mathit{N}\mathit{E}}>{\pi }_{m3}^{NE}>{\pi }_{m1}^{NE}$, ${\pi }_{r1}^{NE}{>\mathit{\pi }}_{\mathit{r}2}^{\mathit{N}\mathit{E}}>{\pi }_{r3}^{NE}$.
${\pi }_{m1}^{NE}>{\mathit{\pi }}_{\mathit{m}2}^{\mathit{N}\mathit{E}}>{\pi }_{m3}^{NE}$, ${\mathit{\pi }}_{\mathit{r}2}^{\mathit{N}\mathit{E}}{>\pi }_{r3}^{NE}>{\pi }_{r1}^{NE}$.	${\mathit{\pi }}_{\mathit{m}1}^{\mathit{N}\mathit{E}}>{\mathit{\pi }}_{\mathit{m}2}^{\mathit{N}\mathit{E}}>{\pi }_{m3}^{NE}$, ${\mathit{\pi }}_{\mathit{r}2}^{\mathit{N}\mathit{E}}{>\mathit{\pi }}_{\mathit{r}1}^{\mathit{N}\mathit{E}}>{\pi }_{r3}^{NE}$.	${\pi }_{m3}^{NE}>{\pi }_{m2}^{NE}>{\pi }_{m1}^{NE}$, ${\pi }_{r1}^{NE}{>\pi }_{r2}^{NE}>{\pi }_{r3}^{NE}$.
${\pi }_{m1}^{NE}>{\pi }_{m3}^{NE}>{\pi }_{m2}^{NE}$, ${\pi }_{r2}^{NE}{>\pi }_{r3}^{NE}>{\pi }_{r1}^{NE}$.	${\mathit{\pi }}_{\mathit{m}1}^{\mathit{N}\mathit{E}}>{\pi }_{m3}^{NE}>{\pi }_{m2}^{NE}$, ${\pi }_{r2}^E{>\mathit{\pi }}_{\mathit{r}1}^{\mathit{E}}>{\pi }_{r3}^E$.	${\pi }_{m3}^{NE}>{\pi }_{m1}^{NE}>{\pi }_{m2}^{NE}$, ${\pi }_{r2}^{NE}{>\pi }_{r1}^{NE}>{\pi }_{r3}^{NE}$.
${\pi }_{m1}^{NE}>{\mathit{\pi }}_{\mathit{m}3}^{\mathit{N}\mathit{E}}>{\pi }_{m2}^{NE}$, ${\mathit{\pi }}_{\mathit{r}3}^{\mathit{N}\mathit{E}}{>\pi }_{r2}^{NE}>{\pi }_{r1}^{NE}$.	${\mathit{\pi }}_{\mathit{m}1}^{\mathit{N}\mathit{E}}>{\pi }_{m3}^{NE}>{\pi }_{m2}^{NE}$, ${\mathit{\pi }}_{\mathit{r}1}^{\mathit{N}\mathit{E}}{>\pi }_{r2}^{NE}>{\pi }_{r3}^{NE}$.	${\pi }_{m3}^{NE}>{\mathit{\pi }}_{\mathit{m}1}^{\mathit{N}\mathit{E}}>{\pi }_{m2}^{NE}$, ${\mathit{\pi }}_{\mathit{r}1}^{\mathit{N}\mathit{E}}{>\pi }_{r2}^{NE}>{\pi }_{r3}^{NE}$.
${\mathit{\pi }}_{\mathit{m}1}^{\mathit{N}\mathit{E}}>{\mathit{\pi }}_{\mathit{m}3}^{\mathit{N}\mathit{E}}>{\pi }_{m2}^{NE}$, ${\mathit{\pi }}_{\mathit{r}3}^{\mathit{N}\mathit{E}}{>\mathit{\pi }}_{\mathit{r}1}^{\mathit{N}\mathit{E}}>{\pi }_{r2}^{NE}$.	${\mathit{\pi }}_{\mathit{m}1}^{\mathit{N}\mathit{E}}>{\pi }_{m3}^{NE}>{\pi }_{m2}^{NE}$, ${\mathit{\pi }}_{\mathit{r}1}^{\mathit{N}\mathit{E}}{>\pi }_{r3}^{NE}>{\pi }_{r2}^{NE}$.	${\mathit{\pi }}_{\mathit{m}3}^{\mathit{N}\mathit{E}}>{\mathit{\pi }}_{\mathit{m}1}^{\mathit{N}\mathit{E}}>{\pi }_{m2}^{NE}$, ${\mathit{\pi }}_{\mathit{r}1}^{\mathit{N}\mathit{E}}{>\mathit{\pi }}_{\mathit{r}3}^{\mathit{N}\mathit{E}}>{\pi }_{r2}^{NE}$.

Note: Parts marked in bold indicate models that can achieve perfect or acceptable equilibrium.

shows that in the scenario without the drainage effect, regardless of whether the shareholding ratio is high or low, equity cooperation makes it difficult to achieve an equilibrium SC-LSC model between the manufacturer and retailer. Only in the interval E2-F2 (as shown in Figure 4) can the two parties achieve a perfect equilibrium in Model 1. However, reaching such an equilibrium condition is a low-probability event in practice. In addition, when the shareholding ratio is moderate, it is very easy for both parties to reach an acceptable equilibrium in Model 1 or Model 2. In short, in a supply chain without a drainage effect, it is difficult for manufacturers and retailers to reach stable SC-LSC cooperation regardless of whether there is equity cooperation. The probability associated with independent SC-LSC is far greater than that associated with cooperative SC-LSC.

5. SC-LSC with Drainage Effects

Although SC-LSC mainly acquires new consumers from outside the supply chain, companies such as Gree Electric and Midea want to enter the LSC market and maintain the original offline channels to ensure product installation and after-sales service. This will cause some brand loyalists to switch from offline channels to live-streaming rooms; that is, it will create the drainage effect of SC-LSC. In 2022, Gree Electric retailers reduced their stake by 1.86% to protest Gree Electric because LSC led to the loss of a large number of offline consumers. In this section, we examine how the drainage effect affects their LSC decisions in the supply chain. The drainage effect coefficient is $\mu$, and the number of offline consumers lost is $r\mu$. Then, the new profit function of both parties can be obtained as follows:

For the independent SC-LSC run by the manufacturer with a drainage effect (Model 1),

$${\pi }_{m1}^D=\left(1-\lambda \right)\left\{-r\mu \left(w-c\right)+(1+br)mt\left[(1-r)p-c\right]-{\displaystyle \frac{1}{2}}k{t}^2\right\},$$

$${\pi }_{r1}^D=-r\mu \left(p-w\right)+\lambda \left\{-r\mu \left(w-c\right)+(1+br)mt\left[(1-r)p-c\right]-{\displaystyle \frac{1}{2}}k{t}^2\right\}.$$

For the cooperative SC-LSC with a drainage effect (Model 2),

$${\pi }_{m2}^D=\left(1-\lambda \right)\left[-r\mu \left(w-c\right)+(1+br)mt(w-c)-{\displaystyle \frac{1}{2}}k{t}^2\right],$$

$${\pi }_{r2}^D=-r\mu \left(p-w\right)+\left(1+br\right)mt\left[\left(1-r\right)p-w\right]+\lambda \left[-r\mu \left(w-c\right)+(1+br)mt(w-c)-{\displaystyle \frac{1}{2}}k{t}^2\right]$$

For the independent SC-LSC run by the retailer with a drainage effect (Model 3),

$${\pi }_{m3}^D=\left(1-\lambda \right)\left[-r\mu \left(w-c\right)+(1+br)mt(w-c)\right]$$

$${\pi }_{r3}^D=-r\mu \left(p-w\right)+\left(1+br\right)mt\left[\left(1-r\right)p-w\right]-{\displaystyle \frac{1}{2}}k{t}^2+\lambda \left[-r\mu \left(w-c\right)+(1+br)mt(w-c)\right]$$

The method for solving the above equilibrium solution is consistent with that described in Section 4.

Table 3. The impact of the drainage effect without equity cooperation.

Variable	Model 1	Model 2	Model 3
$∆r$	$-{\displaystyle \frac{2k(w-c)\mu }{{\left[\left(1+b\right)p-bc\right]}^2{m}^2}}$	$-{\displaystyle \frac{k(p-c)\mu }{bm(kp{t}_2^{NN}+A)}}$	$-{\displaystyle \frac{2k(p-w)\mu }{{\left[\left(1+b\right)p-bw\right]}^2{m}^2}}$
$∆t$	0	$-{\displaystyle \frac{kp{t}_2^{NN}-A}{2kp}}$	0
$∆q$	$-{\displaystyle \frac{(w-c)\mu }{2p}}$	${q_2^{DN}-q}_2^{NN}$	$-{\displaystyle \frac{(p-w)\mu }{2p}}$
$∆{\pi }_m$	$-\mu (w-c)(r_1^{NN}+{\displaystyle \frac{1}{2}}∆r_1)$	${{\pi }_m^{DN}-\pi }_m^{NN}$	$-\mu (w-c)(r_3^{NN}+r_3^{DN}+{\displaystyle \frac{1}{2b}})$
$∆{\pi }_r$	$-\mu (p-w)r_1^{DN}$	${{\pi }_r^{DN}-\pi }_r^{NN}$	$-\mu (p-w)(r_3^{NN}+{\displaystyle \frac{1}{2}}∆r_3)$

Note: $r_1^{NN}={\displaystyle \frac{1}{2}}\left(\theta -{\displaystyle \frac{1}{b}}\right)$, $r_3^{NN}={\displaystyle \frac{(b-1)p-bw}{2bp}}$, A=$\sqrt{kp\left[kp{(t_2^{NN})}^2-2(w-c)(p-c)\mu \right]}$.

summarizes the impact of the drainage effect on the SC-LSC without equity cooperation. The table shows that (1) the drainage effect has no impact on the sales conversion rate of an independent SC-LSC run by the manufacturer or retailer; (2) no matter what SC-LSC model is adopted, the drainage effect will be detrimental to both parties’ product discount, sales volume and profits; and (3) if the wholesale price is low, then the manufacturer is less negatively affected by the drainage effect, while the retailer and the entire supply chain are more negatively affected by the drainage effect. If the wholesale price is relatively high, then this is the case.

Generally, because offline channels are associated with a better product experience, the conversion rate of SC-LSC is lower than that of offline channels. The existence of the drainage effect means that a considerable proportion of SC-LSC sales come from offline channels rather than external markets. This inevitably leads to a decline in the comprehensive conversion rate of online and offline channels. In other words, the drainage effect creates internal friction in the supply chain and weakens the positive effect of LSC. Moreover, Model 2 has the most stringent requirements for the drainage effect. Once the corresponding threshold is exceeded (${\mu }_{limit}^2={\displaystyle \frac{\left[(b-1)p-bc\right](w-c)m^2}{k(p-c)}}$), the two parties will not be able to carry out SC-LSC cooperation.

Since the wholesale price can amplify or weaken the negative impact of the drainage effect, manufacturers can improve their SC-LSC profits by adjusting the wholesale price, as shown in the following proposition.

Proposition 3. 

(1) For Model 1, manufacturers are motivated to adopt low wholesale prices to reduce the negative impact of the drainage effect, which also takes a toll on retailer profits. (2) For Model 2, the manufacturer will adopt a relatively high wholesale price to maximize the SC-LSC profit of the entire supply chain while ensuring its own profit. (3) For Model 3, there is an optimal wholesale price to maximize the manufacturer’s SC-LSC profit.

Counterintuitively, Proposition 3 reflects that for the drainage effect, manufacturers do not promote SC-LSC cooperation by lowering wholesale prices to appease retailers. In other words, it is difficult for manufacturers to maintain the stability of the original offline channels and cooperate to carry out SC-LSC. This is why while Gree Electric claimed that they gave distributors more SC-LSC profits, distributors did not appreciate it and even withdrew from equity cooperation in protest.

Since there is no equity cooperation, in Model 1, the drainage effect will inevitably lead to the loss of profit for retailers, which will lead to strong resistance from retailers. Therefore, in this scenario, we only need to consider the feasibility of Model 2 and Model 3. Figure 5 shows that when the drainage effect is relatively mild and the wholesale price is high, it is most sensible for the two parties to cooperate with SC-LSC. Otherwise, independent SC-LSC is more suitable for retailers. If the drainage effect is too great, that is, most offline consumers will turn to the SC-LSC and retailers will be eliminated from the supply chain.

Table 4. The impact of the drainage effect, wholesale price, and equity cooperation on SC-LSC.

$(\mathit{\lambda }, \mathit{\mu }, \mathit{w}$)	Model	$\mathit{r}$	$\mathit{t}$	$\mathit{q}$	${\mathit{\pi }}_{\mathit{m}}$	${\mathit{\pi }}_{\mathit{n}}$	${\mathit{\pi }}_{\mathit{s}}$
S1(0.1, 0.05$m$, 1600)	Model 1	0.2017	0.0480	2897.0	964,860 *	−74,296	890,564 *
	Model 2	0.1514	0.0201	1010.9	127,478	405,233 *	532,711
	Model 3	0.1041	0.0277	1132.8	370,329	287,105	657,434
S2(0.1, 0.05$m$, 2000)	Model 1	0.1933	0.0479	2810.5	893,764 *	2656	896,420 *
	Model 2	0.1970	0.0475	2821.8	874,053	20,119	894,173
	Model 3	0.0205	0.0127	307.0	206,319	69,300 *	275,619
S3(0.1, 0.1$m$, 1600)	Model 1	0.1933	0.0479	2810.5	893,764 *	−248,637	645,127 *
	Model 2	0	0	0	0	0	0
	Model 3	0.0647	0.0256	844.1	257,323	206,955 *	464,278
S4(0.1, 0.1$m$, 2000)	Model 1	0.1763	0.0475	2623.9	760,678 *	−91,784	668,894 *
	Model 2	0.1680	0.0415	2545.8	575,350	100,254 *	675,605
	Model 3	0	0	0	0	0	0
S5(0.3, 0.05$m$, 1600)	Model 1	0.2017	0.0480	2897.0	750,447 *	140,118	890,564 *
	Model 2	0.1677	0.0214	1146.6	113,563	448,971 *	562,534
	Model 3	0.1219	0.0317	1408.5	360,251	379,465	739,716
S6(0.3, 0.05$m$, 2000)	Model 1	0.1933	0.0479	2810.5	695,149 *	201,270 *	896,420 *
	Model 2	0.1861	0.0526	1225.3	421,465	130,197	551,663
	Model 3	0.0587	0.0189	597.7	301,817	133,842	435,659
S7(0.3, 0.1$m$, 1600)	Model 1	0.1933	0.0479	2810.5	695,149 *	−50,023	645,127 *
	Model 2	0	0	0	0	0	0
	Model 3	0.0859	0.0299	1112.8	263,493	272,942 *	536,434
S8(0.3, 0.1$m$, 2000)	Model 1	0.1763	0.0475	2624.0	591,638 *	77,256	668,894 *
	Model 2	0.1921	0.0467	2730.3	549,326	97,269 *	646,595
	Model 3	0	0	0	0	0	0

Note: The numbers marked with “*” in the table indicate the most profitable model.

analyzes the superimposed impacts of drainage effects and equity cooperation. Since this scenario is too complicated, we can only perform numerical analysis according to the SC-LSC case of Gree Electric. Similarly, in any scenario, the manufacturer tends to choose Model 1, and only in Scenario 6 can both parties reach a game equilibrium in Model 1. Reaching a consensus on cooperative SC-LSC is difficult, and in some scenarios (such as scenario 3), Model 2 is not feasible. From the perspective of other indicators, such as supply-chain profit, price discount, and sales volume, Model 1 is the best choice for addressing drainage effects and equity cooperation. This also forces companies such as Gree Electric to accept the reality that cooperative SC-LSC is unsustainable. If they want to truly switch to online channels such as SC-LSC, they have to face the continued shrinking and resistance of offline channels.

6. Conclusions

This study investigates how manufacturers and retailers in the supply chain conduct LSC independently or cooperatively. Equity cooperation between the two parties and the drainage effect of SC-LSC on conventional offline channels have been comprehensively considered, including four scenario settings: a scenario without a drainage effect and without equity cooperation; a scenario without a drainage effect and with equity cooperation; a scenario with a drainage effect and without equity cooperation; and a scenario with a drainage effect and with equity cooperation. The difference between LSC and conventional e-commerce and the pricing strategy of LSC are also examined. Furthermore, a novel demand function composed of online consumers, price discounts and conversion rates was created for LSC. Some important conclusions and managerial insights are proposed, as follows.

Regardless of the scenario, it is difficult for both parties to reach an agreement on a given SC-LSC model. Although independent SC-LSC run by manufacturers has proven to be the best model in terms of the entire supply chain and consumer benefits, it will inevitably lead to resistance from retailers, especially when the drainage effect is significant. Counterintuitively, even equity cooperation cannot change this situation. A perfect equilibrium in Model 1 can be achieved only when the retailer has a high percentage of the manufacturer’s stake. However, for the two parties to cooperate in SC-LSC, an acceptable equilibrium can be achieved only when the manufacturer abandons independent SC-LSC. Therefore, it is very difficult for a company such as Gree Electric to establish a large presence in SC-LSC and maintain the stability of the original offline channels. For Gree Electric, the choice is between a full shift to online channels or the merger of downstream channels to make the retailers a subdivision of itself. Retailers can either improve their competitiveness in LSC or expand other brand channels to achieve diversified development.

The drainage effect will create internal friction and reduce the sales conversion rate of the entire supply chain. However, this study found that it is unrealistic to expect manufacturers to lower the wholesale price to make up for the losses caused by SC-LSC to retailers. In most cases, manufacturers tend to reduce the losses caused by the drainage effect by adopting high wholesale prices. Therefore, for retailers with weak bargaining power, it will be more difficult to obtain low-priced products in the LSC market. For these retailers, achieving transformation and improving irreplaceability is a new way of surviving. For example, Gree Electric and VIVO retailers can transform from their original incarnations as retail stores to experience stores. They can diversify their agency brands and provide better after-sales services such as installation, maintenance, and data upgrades.

It must be recognized that whether LSC is implemented independently or cooperatively, the key to the success of the supply chain in LSC is to have sufficiently high product margins, enough online consumers, and high enough operational efficiency. Therefore, the supply chain should focus more attention on the accumulation of brands and fans while continuously improving the cost performance of products. In addition, innovation in expression has also proven to be a low-cost and high-efficiency LSC mode, as in the phenomena of Gree’s “Meng Yutong” and New Oriental’s “Dong Yuhui”.

This study’s contribution is that it explores the possibility of LSC cooperation from the perspective of the supply chain. The proposed new demand function and three SC-LSC models can also help us better understand the implications of LSC. This study’s limitation is that we did not distinguish among the power structures and decision-making sequences of the two parties, nor did we compare the three SC-LSC models with external anchors. These factors will obviously enrich this research topic and expand our knowledge in the future. In addition, product quality, information asymmetry, and platform intrusion are all very important issues in LSC. Approaches to the question of how to design a reward-and-punishment mechanism to improve the operational efficiency of SC-LSC will be a very valuable research direction. Moreover, many supply chains currently have LSC channels, traditional offline stores, and online platform channels, so the conflict and coordination issues among these three channels will also be a very interesting topic.