From Clicks to Bricks: Offline Entry Strategies of E-tailers with Private Brands

From Clicks to Bricks: Offline Entry Strategies

of E-tailers with Private Brands

Master’s Thesis

Supply Chain Management

June 22, 2020

Student Name: Qingcheng Li Student Number: s3581837 E-mail: q.li.20@student.rug.nl

Abstract

This paper compares two different offline entry strategies of e-tailers with private brands. We develop a game-theoretical model to investigate a supply chain where e-tailers can enter the offline market by building their own offline stores or by selling their private brands to their offline competitors. Numerical experiments are conducted to examine how market factors influence e-tailers’ channel decisions when entering the offline market. The results show that the e-tailer is more likely to align with their offline competitors when the self-price sensitivity of the products is high. In contrast, when the competition between national brands and private brands or the competition between national brands sold by different retailers is high, e-tailers are more likely to build their own offline stores. We also find that building own offline stores is more likely to become the optimal channel choice of the e-tailer, when the consumers’ online preference is low, when the private brands are highly preferred by the consumers, or when the offline operating cost is close to the online operating cost. Moreover, we find that e-tailers are more likely to use the sharing channel if traditional retailers are willing to make concessions on revenue sharing rates.

Keywords: Private Brands; Offline Entry strategies; Game theory; Channel Selection; Competition

1 Introduction

Private brands are getting increasingly popular in the past two decades (Huang & Feng, 2020). It is reported that private brand sales in the U.S. market have grown from 129 billion US dollars in 2015 to 143 billion US dollars in 2018 (Nielsen, 2019). Many e-tailers such as Amazon, BarkBox and Boxed have introduced their own private brands. Meanwhile, e-tailers are also trying to expand their sales to the offline market, which is called the From-Clicks-to-Bricks strategy. (Jiang et al., 2019) When private brands can generate large profit, is it also profitable for these e-tailers to sell private brands to offline customers?

The research idea of this paper comes from different offline channel strategies taken by Amazon. It is reported that the online market only shared about 10 percent of the retail market in 2018 and most of the purchases still occur in brick-and-mortar stores (Statt, 2018). This is a huge incentive for Amazon to change from “clicks” to “bricks”. In 2010, Amazon started to cooperate with her offline competitors in the United States, such as Walmart, Target and Best Buy to enter the offline market with her best-selling private brand, Kindle (Kumar, 2016). Although some of these retailers terminated the partnership (e.g. Walmart in 2012), Amazon electronic products are still competing with famous national brands (e.g. Fire tablet vs Samsung Tablet, Kindle vs Kebo e-reader and Amazon Fire TV streamer vs Apple TV) in many offline stores in different parts of the world. (e.g. Best Buy in the United States and Media Markt in Germany). However, Amazon’s ambition is not just selling her private brands through competitors’ channel. Brick-and-mortar stores are built by Amazon to further reach offline consumers (Ingraham, 2016). In 2015, Amazon opened her first physical book store Amazon Books, where consumers can also find Amazon electronic devices (DePhills, 2018). In 2018, the first Amazon 4-star store opened in New York. Different from Amazon Books, A wider range of national brands that are best sellers on Amazon.com are provided along with Amazon electronic devices and another Amazon private brand AmazonBasics. This can help Amazon achieve the balance between online and offline shopping (Clough, 2020). Meanwhile, the strategy of using the competitors’ offline channel is not abandoned by Amazon. It is reported that Amazon plans to cooperate with Indian retail group to sell her private brand AmazonBasics to Indian offline market (Mukherjee, 2019).

When entering the offline market, most of the e-tailers will choose from the aforementioned offline channel strategies. In China, e-tailers such as JD.com and Xiaomi Youpin have already opened offline stores for their private brands (Fung Business Intelligence, 2019). Online pet products retailer BarkBox in the United States is selling its private brand to another retail giant Target’s brick-and-mortar stores to better reach offline customers (DeBaun, 2017). Thus, it is interesting to model this kind of supply chain and study the optimal channel strategy in different market environments.

Most of the previous research supports e-tailers’ introduction of offline channels. Bell et al. (2018) find that establishing offline showrooms can benefit online retailers by increasing their sales. Dawes and Nenycz-Thiel (2014) conduct an analysis on brand data in 10 categories in UK market and report that most of the consumers using an online channel of a particular retailer also shop in the retailer’s offline stores. This can be an incentive for online retailers to embrace complete offline stores instead of showrooms. Zhang et al. (2017) conclude that when consumers’ acceptance of online channels is medium, the retailer should introduce both offline and online channels. However, few studies consider the possibility of selling private brands to com-petitors. Moorthy et al. (2018) consider the use of competitor’s channel from manufacturer’s perspective and find that only one manufacturer would use a competitor’s channel under this situation. Nevertheless, their model solely focuses on the interaction between manufacturers with their own outlets and the interaction between different retailers are not investigated. Besides that, Jiang et al. (2019) provide empirical evidence that it is beneficial for online retailers to sell their private brands to offline competitors. Overall, how online retailers choose from these two strategies is still missing in the literature. This paper will use game-theoretic models to find the optimal channel strategy for those e-tailers who want to introduce their private brands to the offline market. The research question of this paper is then addressed as follows:

What is e-tailer’s optimal channel strategy to maximize her profit, when her private brands can be sold in her own offline stores or introduced to her offline competitors?

To better answer the research question, the following sub-question is addressed:

How do market factors influence the e-tailer’s channel decisions?

The contribution of this paper is twofold. Firstly, this paper examines the interaction between retailers under a unique supply chain structure with multiple channels and multiple brands. A linear demand model is developed for this kind of supply chain structure. To the best of our knowledge, it is the first research that models the channel sharing between a e-tailer with a private brand and a traditional retailer. Secondly, the impact of different market factors on e-tailers’ channel decisions when entering the offline market is analyzed through numerical experiments. The findings can help e-tailers making decisions in different market environments.

The remainder of the paper will be structured as follows. In section 2, literature about channel competition and private brand strategy will be reviewed. In section 3, a game-theoretic model will be created. After that, the model will be analyzed in section 4. Numerical experiments will be conducted in section 5, and conclusions will be drawn in section 6.

2 Theoretical Background

This paper is about online retailers’ optimal channel strategy for their private brands. The first stream related to this paper is channel selection and competition and the second one is the private brand strategy. In this section, the theoretical background of the current paper is provided.

2.1 Channel Selection and Competition

There is plenty of literature about channel selection. Most of which is about the manufacturer’s channel strategy, especially the introduction of online direct channel (e.g. Cai, 2010; Chen et al., 2008 and Cattani et al. 2006). In contrast, some studies focus on pure horizontal channel competition from the perspective of retailers. Among these studies, offline retailer’s online channel introduction is a special research theme. Liu et al. (2006) investigate when can an incumbent offline retailer use online channel introduction to deter the market entry of an e-tailer. Bernstein et al. (2008) study whether online entry benefits offline retailer under an oligopoly setting. They find that when competing with outside goods, entering the online market will benefit offline retailers. Nevertheless, little attention is paid to the e-tailer’s offline channel expansion in previous research.

Some studies consider both horizontal channel competition and vertical channel competition with man-ufacturer online direct channel introduced. Hsiao and Chen (2014) divide the consumer market into two segments and find that under retailer competition, symmetric retailers may adopt asymmetric channel struc-tures. Wang et al. (2016) and Zhang et al. (2017) use retailer-led Stackelberg game to analyze channel structure. Wang et al. (2016) conclude that operating both online and offline channel is an optimal choice for both retailer and manufacturer only when the operation costs of the two channels are close to each other. Zhang et al. (2017) find that the retailer and manufacturer’s channel structure would mainly depend on the consumers’ acceptance of different channels. Pure online channel should be applied when consumer acceptance of online channel is high and dual channel strategy should be used when the acceptance is medium. In this paper, we will focus on the e-tailer’s offline channel strategy and the manufacturer’s direct channel is not considered. Moreover, this paper models the use of a competitor’s channel, which is not considered in previous literature except Moorthy et al. (2018) and Bernstein et al. (2008). Moorthy et al. (2018) conduct their study from the manufacturer’s perspective. They find that a manufacturer would welcome her competitor to use her channel because of the showroom effect and the manufacturer with a weaker brand is more likely to introduce a sharing channel. Bernstein et al. (2008) considered the possibility of an offline retailer’s cooperating with an e-tailer to get online presence. Till now, models for online retailer’s sharing channel is still missing in the literature. To fill the gap in the literature, we will compare retailers’ self-built channel and sharing channel with their competitors. Table 1 summarizes previous theoretical modeling research on channel selection and competition.

T able 1: Summary of quan titativ e mo deling researc h on channel selection and comp etition A uthor(s) Supply chain structure Comp etition mo dels Demand functions Decision v ari-ables P arameters studied Researc h themes Cai (2010) A man ufacturer with online di-rect channel, one single-c hannel retailer or tw o single-c hannel retailers Man ufacturer-led Stac k elb erg game Customer Utilit y framew ork Con tract ad-justmen t factor (equiv alen t to wholesale price), Retail Price Relativ e channel p ow er param-eter,Supplier negotiation p ow er, Rev en ue sharing rate Study the optimal channel strat-egy in a supply chain with and without supply chain co ordina-tion Chen et al. (2008) A man ufacturer with online direct channel and a retailer Man ufacturer-led Stac k elb erg game Sto chastic demand considering differen t customer choice and av ailabilit y-based comp etition Wholesale price, deliv ery lead time, service lev el, retail price, av ailabilit y lev el Direct channel cost, retailer incon v enience, unit pro duction cost, pro duct value, mark et size Study man ufacturer’s optimal channel choice under differen t channel en vironmen t Zhang et al. (2017) A man ufacturer and a retailer Retailer-led Stac k elb erg game Utilit y maximiza-tion, piecewise-linear demand Retailer profit margin, whole-sale price, retail price Pro duction cost, consumption value, customer acceptance of the online channel, purc hase cost Study retailer and man ufac-turer’s optimal pricing strategy under differen t channel stratures Liu et al. (2006) An incum b en t of-fline retailer and an e-tailer Hotelling mo del Customer utilit y framew ork Retail price, whether or not to op en online stores Cost of op ening online stores, transp ortation cost, out-of-p o ck et costs parameter Study the optimal channel strat-egy of an incum b en t offline e-tailer to deter the en try of an e-tailer. Bernstein et al. (2008) Multiple retailers Oligop oly mo del. All retailers sim ul-taneously set their prices Multinomial logit (MNL) demand mo del Offline and retail price Consumer valuation, Online-offline channel valuation differ-ence parameter, online channel’s impact on consumer valuation Study whether it is b eneficial for the offline retailer to in tro duce an online channel W ang et al. (2016) One man ufacturer with online di-rect channel and one m ultic hannel retailer Retailer-led Stac k elb erg com-p etition Linear demand func-tion Retail price, di-rect selling price and wholesale price Unit channel op erating cost, Pro duct preference, channel preference, channel mismatc h parameter, pro duct differen tia-tion Study the optimal channel choice for b oth retailer and man ufacturer considering channel op erating costs

Hsiao and Chen (2014)

2.2 Private Brand Strategy

Previous research in the stream of private brand strategy studies the competition between private brand and national brand from the perspective of price competition (e.g. Raju at el. 1995; Kurata et al. 2007 and Ru et al. 2015) and the perspective of quality competition and private brand positioning (e.g. Sayman et al. 2002; Choi and Coughlan 2006 and Chung et al. 2017). Li et al. (2018) consider the impact of both customer return policy and supply chain coordination on the private brand introduction.

Some previous research investigates whether the introduction of the private brand will harm the man-ufacturer. Raju at el. (1995) finds that the private brand introduction will harm the manufacturer under manufacturer-led Stackelberg game while Ru et al. (2015) find that the manufacturer may benefit from the private brand introduction under retailer-led Stackelberg. In contrast, some studies consider the impact of retailer competition on the private brand strategy. As is pointed out by Groznik and Heese (2010), when retailers are competing with each other, they will fall into a “chicken game” regarding the private brand introduction. This makes randomized private brand introduction strategy more likely to occur. Mehra et al. (2018) point out that private brand introduction can be a strategy of an offline retailer to counter the show-room effect caused by powerful online retailers. Choi and Fredj (2013) consider both product differentiation and store differentiation in their model and find that less brand differentiation between private brands and national brands will benefit the retailers. Corstjens and Lal (2000) conclude that the introduction of quality private brands can benefit both the competing retailers by increasing store royalty and store differentiation. Jin et al. (2017) study retailers’ incentive of introducing private brands under different wholesale price setting scheme. They find that if a retailer has already introduced a private brand, and the manufacturer chooses to set a uniform wholesale price for the retailers, the retailer without private brand would not introduce private brands.

Despite the vast literature in the private brand introduction, little research considers the private brand’s channel strategy. As is defined in the literature, private brand is “the only brand for which the retailer must take on all responsibility-from development, sourcing and warehousing to merchandising and marketing.” (Dhar & Hoch, 1997. P. 208). The channel choice of private brands seems to be limited based on the definition. However, as mentioned in the previous section, private brands may be distributed by the retailer’s competitor in practice. This challenges the definition of private brand and provides a new supply chain structure. The empirical research conducted by Jiang et al. (2019) analyzes data from an e-tailer in South Korea. The findings of their study show that selling products to offline competitors can have a positive effect on the online channel of the e-tailer. Thus, instead of studying whether or not to introduce a private brand, this paper will focus on the channel strategy of an e-tailer who has already introduced a private brand in her online channel. Table 2 is a summary of theoretical modelling research on private brand strategy.

T able 2: Summary of q uan titativ e mo delling researc h on priv ate brands strategy A uthor(s) Supply chain structure Comp etition mo dels Demand functions Decision v ari-ables P arameters studied Researc h themes Ra ju et al. (1995) T w o man ufacturer and one retailer Man ufacturer-led Stac k elb erg game Linear demand func-tions Retail price, wholesale price Cross-price sensitivit y, base de-mand of priv ate brand Study the optimal priv ate brand in tro duction strategy under dif-feren t cross-price sensitivit y pa-rameters Kurata et al. 2007 A national brand man ufacturer with direct chan-nel, a priv ate brand man ufac-turer and a chain retailer Nash pricing game Linear demand mo del The wholesale price and retail price of the pri-vate brand and national brand in b oth direct stores and chain stores P oten tial mark et size, self-price sensitivities and cross-price sen-sitivities. Study the impact of mark et-ing decisions on proits and the supply chain co ordination when b oth direct channel and priv ate brands are in volv ed R u et al. (2015) A man ufacturer and a retailer Retailer-led Stac k elb erg game Utilit y Maximiza-tion, piecewise-linear demand Wholesale price, retail price, retail markups Store brand qualit y, unit pro-duction cost, consumer valua-tion Study the priv ate brand in-tro duction of a p ow er retailer (retailer-led Stac kelb erg)

Groznik and Heese (2010)

2.3 Summary of Theoretical Background

To sum up, both the e-tailer’s offline entry strategy and the channel strategy for private brands seem to be new to the literature. This paper will study the impact of the differentiation between private brands and national brands on the offline entry strategy of an e-tailer. Although many researchers study this kind of differentiation, most of them focus on the impact on manufacturer’s online direct channel strategy (e.g. Li et al., 2018 and Amrouche and Yan, 2012) and the retailer’s private brand introduction strategy (e.g. Raju et al. 1995; Jin et al. 2017 and Choi and Fredj., 2013). None of them considers the impact of brand differentiation on the retailer’s channel selection. The current paper also contributes to channel selection literature by investigating a novel channel structure where the e-tailer can choose to build her own offline channel or to align with her offline competitor. The study conducted by Bernstein et al. (2008) is the most relevant one to our study. They examine the offline retailer’s online entry strategy and consider a channel structure that is similar to the structure in our paper. They find that when offline retailers can share a high ratio of revenue with their online competitors and cooperate with them to build online channels, these offline retailers may fall into a prisoner’s dilemma. However, Bernstein et al. (2008) focus on an industry with undifferentiated offline retailers and undifferentiated brands. In contrast to their paper, this paper focuses on the interaction between an e-tailer and her offline competitor and take brand competition and channel operating costs into account. Moreover, although Bernstein et al. (2008) point out that lower channel operating cost can lead to a lower retail price, how can the channel operating cost effect retailer’s channel selection decisions is ignored in their study. As is pointed out by Wang et al. (2016), few previous studies investigate the role of channel operating cost in channel selection. They study both the role of channel operating cost and brand differentiation and find that low product differentiation cannot always benefit the retailer. The retailer should introduce only one channel with the lowest channel operating cost if the operating cost gap between the online channel and offline channel is high. Dual channel strategy can only benefit the retailer when the operating cost gap is low. However, they ignore that aligning with other retailers can also become one of retailers’ channel strategies. The current paper will fill the research gap by using a game-theoretic model to compare e-tailer’s self-built offline channel strategy and sharing channel strategy under the competition between the national brand and the private brand and investigate. The influence of brand differentiation and channel operating costs is also investigated in the current study

3 The Model

3.1 Description

We consider a supply chain with an e-tailer and a traditional retailer. The e-tailer and the traditional retailer operate an online channel and an offline channel respectively and both of them sell a national brand supplied by a national brand manufacturer. The decisions of the national brand manufacturer is not considered in the current study. Besides, the e-tailer has her own private brand and is selling it through her online channel. In order to enter the offline market, the e-tailer now faces two options. One is to sell her private brand through the traditional retailer and the other is to build her own offline stores. We assume that the e-tailer can sell both national brand and private brand through her offline stores. Figure 1 and Figure 2 illustrate the supply chain structure of two different scenarios in our model. All the players in this supply chain are self-interest and aim at maximizing their own profits.

Figure 1: Supply chain structure in the sharing channel scenario

Figure 2: Supply chain structure in the self-built channel scenario

3.2

Assumptions

The following assumptions are used in our model:

1. When operating dual channel, the e-tailer will keep the online price and offline price of a product the same. This assumption is used by some researchers (e.g. Wu et al., 2015 and Liu et al., 2006) and it in line with the practice of most e-tailers in reality when they enter the offline market.

2. Both traditional retailer and e-tailer will face channel operational cost. We assume that the unit online channel operating cost is lower than the unit operational cost of offline channels. This is because operating offline channel usually has a higher cost on store maintenance, higher inventory costs and higher cost on employing offline employees (Bernstein et al., 2008). Other cost parameters, such as production cost and selling cost, would influence the retailers’ decisions in a similar way, so we ignored them in the current paper. 3. The consumer behavior is determined by channel preference, brand preference, retailer preference and price. We assume that all these preferences are independent (Wang et al., 2016). We also assume that the channel preference dominates the other preferences, and the consumer will not switch to a less preferred channel. This assumption divides the market into two separate markets: the online market and the offline market. In practice, consumers may prefer offline channel because of the privacy concern of shopping online and may prefer online channel because of high transportation cost to reach offline stores. (Hsiao & Chen, 2017)

4. In the online market and offline market, consumers will choose from different brands and different retailers. We assume that the consumer demand of a certain brand in a certain retailer is determined by the base demand and the price competitions. (Jin et al., 2017; Raju et al., 1995 and Choi and Fredj, 2013). For the base demand, we first normalize the category base demand to 1, which means the demand of the category is 1 unit if all the products are free of charge. Then, proportions are used to describe consumer’s preference on different retailers and brands. This kind of preference parameter is described as the absolute difference in demand in the paper of McGuire and Sataelin (1983). For the price competition, we only focus on the differentiation between national brands provided by different retailers and the differentiation between the

national brand and private brand sold by the same retailer. In line with Choi (1996), Choi and Fredj (2013) and Jin et al. (2017), we ignore the differentiation between e-tailer’s private brand and the national brand sold by the traditional retailer.

3.3 Decision Variables and Parameters

In our model, we consider a product category with a base demand D. We use p1e and p1t to denote the retail price of the national brand sold by e-tailer and traditional retailer respectively. p2 denotes the price of private brand set by the e-tailer. Moreover, if the e-tailer chooses to use the offline channel provided by the traditional retailer, a slotting fee must be paid to the traditional retailer and a certain percentage of revenue must be shared with her. The traditional retailer will decide the revenue sharing rate r. We use F to denote the slotting fee paid to the traditional retailer. If the e-tailer decides to build the offline channel by herself, she will face the cost of building offline stores. C denotes the e-tailer’s fixed cost in building the offline channel. We assume that both F and C are exogenous.

Several parameters are used to describe the demand of the consumer. The price sensitivity of a certain brand and the cross-price sensitivity between national brands and private brands are denoted as α and β respectively. The customer also has a cross-price sensitivity between the national brands sold by e-tailer and traditional retailer, which is denoted as γ. We denote ρ as the proportion of consumers who prefer online channels, θ as the proportion of consumers who prefer national brands and δ as the proportion of consumers who prefer to shop in the channels provided by the e-tailer. The notation of decision variables and parameters are summarized in table 3

Table 3: Parameters and Decision variables

Symbols Descriptions Decision Variables:

p1e Retail price of the national brand sold by e-tailer

p1t Retail price of the national brand sold by traditional brand

p2 Retail price of the private brand

r Revenue sharing rate

Parameters:

D Base catagory demand

α Self-price sensitivity

β Cross-price sensitivity between national brand and private brand in the same store

γ Cross-price sensitivity between the same national brand sold by e-tailer and traditional retailer

ρ Initial proportion of consumers who prefer online channels

θ Initial proportion of consumers who prefer national brands

δ Initial proportion of consumers who prefer the e-tailer.

F Slotting fee of using the competitor’s offline channel

ce Online channel operating cost per unit of product

ct Offline channel operating cost per unit of product

C The fixed cost of building offline channel

3.4

Demand Functions

The sequence of consumer’s purchasing decision can be described as follows. First, the consumer will choose a preferred channel. The initial demand of online market is ρ and the initial demand of offline market is 1− ρ. After that, the consumer will choose from the national brand and the private brand in the preferred channel. In online market, the initial proportion of consumers who prefer the national brand is θ. Based on our assumption that the consumer preferences are independent, the base demand of national brand online is

ρθ, and the base demand of private brand online is ρ(1−θ). Similarly, in the offline market, the base demand

of national brand online is (1− ρ)θ and the base demand of private brand online is (1 − ρ)(1 − θ).

The demand will decrease if the price of the product increases. α describes consumer’s the sensitivity on price. The retailer will lose more consumers when increasing the price if α gets higher. Consumers will also switch to another brand depends on the price and the quality of different brands. If the cross-brand sensitivity β gets higher, the consumer is more likely to switch to another brand when the price of a brand gets higher. Our demand function is an extension of the linear demand model used by Jin et al. (2017). This linear demand function is developed by Raju et al. (1995) and also used in other studies (e.g. Choi and Fredj, 2013). The notations of the demands in our model are superscripted by C and S, which stands for the sharing channel scenario and the self-built channel scenario respectively. In the sharing channel scenario, the demand function can be addressed as follows:

The demand of the national brand online:

DC_E11= ρθD− αp1e− β (p1e− p2) (3.1)

The demand for the private brand online:

DC_E12= ρ(1− θ)D − αp2− β (p2− p1e) (3.2)

The demand for the private brand offline:

DCE22= (1− ρ) (1 − θ) D − αp2− β (p2− p1t) (3.3) The demand for the national brand offline:

D_TC= (1− ρ) θD − αp1t− β (p1t− p2) (3.4)

In the self-built channel scenario, the national brand provided by e-tailer and traditional retailer competes with the private brand provided by e-tailer. Previous literature (e.g. Raju et al., 1995; Choi and Fredj, 2013 and Jin et al., 2017) assume that the retailers are undifferentiated, which means that the base demand for the same brand in different retailers are the same. In our model, we loosen this assumption by using parameter

δ. If all the prices are zero, δ of the consumer will choose the e-tailer and 1− δ of them will choose the

traditional retailer. Under this scenario, there is also competition between national brands sold by different retailers. The higher parameter γ is, the fiercer price competition between national brands is. As is assumed in section 3.2, we only focus on the differentiation between national brands provided by different retailers and the differentiation between the national brand and private brand sold by the same retailer. The demand function in the self-built channel scenario is addressed as follows:

The demand for national brand online:

DS_E11= ρθD− αp1e− β (p1e− p2) (3.5)

The demand for the private brand online:

D_E12S = ρ (1− θ) D − αp2− β (p2− p1e) (3.6)

The demand for the private brand offline:

DS_E22= (1− ρ) (1 − θ) D − αp2− β (p2− p1e) (3.7) The demand for the national brand offline sold by the e-tailer:

DE21S = δ (1− ρ) θD − αp1e− β (p1e− p2)− γ (p1e− p1t) (3.8) The demand of the national brand offline sold by the traditional retailer:

DS_T = (1− δ) (1 − ρ) θD − αp1t− γ (p1t− p1e) (3.9)

3.5 Objective Functions and Constraints

In our game-theoretic model, we focus on the retailers’ profits. The notations of the profits in our model are also superscripted by C and S, which stands for the sharing channel scenario and the self-built channel scenario respectively. The profit functions of the retailers under each scenario in our model are addressed as follows:

1. Sharing channel scenario:

In the sharing channel scenario, we consider a two-stage game model. In the first stage, the traditional retailer will set her retail price for the national brand and decide a revenue sharing rate for the e-tailer to use her offline channel. In the second stage, the e-tailer will set the retail price for both the national brand and the private brand to maximize her profit.

The e-tialer’s profit:

ΠC_E(p1e, p2) = (p1e− ce)DE11C + (p2− ce)DCE12+ (1− r) (p2− cs)DCE22− F (3.10)

The traditional retailer’s profit:

Π_TC(p1t, r ) = (p1t− cs) DTC+ r(p2− cs)DE22C + F (3.11)

2. Self-built channel scenario:

In the self-built channel scenario, the traditional retailer will set the retail price of the national brand in her stores. After that, the e-tailer will first decide the retail price of national brand to maximize her profit. The profit function is addressed as follows:

The e-tialer’s profit:

ΠS_E(p1e, p2) = (p1e− ce)DE11S + (p1e− cs)DE21S + (p2− ce)DSE12+ (p2− cs)DE22S − C (3.12)

The traditional retailer’s profit:

ΠST(p1t) = (p1t− cs) DTS (3.13)

In our model, parameter α, β, γ, ρ, θ and δ range from 0 to 1. The decision variable r also ranges from 0 to 1. Both the two scenarios have constraints that the online operational cost is lower than offline operational cost, the retail price is higher than operational cost and all the prices and demands are non-negative:

0 < α, β, γ, ρ, θ, δ, r < 1

p1, p2> ct> ce> 0

DC

E11, DCE12, DCT, DE22C > 0

DS

E11, DSE12, DST, DE21S , DE22S > 0

4 Model Analysis

In this section, we analyze the concavity of our profit functions and obtain optimal decisions of the e-tailer and the traditional retailer in two different scenarios (i.e. self-built channel scenario and sharing channel scenario). We consider the traditional retailer as the Stackelberg leader.

4.1 Self-built Channel Scenario

Using backward induction, the e-tailer sets p1eand p2to maximize her profit in the second stage of the game. Analyzing the Hessian Matrix of ΠS

E, we obtain the following theorem:

Theorem 1. ΠS

E is jointly concave in p1e and p2.

The proof of Theorem 1 can be found in Appendix A. //Solving first order condition, we obtain the optimal p1eand p2:

p∗_1e=Dβ + α 2_c e+ α2ct− Dβθ + 2αβce+ 2αβct+ αctγ + βctγ + αγp1t+ βγp1t 2 (4αβ + αγ + βγ + 2α2₎ ++Dαδθ + Dβδθ + Dαρθ + Dβρθ− Dαδρθ − Dβδρθ 2 (4αβ + αγ + βγ + 2α2₎ (4.1) p∗₂=2Dα + 2Dβ + Dγ + 2α 2_c e+ 2α2ct− 2Dαθ − 2Dβθ − Dγθ + 4αβce+ 4αβct+ αceγ 4 (4αβ + αγ + βγ + 2α2₎ +αctγ + 2βctγ + 2βγp1t+ 2Dβδθ + 2Dβρθ− 2Dβδρθ 4 (4αβ + αγ + βγ + 2α2₎ (4.2)

Anticipating e-tailer’s decisions, the traditional retailer decides p1t in the first stage. Substitute p∗1einto πTS

and analyze the second order derivative, we obtain:

Theorem 2. ΠS

T is concave in p1t

The proof of theorem 2 can be found in Appendix A Solving the first order condition, the optimal p1t is obtained p∗_1t= ct ( α− γ ( αγ+βγ 2(4αβ+αγ+βγ+2α2₎− 1 )) + s0+ Dθ (δ− 1) (ρ − 1) 2α− 2γ ( αγ+βγ 2(4αβ+αγ+βγ+2α2₎− 1 ) (4.3) Where s0= γ(Dβ + α2_c e+ α2ct− Dβθ + 2αβce+ 2αβct+ αctγ + βctγ ) 2 (4αβ + αγ + βγ + 2α2₎ +γ (Dαδθ + Dβδθ + Dαρθ + Dβρθ− Dαδρθ − Dβδρθ) 2 (4αβ + αγ + βγ + 2α2₎

4.2

Sharing Channel Scenario

Using backward induction, the e-tailer make decisions on p1e and p2 in the second stage. Analyzing the Hessian Matrix of ΠC

E, we obtain:

Theorem 3. ΠC

E is jointly concave in p1e and p2

The proof of theorem 3 can be found in Appendix A Solving first order condition, we obtain the optimal

p∗₂=Dt3+ cet2− r ( ctt12+ Dt0t1 ) + ctt12+ βp1tt1− βp1trt1 2t2− 2rt12+ 2t12 (4.5) Where: t0= (1− θ)(1 − ρ) t1= α + β t2= α + 2β t3= (1− θ)(α + β) + βρθ

For the optimization of the traditional retailer’s profit in the sharing channel scenario, we first find the optimal p1t and then examine the optimal r. This optimization method is also used by Liu et al. (2007). Analyzing the second order derivative of ΠC

T(p1t), the following theorem is obtained:

Theorem 4. For any given r, the ΠC

T(p1t) is concave in p1t

Solving first order condition and collecting the coefficients of r, we obtain:

p∗_1t(r) =m1r

3_{+ m2r}2_{+ m3r + m} 4

m5r3+ m6r2+ m7r + m8

(4.6)

The expressions of m1 to m8 are complicated and we do not display them here in this paper.

Substitute p∗1t(r) into ΠCT we can change the problem into a single variable optimization problem. However,

the second order derivative of πC

T(r) is still very complicated. It is almost impossible to obtain the analytical

results, so we resort to numerical experiments in order to test the concavity of the profit function and to get managerial implications from our model.

5 Numerical Experiments

5.1 Experiment Design

Consider a case where α = 0.005, β = 0.005, γ = 0.005, ρ = 0.4, θ = 0.8, δ = 0.2, ce= 3, ct = 12, C = 0, F =

0, D = 1. We set the price sensitivity parameter α, β, and γ and the channel operating cost ce based on

the numerical example used in Cai et al. (2009). As is assumed in section 3.2, the offline channel operating cost is higher than the online operating cost. We set parameter ct to 12 which is 4 times higher than ce.

Parameter θ is set to 0.8 because of the characteristic of private brands. According to Jin et al. (2017), the sales of private brands usually account for about 23 percent of the total sales in the United States. For parameter ρ, we set it to 0.4. According to a report conducted by McKinsey, the online sales of some product categories (e.g. Consumer electronics) have already accounted for about 40% of the total category sales in China in 2016 (Wang, Lau & Gong, 2016). For parameter δ, we assume that the e-tailer is inexperienced in building offline stores and the traditional retailer’s offline store is more preferred by consumers. This parameter is set to 0.2 in our numerical example. Because parameters F and C are fixed costs and do not influence the retailer’s pricing decisions, we ignore them in our numerical experiments. The base demand

D is normalized to 1 based on our assumptions in section 3.2. In order to compare two aforementioned

scenarios, our numerical experiments ensure that both scenarios are feasible for the retailers. Particularly, when the revenue sharing rate equals 0 or 1 under the sharing channel scenario, the sharing channel is occupied completely by the e-tailer or the traditional retailer respectively, which is not possible in practice. We will also avoid this situation in our experiments. The purpose of the numerical experiment is threefold. Firstly, the concavity of the e-tailer’s profit function under sharing channel, which is not proven in section 4, is tested by the experiments. Secondly, we investigate how the changes of price sensitivity parameters, the market share parameters and the operating cost parameter will influence the retailer’s channel decisions. Thirdly, the possibility of supply chain coordination in the current supply chain structure is studied.

5.2

Concavity of the E-tailer’s Profit Function

We test the concavity of the e-tailer’s profit function by changing one parameter one time and set different values of r in the function of ΠC

T(r). The numerical results show that in most of our experiment cases the

second order derivatives are negative. This indicates that there exists an optimal revenue sharing rate for the traditional retailer to maximize her profit in our numerical experiments. Tables of the numerical results can be found in Appendix B.

5.3

The Impact of the Market Factors on the Retailer’s Channel Decisions

In most of our numerical experiments, the traditional retailer’s profit in the sharing channel scenario is higher than it is in the self-built channel scenario. Figures that show the results can be found in the Appendix C. It is worth pointing out that the limitation of our demand model may lead to this situation. In our model, we assume that there is no channel switching between online market and offline market. When building her offline stores, the e-tailer invades the traditional retailer’s offline market while the traditional retailer does not have any means to influence the online market. In the sharing channel scenario, however, the traditional retailer can benefit from the sales of the private brands, which makes the sharing channel a better choice for the traditional retailer. As a result, the whole channel structure in our model will depend on the channel decisions of the e-tailer and we only focus on the e-tailer’s channel decisions hereafter in this subsection.

5.3.1 The Impact of Price Sensitivity Parameters

As is shown in figure 3, parameter α has a negative impact on the e-tailer’s profits. When the self-price sensitivity of the product is high, the e-tailer will choose the sharing channel scenario as the optimal offline channel strategy. This is because, in the self-built channel scenario, the e-tailer sells the national brand in both the online and the offline market while in the sharing channel, the e-tailer only sells the national brand online. This means that the increase of self-price sensitivity will have a greater negative impact on the demand of the national brand sold by the e-tailer. In the sharing channel scenario, however, the e-tailer is

better off because it is easier for the e-tailer to increase the price to counter the negative effect of the increase of price sensitivity.

Figure 3: Impact of α on e-tailer’s profits

Parameter β has a similar effect on the e-tailer’s profit. However, we can see from Figure 4 that the e-tailer is more sensitive to the changes of β in the sharing channel scenario and the sharing channel is more preferred by the e-tailer when β is relatively low. To find the reason for this, we further investigate how β influences the profit of different products in different channels. Let πs

e11, πe12s , πse22, πe21s denote the profits of different

brands sold in different channels of the e-tailer in the self-built channel scenario and let πc

e11, πce12, πe22c denote

the profits of different brands sold in different channels of the e-tailer in the sharing channel scenario. As can be seen in Figure 5, β has a negative impact on the profit of national brand sold online under both scenarios. This negative impact is greater in the sharing channel scenario. To avoid too much loss on the profit of the national brand sold online, the e-tailer will choose the self-built channel scenario when the competition between the national brand and the private brand is intense. In reality, BarkBox, a pet product e-tailer who sells both national brands and her private brands online in the United States, has chosen to cooperate with her offline competitor Target to sell her private brands(DeBaun, 2017). It is reported that the pet product market has a low competition level (Salpini, 2018). This can be one of the reasons why BarkBox does not build offline stores when entering the offline market.

Figure 4: Impact of β on e-tailer’s profits

In the self-built channel scenario

In the sharing channel scenario

Figure 5: Impact of β on different products sold in e-tailer’s different channels in two different scenarios

that stronger competition between national brands is more preferred by the e-tailer when building her own offline stores. This is because when the national brands’ competition is fierce, more consumers will switch from the traditional retailer to the e-tailer because the price of the national brand sold by the e-tailer is lower. This has a positive impact on the demand of the national brand sold by the e-tailer. This finding can explain why e-tailers who are building their offline stores in practice, are trying to create a unique shopping environment, which is different from traditional offline shopping, to intensify the competitions between national brands and create strong store differentiation. Take Amazon 4-star offline stores, for example, the sections in the stores are carefully arranged to provide the customers with a shopping experience that is similar to online shopping. (Berthiaume, 2018)

Figure 6: Impact of γ on e-tailer’s profits

5.3.2 The Impact of Market Share Parameters

In our model, the e-tailer has a larger offline market share in the self-built channel scenario because both the national brand and private brand are sold offline. As a result, in the self-built channel scenario, the e-tailer is more sensitive to changes of the initial proportion of online consumers. This can be seen in Figure 7. We also find that the impact of parameter ρ is effected by parameter α. Figure 7 shows that when α equals 0.005, self-built channel scenario is always the optimal channel choice of the e-tailer. When α increases to 0.006, however, sharing channel scenario becomes the e-tailer’s optimal choice if ρ exceeds 0.392. We conclude that when ρ is relatively low, the self-built channel is preferred by the e-tailer. This is because when building her own offline stores, the e-tailer can sell both the national brand and the private brand offline, thus scrambling for more offline market share. When ρ is high, the offline market is less important to the e-tailer and she will avoid intense competition with the traditional retailer offline. In real life, e-tailers also prefer to have products with low online acceptance on sale in their offline stores. Chinese e-tailer JD.com, for example, is building her own offline stores that sell both her private brands and national brands in China. Most of her private brand products fall into the categories of home decor, household products and apparel.1 _These product categories is reported to have lower online preference.(Wang, Lau & Gong, 2016)

As is illustrated in Figure 8, parameter α has a great impact on the effect of θ. When alpha is increased to 0.006, the sharing channel scenario becomes the optimal channel strategy of the e-tailer when θ exceeds 0.77. When theta is relatively low, building own offline stores is more likely to be the optimal channel choice of the e-tailer. This is because when more consumers prefer the private brand, the private brand can generate more profit. Instead of sharing the profit of the private brand with the traditional retailer, the e-tailer prefers to sell her private brand exclusively. This finding can explain why Amazon starts to build her own offline stores. In the past, Amazon only aligns with her offline competitors to sell her private electronic brands. Recently,

1_{This conclusion is made based on the information of JD.com’s private brand, derived from https://www.joybuy.com/}

search?bucket=-2&keywords=\%4A\%2E\%5A\%41\%4F&arriveCountry=2456&pageSize=48&qpExclude=0&filterTypes=\%65\%78\ %70\%61\%6E\%64\%2C\%72\%65\%64\%69\%73\%73\%74\%6F\%72\%65&page=1

α = 0.005 α = 0.006

Figure 7: The impact of ρ on e-tailer’s profits

it is reported that Amazon’s private brands have high customer satisfaction and Amazon starts to sell her private brands through her offline stores. It is worth noting that some categories of Amazon’s private brands which are less preferred by the customers (e.g. clothing) are not available in Amazon’s brick-and-mortar stores (Berthiaume, 2018 and ScrapeHero, 2019). Figure 8 also shows that the profit of the e-tailer in the sharing channel goes up when θ increases. However, in the self-built channel scenario, the profit of the e-tailer decreases at first and then starts to increase. Figure 9 shows the impact of θ on different products sold in different channels under the self-built channel scenario. It can be seen that when θ is low, the negative effect on the private brand dominates the positive effect on the national brand. When θ increases, the positive effect on the national brand dominates the negative effect on the private brand. This can explain why the e-tailer’s profit decreases at first and then increases, as θ increases.

α = 0.005 α = 0.006

Figure 8: The impact of θ on e-tailer’s profits

Figure 9: Impact of θ on different brands sold in e-tailer’s different channels under the self-built channel scenario when α equals 0.005

Figure 10 shows that as the parameter δ increase, the profit of e-tailer in the self-built channel also increases. The explanation is straight forward. When more customers prefer to shop in the e-tailer’s offline stores, the offline channel operated by the e-tailer is more profitable.

Figure 10: The impact of δ on e-tailer’s profits

5.3.3 The Impact of Operating Cost Parameter

Figure 11 shows that when the offline operating cost is high, the sharing channel scenario is more likely to be the optimal offline entry strategy of the e-tailer. In the sharing channel scenario, e-tailer’s private brand compete directly with the traditional retailer’s national brand. The increase in offline operating cost will force the traditional retailer to increase the price of the national brand. As a result, more consumer will switch to the private brand in the offline channel, thus eliminating the negative impact of the increase of operating cost on the e-tailer’s profit. As a result, the e-tailer is better off in the sharing channel when the offline operating cost is relatively high. Figure 12 indicates that when offline channel operating cost is close to the online channel operating cost, the positive effect on the profit of private brand sold online dominates the negative effect on the profits of the private brand sold offline and the national brand sold online. As the offline operating cost increases, the negative effect dominates the postive effect. This can explain why the total profit of the e-tailer shown in Figure 11 increases at first and then decreases.

Figure 11: Impact of offline channel operating cost on e-tailer’s profits

Figure 12: Impact of offline channel operating cost on different products sold in different channels of the e-tailer under sharing channel scenario

5.4

Supply Chain Coordination

As is discussed in section 4.2. Theorem 4 implies that the traditional can choose a revenue sharing rate first and then make pricing decisions. This allows the traditional retailer to coordinate the supply chain by setting an appropriate revenue sharing rate. Before claiming the optimal p1t, the traditional retailer can anticipate the e-tailer’s decisions and compare the profit of the e-tailer under sharing channel and self-built channel. By solving equation ΠC

E(r) = Π S

E, the traditional retailer can obtain a revenue sharing rate that makes the e-tailer

indifferent to both channel scenarios, thus stimulating the e-tailer to use the sharing channel and creating a win-win condition. Table 4 shows an example of the supply chain coordination. The parameter setting in this example is the same with the setting in section 5.1. As can be seen in the table, with coordination, the e-tailer is indifferent to both scenarios. The sharing channel is still the optimal channel choice, although the profit decreases with coordination. The total profit of the supply chain also increases when the traditional retailer chooses to coordinate with the e-tailer by cutting the revenue sharing rate under the sharing channel scenario. We conclude that the e-tailer is more likely to choose the sharing channel scenario as optimal offline entry strategy and a win-win condition supply chain will be made if the traditional retailer is willing to make concession on the revenue sharing rate of building the sharing channel.

Table 4: An example of supply chain coordination

Self-built channel scenario Sharing channel scenario without coordination

Sharing channel scenario with coordination

Revenue sharing rate N/A 0.93 0.61

E-tailer’s profit 4.19 3.95 4.19

Traditional retailer’s profit 3.41 5.67 5.59

Total profit in the supply chain 7.60 9.62 9.78

6 Conclusion

This paper develops a game-theoretical model for a unique supply chain where the e-tailer can either build her own offline stores or use her competitor’s offline channels. Numerical experiments are conducted to get managerial implications. Our numerical results suggest that when entering the offline market, it is more profitable for e-tailers to use their competitors’ channel to sell private brands if the consumers’ online preference is high. The sharing channel scenario is also more suitable for product categories with high self-price sensitivities and low self-price competition between national brands and private brands. When building offline stores, it is better for e-tailers to make their stores more differentiated and preferred by consumers. These findings can guide the e-tailers who are interested in entering the offline market in practice.

The theoretical contribution of this paper is twofold. Firstly, we develop a linear demand model which describes the consumer demands in a supply chain with retailer competition and multiple brands. Although the model is still too complex to obtain analytical results, it is a valuable attempt and can be used as a basis for future research. Secondly, we study the e-tailer’s channel decisions during the offline entry. Moorthy et al. (2018) conclude that when selling a competitor’s brand, the competition between brands may benefit the retailer. In contrast, we find that the competition between the national brand and the private brand has a negative impact on the retailer’s profit. The competitor’s channel is less preferred by the e-tailer when the competition between the national brand and private brand is high. Wang et al. (2016) study the impact of channel operating cost on the retailer’s channel decisions. However, they only focus on the normal dual channel strategy. In the current paper, we find that for e-tailer’s who is trying to entering the offline market, using competitor’s channel to sell private brands offline can be an optimal dual channel choice when offline channel operating cost is high. We also investigate the coordination role of the revenue sharing rate, which is ignored by Bernstein et al. (2008). Our finding suggests that negotiation is important when building a sharing channel. When the traditional retailer agrees to make a concession in the revenue sharing rate, the e-tailer is more likely to choose the sharing channel scenario, which will be a win-win condition and benefit the whole supply chain.

We are aware that this paper has some limitations. One of them is that we ignore the competition between online channel and offline channel and assume that there is no channel switching between online market and offline market in our demand model. An more general model for the supply chain in this paper need to model the brand competition, channel competition and store competition at the same time which makes the selection of demand model one of the most difficult parts of this study. However, in order to get analytical results, the validity and complexity of the model need to be balanced. Groznik and Heese (2010) provide a two-dimensional spatial model where customers can choose from the physical stores of different retailers based on store preference and spatial difference. Applying this model to the supply chain in the current study can also make the online-offline switching possible. However, as is stated in their research, this model may generate tens of cases with different consumer demands, which makes it difficult to obtain analytical results. For future research, more advanced demand functions need to be developed for this kind of supply chain.

Our current work might be extended in several ways. What is the national brand manufacturer’s optimal decisions when e-tailer is entering the offline market would be a fruitful direction. Furthermore, the traditional retailer in our model does not own private brands. However, some traditional retailers in reality, such as Walmart, have already introduced their private brands. It would be interesting to investigate whether it is still profitable for the traditional retailer to align with the e-tailer if the traditional retailer also owns a private brand. When entering the offline market is becoming increasingly popular among the e-tailers in practice, more empirical research can also be conducted to get insights into the factors that the retailers will consider

when building sharing channels. For example, it is interesting to study whether selling a private brand of a competitor would damage the store image.

References

[1] Amrouche, N., & Yan, R. (2012). Implementing online store for national brand competing against private label. Journal of Business Research, 65(3), 325-332

[2] Bell, D. R., Gallino, S., & Moreno, A. (2018). Offline showrooms in omnichannel retail: Demand and operational benefits. Management Science, 64(4), 1629-1651.

[3] Berthiaume, D. (2018). Amazon expands curated brick-and-mortar format to Seattle. Retrieved 18 June 2020, from https://chainstoreage.com/real-estate/amazon-expands-curated-brick-and-mortar-format-to-seattle

[4] Bernstein, F., Song, J. S., & Zheng, X. (2008). “Bricks-and-mortar” vs.“clicks-and-mortar”: An equilib-rium analysis. European Journal of Operational Research, 187(3), 671-690.

[5] Cai, G. G., Zhang, Z. G., & Zhang, M. (2009). Game theoretical perspectives on dual-channel sup-ply chain competition with price discounts and pricing schemes. International Journal of Production Economics, 117(1), 80-96.

[6] Cai, G. G. (2010). Channel selection and coordination in dual-channel supply chains. Journal of Retailing, 86(1), 22-36.

[7] Chen, K. Y., Kaya, M., & Özer, Ö. (2008). Dual sales channel management with service competition. Manufacturing & Service Operations Management, 10(4), 654-675.

[8] Choi, S. C. (1996). Price competition in a duopoly common retailer channel. Journal of retailing, 72(2), 117-135.)

[9] Choi, S. C., & Coughlan, A. T. (2006). Private label positioning: Quality versus feature differentiation from the national brand. Journal of Retailing, 82(2), 79-93.

[10] Choi, S., & Fredj, K. (2013). Price competition and store competition: Store brands vs. national brand. European Journal of Operational Research, 225(1), 166-178.

[11] Chung, H., & Lee, E. (2017). Store brand quality and retailer’s product line design. Journal of Retailing, 93(4), 527-540.

[12] Clough, A. (2020). Amazon.com is planning physical stores in South Florida. One of them is in a mall. Retrieved from https://eu.tcpalm.com/story/news/newswire/2020/02/27/amazon-4-star-physical-stores-coming-3-south-florida-locations-palm-beach-miami/4891767002/

[13] Cattani, K., Gilland, W., Heese, H. S., & Swaminathan, J. (2006). Boiling frogs: Pricing strategies for a manufacturer adding a direct channel that competes with the traditional channel. Production and Operations Management, 15(1), 40.

[14] Dawes, J., & Nenycz-Thiel, M. (2014). Comparing retailer purchase patterns and brand metrics for in-store and online grocery purchasing. Journal of Marketing Management, 30(3-4), 364-382.

[15] DePhills, L (2018). AMAZON’S EXTRAORDINARY 25-YEAR EVOLUTION. Retrieved from https://edition.cnn.com/interactive/2018/10/business/amazon-history-timeline/index.html

[16] DeBaun, D. (2017). Target continues online retail partnership trend with BarkBox company agreement. Retrieved from https://www.bizjournals.com/twincities/news/2017/08/09/target-continues-online-retail-partnership-trend.html

[17] Dhar, S. K., & Hoch, S. J. (1997). Why store brand penetration varies by retailer. Marketing science, 16(3), 208-227.

[18] Fung Business Intelligence (2019). Venturing into private label business – the next big thing for e-commerce operators Eng: Fung Business Intelligence. Retrieved from https://www.fbicgroup.com/?q=publication/venturing-private-label-business-–-next-big-thing-e-commerce-operators-eng

[19] Groznik, A., & Heese, H. S. (2010). Supply chain interactions due to store-brand introductions: The impact of retail competition. European Journal of Operational Research, 203(3), 575-582.

[20] Huang, Z., & Feng, T. (2020). Money-back guarantee and pricing decision with retailer’s store brand. Journal of Retailing and Consumer Services, 52, 101897.

[21] Hsiao, L., & Chen, Y. J. (2014). Strategic motive for introducing internet channels in a supply chain. Production and Operations Management, 23(1), 36-47.

[22] Ingraham, N. (2016). Target starts selling Kindles again after a

four-year break. Retrieved from https://www.engadget.com/2016/08/05/

target-starts-selling-kindles-again-after-a-four-year-break/?guccounter=1& guce_referrer=aHR0cHM6Ly93d3cuZ29vZ2xlLmNvbS8&guce_referrer_sig=AQAAAGZPDu_ m1Y_WWH_cUJIHeLh3Q8LNx7fV21CsqNR0htuxL8NR5l1B9hNjfRVSgbi6YcNVQZfRYcqW4Ii_ o7EmIoZ6Iv27mIK5tbcMqDWSuur7mtkqCu7dOfwdYK2WmS6Gf2zM9yoPQmc8gsOpiNR46uV_7Z_ BZjeBht5oevmeOxky

[23] Jiang, Y., Kim, J., Choi, J., & Kang, M. Y. (2019). From clicks to bricks: The impact of product launches in offline stores for digital retailers. Journal of Business Research.

[24] Jin, Y., Wu, X., & Hu, Q. (2017). Interaction between channel strategy and store brand decisions. European Journal of Operational Research, 256(3), 911-923.

[25] Kurata, H., Yao, D. Q., & Liu, J. J. (2007). Pricing policies under direct vs. indirect channel competition and national vs. store brand competition. European Journal of Operational Research, 180(1), 262-281. [26] Kumar, K. (2016). Target resumes selling Amazon products after four-year hiatus.

Re-trieved from http://www.startribune.com/target-resumes-selling-amazon-products-after-four-year-hiatus/389184711/?refresh=true

[27] Li, W., Chen, J., & Chen, B. (2018). Supply chain coordination with customer returns and retailer’s store brand product. International Journal of Production Economics, 203, 69-82.

[28] Lieber, C. (2018, September 28). Amazon’s new store, Amazon 4-Star, is both a dream and a nightmare. Retrieved from https://www.vox.com/the-goods/2018/9/28/17911934/amazon-4-star-store-nyc

[29] Liu, L., Parlar, M., & Zhu, S. X. (2007). Pricing and lead time decisions in decentralized supply chains. Management science, 53(5), 713-725.

[30] Liu, Y., Gupta, S., & Zhang, Z. J. (2006). Note on self-restraint as an online entry-deterrence strategy. Management Science, 52(11), 1799-1809.

[31] McGuire, T. W., & Staelin, R. (1983). An industry equilibrium analysis of downstream vertical integra-tion. Marketing science, 2(2), 161-191.

[32] Mehra, A., Kumar, S., & Raju, J. S. (2018). Competitive strategies for brick-and-mortar stores to counter “showrooming”. Management Science, 64(7), 3076-3090.

[33] Moorthy, S., Chen, Y., & Tehrani, S. S. (2018). Selling your product through competitors’ outlets: Channel strategy when consumers comparison shop. Marketing Science, 37(1), 138-152.

[34] Mukherjee, W. (2019, September 5). Amazon coming offline with basics first and hits later. Retrieved from https://economictimes.indiatimes.com/industry/services/retail/amazon-coming-offline-with-basics-first-and-hits-later/articleshow/70986017.cms

[35] Nielsen. (2019). The Rise of Premium Private Label and its Impact on Discount Retailers. Re-trieved from https://www.nielsen.com/us/en/insights/article/2019/the-rise-of-premium-private-label-and-its-impact-on-discount-retailers/

[36] Raju, J. S., Sethuraman, R., & Dhar, S. K. (1995). The introduction and performance of store brands. Management science, 41(6), 957-978.

[37] Ru, J., Shi, R., & Zhang, J. (2015). Does a store brand always hurt the manufacturer of a competing national brand?. Production and operations management, 24(2), 272-286.

[38] Sayman, S., Hoch, S. J., & Raju, J. S. (2002). Positioning of store brands. Marketing science, 21(4), 378-397.

[39] ScrapeHero (2019).Amazon Steps Up Its Private Label Strategy, Retrieved 19 June 2020, from https://www.scrapehero.com/amazon-steps-up-its-private-label-strategy/

[40] Statt, N. (2018). Why Amazon’s future depends on moving from the internet to the physical world. Retrieved from https://www.theverge.com/2018/11/2/18049672/amazon-go-offline-retail-future-competition-walmart-food-drink-grocery-sales

[41] Salpini, C., (2018). The Bite In Barkbox’s Retail Gameplan .Retrieved from: https://www.retaildive.com/news/the-bite-in-barkboxs-retail-gameplan/517974/

[42] Wang, W., Li, G., & Cheng, T. C. E. (2016). Channel selection in a supply chain with a multi-channel retailer: The role of channel operating costs. International Journal of Production Economics, 173, 54-65. [43] Wang, W., Lau, A. & Gong, F. (2016). McKinsey iConsumer China 2016 survey

[44] Wu, H., Cai, G., Chen, J., & Sheu, C. (2015). Online manufacturer referral to heterogeneous retailers. Production and Operations Management, 24(11), 1768-1782.

[45] Zhang, P., He, Y., & Shi, C. V. (2017). Retailer’s channel structure choice: Online channel, offline channel, or dual channels?. International Journal of Production Economics, 191, 37-50.

Appendix

A

Proof of Theorems

Proof of theorem 1: The Hessian Matrix of ΠS

E is shown as follows: ∂ΠS E 2 ∂2_p 1e =−4α − 4β − 2γ < 0 (A.1) ∂ΠS E 2 ∂2_p 2 =−4α − 4β < 0 (A.2) ∂ΠS_E2 ∂2_p 1e ∂ΠS_E2 ∂2_p 2 − ( ∂ΠS_E2 ∂p1e∂p2 )2= (4α + 4β) (4α + 4β + 2γ)− 16β2> 0 (A.3) The matrix shows that ΠS

E is jointly concave in p1e and p2.

Proof of theorem 2: ∂ΠS T 2 ∂2_p 1t = 2 γ ( α γ + β γ 2 (4 α β + α γ + β γ + 2 α2₎− 1 ) − 2 α < 0 (A.4)

The sign of the second order derivative shows that ΠS

T is concave in p1t

Proof of theorem 3: The Hessian of ΠC

E is shown as follows: ∂ΠC E 2 ∂2_p 1e =−2α − 2β < 0 (A.5) ∂ΠC_E2 ∂2_p 2 = 2 (α + β) (r− 1) − 2β − 2α < 0 (A.6) ∂ΠC E 2 ∂2_p 1e ∂ΠC E 2 ∂2_p 2 − ( ∂ΠCE 2 ∂p1e∂p2 )2= (2α + 2β) (2α + 2β− 2 (α + β) (r − 1)) − 4β2> 0 (A.7) The Hessian Matrix shows that ΠC

E is jointly concave in p1eand p2.

Proof of theorem 4: ∂ΠC_T2 ∂2_p 1t = 2β (s1− 1) − 2α − s1(β (s1− 1) + αs1) < 0 (A.8) Where s1= βt1− βrt1 2t2− 2rt12+ 2t12 t0= (1− θ)(1 − ρ) t1= α + β t2= α + 2β t3= (1− θ)(α + β) + βρθ βt1− βrt1< 2t2− 2rt12+ 2t1 → s1− 1 = βt1− βrt1 2t2− 2rt12+ 2t12 − 1 < 0 → ∂πTC 2 ∂2_p 1t < 0

The second order derivative of ΠC

T(p1t) shows that Π C

B The Second Order Derivative of Π

(r)

Table B.1: The second order derivative of ΠC

T(r) with different values of α and r

r\α 0.005 0.0053 0.0056 0.0059 0.0062 0.0065 0.0068 0.0071 0.0074 0.0077 0.008 0 -0.09104 -0.12574 -0.15384 -0.17682 -0.1958 -0.21162 -0.22494 -0.23625 -0.24595 -0.25435 -0.2617 0.1 -0.13173 -0.17027 -0.2016 -0.22735 -0.24873 -0.26667 -0.28189 -0.29492 -0.30621 -0.31608 -0.32481 0.2 -0.18552 -0.22894 -0.2644 -0.2937 -0.31819 -0.3389 -0.35662 -0.37193 -0.38532 -0.39715 -0.40772 0.3 -0.25791 -0.30762 -0.34844 -0.38239 -0.41099 -0.43537 -0.45641 -0.47477 -0.49099 -0.50546 -0.5185 0.4 -0.35737 -0.41533 -0.46324 -0.50339 -0.53748 -0.5668 -0.59233 -0.61483 -0.63489 -0.65296 -0.66941 0.5 -0.49728 -0.56632 -0.62379 -0.67233 -0.7139 -0.74998 -0.78169 -0.80989 -0.83526 -0.85832 -0.87948 0.6 -0.69965 -0.78387 -0.8545 -0.91466 -0.96662 -1.01211 -1.05245 -1.08864 -1.12146 -1.15152 -1.17933 0.7 -1.0019 -1.10745 -1.19665 -1.27323 -1.33993 -1.39881 -1.45143 -1.49901 -1.54248 -1.58258 -1.61988 0.8 -1.47053 -1.60696 -1.72309 -1.82351 -1.91161 -1.98994 -2.06043 -2.12459 -2.18356 -2.23827 -2.28944 0.9 -2.22931 -2.41216 -2.56866 -2.70476 -2.82484 -2.93219 -3.02932 -3.11816 -3.20021 -3.27666 -3.34844 1 -3.52052 -3.77676 -3.99665 -4.18838 -4.35804 -4.51016 -4.6482 -4.77483 -4.89212 -5.0017 -5.10486

Table B.2: The second order derivative of ΠC

T(r) with different values of β and r

r\ β 0.001 0.0015 0.002 0.0025 0.003 0.0035 0.004 0.0045 0.005 0.0055 0.006 0 -0.2081 -0.20893 -0.20322 -0.19226 -0.17712 -0.15874 -0.13786 -0.11512 -0.09104 -0.06603 -0.04044 0.1 -0.25736 -0.2585 -0.25263 -0.24108 -0.22497 -0.20525 -0.18272 -0.15803 -0.13173 -0.10425 -0.07597 0.2 -0.32172 -0.32311 -0.31696 -0.30464 -0.28734 -0.26602 -0.2415 -0.21448 -0.18552 -0.15509 -0.12359 0.3 -0.40709 -0.40863 -0.40205 -0.38877 -0.37001 -0.34676 -0.31987 -0.29006 -0.25791 -0.22393 -0.18854 0.4 -0.52225 -0.5238 -0.51663 -0.50218 -0.48167 -0.45614 -0.42644 -0.39331 -0.35737 -0.31913 -0.27905 0.5 -0.68059 -0.68198 -0.67406 -0.65826 -0.63577 -0.60762 -0.57467 -0.53769 -0.49728 -0.45401 -0.40832 0.6 -0.90309 -0.90414 -0.89544 -0.87826 -0.85372 -0.8228 -0.78635 -0.74509 -0.69965 -0.65059 -0.59838 0.7 -1.2236 -1.22429 -1.21512 -1.19704 -1.17094 -1.13766 -1.09795 -1.0525 -1.0019 -0.94667 -0.88729 0.8 -1.69879 -1.6996 -1.69129 -1.67404 -1.64827 -1.6145 -1.57328 -1.52513 -1.47053 -1.40995 -1.34378 0.9 -2.42756 -2.4305 -2.42693 -2.41541 -2.39536 -2.36661 -2.32922 -2.28338 -2.22931 -2.16726 -2.09748 1 -3.59104 -3.60246 -3.61405 -3.621 -3.62059 -3.61116 -3.59172 -3.56164 -3.52052 -3.46809 -3.40419

Table B.3: The second order derivative of ΠC

T(r) with different values of ρ and r