Online coopetition between hotels and online travel agencies: From the perspective of cash back after stay

Tourism Management Perspectives 12 (2014) 104–112

Contents lists available at ScienceDirect

Tourism Management Perspectives journal homepage: www.elsevier.com/locate/tmp

Online coopetition between hotels and online travel agencies: From the perspective of cash back after stay Xiaolong Guo a, Xiabing Zheng a, Liuyi Ling a, Chenchen Yang b,⁎ a b

School of Management, University of Science and Technology of China, 96 Jinzhai Road, Hefei 230026, PR China School of Economics, Hefei University of Technology, 193 Tunxi Road, Hefei 230009, PR China

a r t i c l e

i n f o

Article history: Received 31 August 2014 Accepted 6 September 2014 Available online xxxx Keywords: Hotel pricing Online marketing Online travel agencies Cash back E-commerce

a b s t r a c t In the online marketplace, many hotels are concentrating on increasing their market share by establishing cooperation with online travel agencies (OTAs). Meanwhile, hotel websites and OTAs are marketing the hotel rooms at the same price as a result of the strong competition for the same pool of customers. Therefore, it is necessary to balance the cooperation and competition between hotels and OTAs. This study investigates the online coopetition (cooperation and competition) through an economical game analysis of an online supply chain consisting of a hotel and an OTA. It first provides an optimal solution to determine the unit commission fee of the hotel to maintain the cooperation. Afterwards, it studies the pricing process of the OTA to determine the cash back value for the customers with respect to the OTA's maximal profit. Moreover, the deeper analysis of the cooperative model demonstrates that a quantity discount contract based on the revenue sharing could eliminate the competition and coordinate the participants in the online supply chain. © 2014 Elsevier Ltd. All rights reserved.

1. Introduction With the rapid development of information technology, the traditional hospitality and tourism industry has entered the E-commerce age. According to a blog post from Thomson (2012), 57% of the rooms from the top 30 brand hotels are booked from the online distribution channels, including the host websites of the hotels and their third party websites such as OTAs in late 2010. At the same time, tourists are becoming increasingly dependent on online travel agencies (OTAs) for their journey plan (Guo, Ling, Dong, & Liang, 2013). The two most popular OTAs in mainland China, eLong, Inc. (http://www.elong. com) and Ctrip.com (http://www.ctrip.com), provide online reservation services for the travelers and market distributions for various cooperating hotels and airlines, globally. Regarding the experience of purchasing hotel rooms from these two OTAs, an interesting issue is raised: in order to attract new customers and maintain loyal customers, the OTAs provide cash back to the travelers who will make reservations from the given website. A cash back example from eLong, Inc. is shown in Fig. 1, indicating that the cash back value reaches 10% of the average daily room rate (CNY100/ (HK$1262 * 0.795CNY/HK$)). The cash back strategy is similar to loyalty programs such as mileage service provided by the airlines to prevent customers from switching to other airlines (Kim & Li, 2009). However, the significant difference ⁎ Corresponding author. Tel.: +86 55163603461. E-mail addresses: [email protected] (X. Guo), [email protected] (X. Zheng), [email protected] (L. Ling), [email protected] (C. Yang).

http://dx.doi.org/10.1016/j.tmp.2014.09.005 2211-9736/© 2014 Elsevier Ltd. All rights reserved.

between cash back strategy and mileage service is that the cash back can be returned to the bank account immediately when the customer pays the bill at the hotel implying that the trade is successfully completed. In other words, the cash back only works for this “one-time deal” and this strategy cannot stop the customers from switching to other OTAs in future transactions. Since this is a newly emerged issue in the online marketing channels and has been paid little attention by academic researchers, the purpose of this paper is to analyze the coopetition between the hotels and OTAs regarding the cash back effects. Through a leader–follower game model consisting of a single hotel and an OTA, this paper considers the following questions: (1) What is the difference between the cash back policy and direct price discount strategy? (2) Which channel do the travelers choose between the hotel-host website and OTA agent website when they make reservations, and how do they make these choices? (3) How does the OTA determine the optimal cash back value for its customers? (4) Which contract can eliminate the competition and coordinate the participants in the online supply chain? A utility function for travelers making reservations from different distribution channels (the hotel-host website and the OTA website) is introduced to explain the difference between cash back and direct price discount. According to the purchasing utilities, the travelers make their decisions of the booking channels. Given the travelers' channel choice and the market size of the two distribution channels, the hotel and the OTA determine their optimal market decisions with respect to their own maximal profits. From the analysis of the decision process for the leader–follower game, we find that the classic principal–agent model cannot achieve full channel coordination.

X. Guo et al. / Tourism Management Perspectives 12 (2014) 104–112

105

Fig. 1. An example of cash back from eLong, Inc.

Fortunately, we are able to introduce a new contract based on the expanded market size of the hotel under the cooperation with the OTA. The remainder of the study is organized as follows. After introducing the related literature as research background, we describe the problem about the online supply chain and marketing distribution channels in Section 3. Afterwards, we provide the optimal decisions of the participants in both integrated and decentralized scenario in Section 4. Next, in Section 5, we show a numerical example and introduce a new contract based on realized sales of hotel rooms and revenue sharing to coordinate the online supply chain in the decentralized scenario. Finally, we conclude this study by summarizing the findings and discussing the directions for future research in the last section.

is the best pricing strategy for hotels in the case of high costs and rigid changes in demand. Guo et al. (2013), which is the most relevant research to ours, study the cooperation contract among the third-party websites and hotels. Employing a theoretical game analysis, they have provided the equilibrium decisions of the hotels and third-party websites. However, the competition between the two players and the cash back strategy are not taken into account in their study, and there is little understanding of this issue in existing literature. Therefore, to fill this research gap, the present paper aims at investigating the coopetition relationship (Casadesus-Masanell & Yoffie, 2007) between hotels and OTAs with cash back effect.

2. Research background

3. Online supply chain and marketing channels description

In recent years, studies on the online marketing issue about tourism products have increasingly acknowledged the existing literature (Zhang, Song, & Huang, 2009). Shon, Chen, and Chang (2003) study the cooperation between airlines and OTAs. Yoon, Yoon, and Yang (2006) verify the impact of e-commerce on the distribution of flight tickets form adopting Korea as an example. Göymen (2000) points out that cooperating with third parties is beneficial for hotels' development. Likewise, Schulz (1994) demonstrates that hotels and travel agencies or other third-party companies are coming to realize the advantages of collaboration over competition. Ling, Guo, and Liang (2009, 2011) investigate the optimal marketing strategy of hotels when cooperating with travel agencies (for both offline and online) from the pricing perspective. Guo and He (2012) present the optimal pricing strategy of hotels for different tourism packages under cooperation with tour operators. Pricing strategy, an important marketing element and an effective management leverage, is always an important research topic. Followed by the prior pricing model for hotel rooms to optimize the profitability of hotels provided by Gu (1997), many researchers continue to work on this topic in the hospitality industry. Lai and Ng (2005) propose an optimization model for hotel pricing in the circumstances of demand uncertainty. Likewise, Schwartz (2006) proposes a booking project to increase hotel revenue through analyzing the relationship between room booking and hotel revenue, and Ling, Guo, and He (2012) propose an optimal hotel pricing model for long-term stay. Further, demand is influenced by the product price heavily and is actually a key factor affects the decision maker's choices. Pan (2007) explains how market demand and hotel capacity affect the pricing strategy of the hotels. Moreover, van der Rest and Harris (2008) demonstrate that discount

This section gives the description of the online supply chain consisting of a hotel and an OTA, and the marketing channels hosted by the two participants. Since the purpose of this paper is to analyze the online coopetition between the two players (a hotel and an OTA), we only focus on the travelers who make reservations through these two online distribution channels, while the other ones who book rooms by phone or in person are ignored in this paper. Thus, in this situation, we suppose the hotel only has two types of customer source: one books hotel rooms through the hotel website and the other from the OTA. In accordance with the prior studies (e.g., Guo et al., 2013; Ling et al., 2011), we use t-travelers to denote the travelers who book the hotel rooms through the hotel website and w-travelers for the customers from the OTA. 3.1. Cooperation description Considering the significant market share of online sales of hotel rooms, a hotel with capacity of C rooms would like to cooperate with an OTA to expand its online market to increase its occupancy rate for generating more profit. Suppose that all the C rooms of the hotel are identical and each room only accommodates one customer without any influence in the findings of the model (Guo et al., 2013; Ling et al., 2011). Suppose the daily fixed operational cost of the hotel is Fc, and the daily variable cost for each room is zero, which has no effect on our findings. In the cooperative relationship of the hotel and the OTA, the OTA distributes the hotel rooms at the same price, p0, to their customers as the hotel website, and receives a commission fee, ω, for each sold room from the hotel (Guo et al., 2013; Ling et al., 2011; Toh, DeKay, &

106

X. Guo et al. / Tourism Management Perspectives 12 (2014) 104–112

Raven, 2011). That is to say, the customers need to pay the same money for the two booking channels (the hotel website and the OTA). While compared with the information of the hotel rooms published by the OTA, the customers can get more detailed descriptions from the hotel's host website, and the customers, especially the Chinese consumers, “place more emphasis on the visual appearance and prefer pictures and photos” (Ding, 2011). Additionally, booking from the hotel's host channels gives more guarantee for room availability since most hotels adopt an overbooking strategy to hedge no-shows and late cancelations. So many travelers only get information from the OTA and then switch to the hotels' host websites to make reservations (Toh, DeKay et al., 2011; Toh, Raven & DeKay, 2011; Wu, Law, & Jiang, 2013). As a result, in order to attract customers to make reservations from the OTA, the OTA gives cash back to the w-travelers, which will only be transferred to the online account of the customers after their stay at the target hotel. We denote the value of the cash back as φ in this paper. Furthermore, the operational cost of the OTA is ignored without any influence of the outcomes of the model. 3.2. Customer utilities Considering the service difference between the hotel website and the OTA, we denote s1 as the purchasing experience of the t-travelers making reservations from the hotel's host website, and s2 as the purchasing experience of the w-travelers booking hotel rooms from the OTA. Since most hotels adopt an overbooking strategy to hedge no-shows and late cancelations, making reservation from the hotel website gives more guarantee for room availability. Furthermore, according to the findings from Ding (2011), the (Chinese) travelers “tend to place more emphasis on visual pictures and photos over text description”. Hence, it is reasonable to assume s1 N s2 and Δs = s1 − s2 is the purchasing experience difference between the two booking channels of the travelers. In order to make up the purchasing experience disadvantage and attract the travelers to make hotel reservations from the OTA, the OTA provides cash back, φ, for their customers. Because the cash back is transferred to the customers' online account of the OTA only when the w-travelers finished staying at the hotel, the cash back is only an

expected value of the w-traveler and different customers have various preferences for this expectation. We denote the reference as θ, and for each customer, θ is a random value following Uniform distribution normalized to [0, 1] (García & Tugores, 2006; Guo & He, 2012). Suppose v is the basic utility of the accommodation of the hotel and homogeneous for both distribution channels. According to the description given so far, the utility of the travelers booking from both channels can be obtained as follows, U 1 ¼ v þ s1 ;

ð1Þ

U 2 ¼ v þ s2 þ θφ;

ð2Þ

where U1 is the purchasing utility of the travelers who make reservations through the hotel website and U2 is the utility of the travelers who purchase from the OTA. 3.3. Number of customers and profits Through the cooperation with the OTA, the information of the hotel is distributed in two channels: the hotel website and the OTA. As a result, the hotel receives the customers from the two sources. As shown in Fig. 2, the hotel faces a demand of x = μ + ε, where μ is the expected number of customers; and ε is a stochastic variable, with zero mean and finite variance, with respect to the fluctuant extent of the number of customers. Among the customers, there are μ1 of them know the hotel from the hotel's host distribution channel and the others (μ2 + ξ) know it from the OTA (Bastakis, Buhalis, & Butler, 2004; Ling et al., 2011). Accordingly, ε1 and ε2 are the stochastic components of the two demands. The customer source of the OTA can be expanded by the OTA's sales effort like advertising, conditional rebate, positive evaluation (Chevalier & Mayzlin, 2006), service bundling (Kim, Bojanic, & Warnick, 2009) and so on in addition of the cash back strategy mentioned above. That is to say, the demand of the hotel can be increased by the OTA's effort. Suppose that the basic expected number of the customers from the OTA without OTA effort is μ2 and it can be increased by an amount ξ through a sales effort provided by the OTA. The effort causes the OTA

Fig. 2. Demand description and customer choices.

X. Guo et al. / Tourism Management Perspectives 12 (2014) 104–112

an investment g(ξ), which is convex increasing in ξ and g(0) = 0. We assume a quadratic function g(ξ) = kξ2/2 with k N 0 to express the investment, which has been widely used in the previous literature (e.g., Guo et al., 2013; Huang & Li, 2001; Little, 1979). Here k can be considered as the investment factor for the sales effort: the bigger k is, the higher the investment required to attract a customer. As a result, the hotel receives a demand with main value μ = μ1 + μ2 + ξ = μ0 + ξ (here we denote μ0 ≡ μ1 + μ2), with an investment, g(ξ) = kξ2/2, for the sales effort of the OTA. We denote that the total demand of the hotel without the OTA's sales effort follows a stochastic distribution with probability density function (pdf), f(⋅), and cumulative distribution function (cdf) F(⋅). The mean value of the total demand is μ0 ≡ μ1 + μ2 and the variance is ε. Suppose the OTA's sales effort only influences the mean value of the demand. Then the pdf of the demand with OTA's effort is fe(x) = f(x − ξ) and the cdf is Fe(x) = F(x − ξ), where ξ is the increased number of customer as a result of the OTA's sales effort. Therefore, under the cooperation with the OTA, the hotel has an expected number of customers shown as follows, Z X¼

C 0

Z xf e ðxÞdx þ

∞ C

Z C f e ðxÞdx ¼ C−

C

F ðx−ξÞdx;

ð3Þ

0

where the first term is the number of accommodated customers when the number of customers is smaller than the hotel's capacity, while the second term is the one when the number of customers is larger than C. According to the purchasing utilities of the customers from the two booking channels, a traveler will make his/her reservation from the OTA only when he/she gets higher utility from the OTA, that is, U2 N U1. Otherwise, he/she will book the hotel room from the hotel website. From Eqs. (1) and (2), we know that a traveler will chose the OTA when θ N Δs/φ. That is to say, among the X customers of the hotel, x1 = Δs/φ ⋅ X of them are t-travelers and the rest x2 = (1 − Δs/ φ) ⋅ X are w-travelers. As the numbers of t-travelers and w-travelers and the costs are observed, the profits of the hotel and the OTA are realized as follows:  πh ¼ p0 X−ωx2 −F c ¼

   Δs  C p0 −ω 1− C−∫0 F ðx−ξÞdx − F c ; φ

  Z Δs C− πo ¼ ðω−φÞx2 −g ðξÞ ¼ ðω−φÞ 1− φ

107

distribution channel; (2) the OTA takes the contract and determines the optimal effort level to increase the total demand of the hotel and the optimal cash back for attracting the travelers to make reservations through its booking system; (3) both the hotel and the OTA accept the reservations at the same daily room rate, p0; (4) the hotel provides accommodation service for both t-travelers and w-travelers at the target day; and finally (5) the OTA receives the commission fee from the hotel according to the contract terms and the realized number of wtravelers. According to the event sequence presented above, we can know that the decision variable of the hotel is the unit commission fee for the OTA and the decision variables of the OTA is the effort level and the cash back. In the following section, we will further analyze how to achieve the optimal decisions and show the equilibrium. 4. Decision analysis In practice, the hotel cannot control the details of the cash back value and cannot decide how the OTA sets its sales effort level, either. Nevertheless, the hotel could provide an appropriate incentive through the commission fee to induce the OTA to make desired decisions of the hotel. In order to provide a benchmark for the analysis of the realistic cooperative problem, we first present the integrated model in the following subsection, by assuming that the hotel and the OTA are integrated as a single player, and then analyze the private action case in the subsequent subsections. 4.1. Benchmark model: the integrated decision This subsection presents the problem under the assumption that the hotel and the OTA act as a single player. We set this as the benchmark case of the problem and the situation often offers the first-best solution because all the decisions are achieved according to the maximal profit for the whole chain. The problem is max fφ;ξg

 Z Π ¼ πh þ πo ¼ ðp0 −ðφ−ΔsÞÞ C−

 2 kξ −F c ; F ðx−ξÞdx − 2 0 C

ð4Þ

 kξ2 ; ð5Þ F ðx−ξÞdx − 2 0 C

where the subscript h denotes the hotel and o represents the OTA. 3.4. Event sequence and decision variables In the above Stackelberg game with the hotel as a leader and the OTA as a follower, the event sequence is shown in Fig. 3. The detailed sequence of events is as follows: (1) the hotel offers the OTA a contract consisting of the unit commission fee of each sold room from the OTA's

s:t:

0 ≤Δs=φ ≤1 and φ; ξ ≥0:

Obviously, the objective total profit of the integrated player is a decreasing function of the value of cash back. As a result, it is the best choice to set φ as small as possible, so the optimal decision on cash back is setting φ equals Δs. The following proposition gives the first-best solution for the integrated decision maker, Proposition 1. In the centralized scenario, the optimal decisions which leads to the maximal profit for the whole supply chain is: φ

ξ

FB

FB

¼ Δs;

ð6Þ

¼ arg f F ðC−ξÞ ¼ kξ=p0 þ F ð−ξÞg:

ð7Þ

ξ

The solution {φFB, ξFB} is unique, and the expected number of w-travelers is zero. Consequently, the total expected profit for the supply chain is, ΠFB = p0(C − ∫ C0F(x − ξFB)dx) − k(ξFB)2/2 − Fc. Corollary 1. (i) ∂ξFB/∂C N 0, (ii) ∂ξFB/∂p0 N 0, (iii) ∂ξFB/∂k b 0.

Fig. 3. Event sequence of the cooperative model.

Corollary 1 tells that the hotel capacity and room rate have positive influence on the optimal effort level and the expected demand decreases the optimal level. However, it is decreasing with the investment factor.

108

X. Guo et al. / Tourism Management Perspectives 12 (2014) 104–112

4.2. Private actions: the OTA's decision

Table 1 Parameters for the numerical example.

After the analysis of the controllable situation, now we turn to the realistic scenario as expected in the cooperation relationship. Given the unit commission fee, ω, the OTA decides {φ, ξ} to maximize its expected profit. Hence, the OTA solves   Z Δs C− max πo ¼ ðω−φÞ 1− φ fφ;ξg

 kξ2 : F ðx−ξÞdx − 2 0 C

The OTA's problem is joint concave in φ and ξ, and the following proposition gives the OTA's optimal response to the unit commission fee given by the hotel. Proposition 2. There is a unique optimal response for the OTA in which it chooses φ* and ξ* as 

φ ¼

pffiffiffiffiffiffiffiffiffiffi Δsω; 



ξ ¼ arg

F ðC−ξÞ ¼ kξ=

ξ

ð8Þ  pffiffiffiffi pffiffiffiffiffiffi2 ω− Δs þ F ð−ξÞ :

ð9Þ

Corollary 2. ∂φ*/∂ω N 0, ∂ξ*/∂ω N 0. Corollary 2 explains the optimal responses of the OTA to the hotel's contract term, the unit commission fee, ω. With respect to the unit commission fee, ω, we find that if the hotel offers a high ω, the OTA will increase the optimal cash back, φ, for the w-travelers to encourage the customers to make room reservations from its website; at the same time, the OTA will also increase the effort level, ξ, to expand the demand of the hotel. 4.3. Private actions: the hotel's decision Knowing that the OTA will respond its contract terms by choosing φ and ξ according to Eqs. (8) and (9), the hotel determines ω according to its maximal profit by solving max ω

s:t:

Z  pffiffiffiffiffiffiffiffiffiffi πh ¼ p0 −ω þ ωΔs C−

C 0

 F ðx−ξÞdx −F c ;

ð10Þ

pffiffiffiffi pffiffiffiffiffiffi2 F ðC−ξÞ ¼ kξ= ω− Δs þ F ð−ξÞ; and ω ≥0:

We denote the optimal solution under this scenario with superscripts SB (second-best). While with respect to the unit commission fee, ω, unfortunately, Eq. (10) is not generally quasiconcave, and consequently, is not unimodal. In order to get additional insights from circumventing this difficulty, we focus on the numerical analysis in the next section.

Parameter

C

p0

Fc

μ0

a

Δs

k

Basic value

200

100

10,000

150

60

10

0.1

Note: C: room capacity of the hotel; p0: average room rate of hotel rooms for the both distribution channels; F: the daily fixed operational cost of the hotel; μ0: the mean value of the number of hotel customers from both distribution channels; a: the demand fluctuation of the hotel; Δs: the consumer purchasing experience difference between the two booking channels; k: the cost coefficient of the sales effort of the OTA.

serves to illustrate the results found in the previous sections and to show the related findings for the online marketing of hotel rooms. Furthermore, a Uniform distribution on [μ0 − a, μ + a] is employed to describe the hotel demand without OTA's sales effort. Then we have f(x) = 1/(2a) and F(x) = (x − μ0 + a)/(2a). The basic parameters for this section are shown in Table 1. Based on the above parameters, the optimal solutions for the hotel and the OTA in both centralized and decentralized scenarios can be calculated by the simulation implemented in Wolfram Mathematica® 8.0.1.0. Table 2 summarizes the fist-best and second-best solutions for the players. From the results of Table 2, we know that the integrated player in the centralized scenario don't want to offer a cash back more than the purchasing experience, as a result, all the customers will make room reservations through the hotel's host website. There are 199.42 customers expected to stay in the hotel with an OTA effort level at 98.21, which leads almost full utilization of the hotel rooms (99.71%) and maximal expected profit at 9459.8. While under the second-best situation, the hotel and the OTA make their private actions with respect to their own profit. The OTA would like to offer a considerable cash back for the w-travelers and its sales effort level is smaller than the centralized scenario since the unit commission fee is always smaller than the average room rate. As a result, the hotel receives 104.58 expected t-travelers and 67.74 expected w-travelers, and its expected room utilization goes down. Finally, the total profit of the supply chain is smaller than the centralized scenario as a result of double marginalization. In addition, in order to provide a detailed illustration of the impact of the basic parameters, Table 3 summarizes the sensitivity of the second-best results with respect to the room capacity, average room rate, expected demand value and purchasing experience difference in the decentralized scenario. The results of Table 3 tell that the optimal unit commission fee provided by the hotel to the OTA increases with its vacancy rate (high room capacity, C, and/or small expected customer number, μ0), average room rate and the expected purchasing experience difference of the customers. Correspondingly, according to the increasing unit commission fee, the cash back value provided to the w-travelers and the sales effort level of the OTA are both increasing. These findings are in line with Corollaries 1 and 2 in precious sections. Furthermore, the results could also be verified by the practical examples (http://hotels.elong. com, http://hotels.ctrip.com), three selected examples from eLong, Inc. are shown in Fig. 4.

5. Special cases This section first gives the numerical analysis about the problem in Section 5.1. Considering the performance gap between the first-best solution and the second-best one, Section 5.2 introduces a coordination contract to impel the decentralized equilibrium achieving full channel coordination and to provide the supply chain maximal profit as the benchmark. 5.1. Experiments In this subsection, we consider a numerical study involving a hotel with C = 200 identical rooms and an OTA. This numerical example

Table 2 Optimal solutions and profits in different situations. φ

ξ

πh

πo

Π

Type of situations

ω

First-best Second-best

– 10 98.21 199.42 0 – – 9459.8 27.15 16.48 28.49 104.58 67.74 5392.97 682.33 6075.3

x1

x2

Note: ω: the unit commission fee provided to the OTA from the hotel; φ: the cash back value offered to the w-travelers from the OTA; ξ: the effort level of the OTA to expand the demand; πh: the profit of the hotel; πo: the profit of the OTA; Π: total profit of the whole supply chain including the hotel and the OTA.

X. Guo et al. / Tourism Management Perspectives 12 (2014) 104–112

109

Table 3 Sensitivity analysis of the results in decentralized scenario. Notation

C

p0

μ0

Δs

ω

φ

ξ

x1

x2

πh

πo

Basic example

200

100

150

10

27.15

16.48

28.49

104.58

67.74

5392.97

682.33

6075.3

Varied parameter and values

180 220

5 20

21.14 31.03 25.38 28.70 30.01 23.62 20.16 38.48

14.54 17.61 15.93 16.94 17.32 15.37 10.04 27.74

13.19 42.35 24.93 31.51 37.06 19.37 32.71 21.97

106.90 106.72 106.62 102.91 98.91 112.50 87.21 120.92

48.52 81.27 63.23 71.42 72.42 60.41 87.90 46.79

4515.98 6277.53 3681.67 7126.54 4960.41 5864.31 5739.24 4970.88

311.50 1000.41 566.30 790.15 849.91 479.87 835.83 478.21

4827.48 7277.93 4247.97 7916.69 5810.32 6344.17 6575.08 5449.09

90 110 140 160

5.2. Online supply chain coordination According to the numerical results presented in Table 2, we know that the performance of the decentralized solution is worse than the integrated one. Compared with the integrated profit, there is a more than 35% profit loss in the decentralized scenario in the given example, which is significant for the supply chain participants. Consequently, it is necessary to design a coordination contract to make the online supply chain perform at a higher level to generate more profits.

Π

However, considering that the cash back is an outflow from the supply chain to the customer, the best choice for the supply chain is to set the cash back as small as possible and drive all the customers make reservations through the hotel website. As a result, the present cooperative contract, the hotel pays the OTA according to the number of w-travelers, cannot provide a full channel coordination, as a consequence of the fact that the OTA will gain negative profit when φ ≤ Δs, which is a necessary condition when full coordination achieved. Hence, as a leader in the cooperation, the hotel needs to introduces a new contract to coordinate the online channel.

Fig. 4. Cash back of eLong, Inc for different hotels in Beijing, Hong Kong and Macau.

110

X. Guo et al. / Tourism Management Perspectives 12 (2014) 104–112

Consequently, considering the hotel demand can be expanded with the cooperation with OTA (Anderson, 2011), and the expected occupancy rate and profit will also be increased, it is reasonable for the hotel to pay the OTA based on the expectation of the realized room demand instead of the number of w-travelers. Proposition 3. The online supply chain can be fully coordinated if the hotel provides a unit commission fee for the OTA based on the realized number of customers including both t-travelers and w-travelers as follows, Z βk ϖ ¼ ð1−βÞp0 þ

C

arg ξ

0

!2 F ðx−ξÞdx ¼ C−q −2ð1−βÞF c 2q

: ð11Þ

where q is the total number of customers for the hotel, and β is a constant belongs to the interval [0, 1]. The expected profits of the hotel and the OTA, under this coordination contract, are πh = βΠFB and πo = (1 − β)ΠFB. Proof. On one hand, the profit of the OTA according to the new contract is πo = ϖX − kξ2/2; on the other hand, the total expected number of customer for the hotel is q = X and is given in Eq. (3). By substituting Eqs. (11) and (3) into the OTA's objective function, we have, πo = (1 − β)(p 0 X − kξ 2 /2 − F c ) = (1 − β)Π. Through first-order condition, the OTA maximize its profit by selecting the  sales effort level at ξ ¼ arg f F ðC−ξÞ ¼ kξ=p0 þ F ð−ξÞg. Then, we ξ can further get the profits of the hotel and the OTA as, πh = βΠFB FB and πo = (1 − β)Π . ■ This contract is deduced from quantity discount based on revenue sharing, which is a popular pricing strategy provided by the suppliers to influence buyer's purchasing behavior through providing economic incentives (Shin & Benton, 2007), and is also widely used among the service suppliers (e.g., Campo & Yagüe, 2007; Guo & He, 2012). Under the coordination contract, the constant, β, is the key parameter which influences the share of the profits between the hotel and the OTA. In order to inspire the players to participate in this contract, the hotel should give the OTA as well as itself more profits than that without SB 1 coordination, i.e., (1 − β)Π N πSB o and βΠ N πh . 6. Conclusions and future research This paper studies the coopetition relationship between hotels and OTAs about the online marketing channels of hotel rooms by investigating an online supply chain consists of a hotel and an OTA. We provide the optimal decision model for the hotel to determine the commission fee, and for the OTA to determine the cash back amount. Additionally, as Jaramillo and Mulki (2008) indicated, sales effort is always positively encouraged in the business activities, and there is no exception for the service industry. Appropriate sales effort of the OTA could expand the demand of the hotel, increase the occupancy rate, and ultimately increase the profit of the hotel and OTA itself. Given the numerical examples, we find that the second-best equilibrium in the decentralized scenario incurs a 35% profit loss for the supply chain compared with the integrated one, which is a result of double marginalization effect. Furthermore, the principal– agent contract consists of a unit commission fee for each sold room through the OTA cannot coordinate the participants of the online supply chain. Fortunately, a quantity discount contract based on revenue sharing can coordinate the OTA, in which the hotel pays the OTA commission fees based on the total number of realized customers including both the t-travelers and w-travelers instead of only the w-travelers.

1

Particularly, the value of β can be determined by standard Nash bargaining model (Nash, 1950) or Shapley Value (Shapley, 1953).

This paper may be the first work focusing on the cash back issue in hotel online marketing channels through cooperating with OTAs. Further research may continue to explore the following aspects by considering the cash back effect. From an analytical perspective, firstly, the network research consists of multi-hotels and multiOTAs in the cooperation is worth considering. Secondly, considering the hotel room as a part of tourism package products with cash back effect is also an interesting topic. Thirdly, an optimal overbooking strategy of a hotel under the cooperation with this type of OTAs will also be welcomed. There are also some worthy future studies for empirical researchers on this topic. The OTAs do not provide cash back on all the hotels. For example, neither eLong nor Ctrip provides cash back for the tourists of Hotel ICON that is a famous hotel established by School of Hotel and Tourism Management Hong Kong Polytechnic University. This is to say, the practice tells us that the cash back policy is not always the best strategy for all the OTAs and hotels. Hence, to verify the effective factors of cash back application is important for the managers and the results are expected by both the OTAs and hotels. Acknowledgments This work was supported by the National Natural Science Foundation of China (Nos. 71301156; 71271197), the Foundation for Innovative Research Groups of the National Natural Science Foundation of China (No. 71121061), the Foundation for International Cooperation and Exchange of the National Natural Science Foundation of China (No. 71110107024), the China Postdoctoral Science Foundation (Nos. 2014M560523; 2013M541845). Appendix A. Proofs A.1. Proof of Proposition 1 Firstly, considering that the objective total profit of the integrated player is a decreasing function of the value of cash back. As a result, it is the best choice to set φ as small as possible, so the optimal decision on cash back is setting φ equals Δs. Secondly, from the first-order condition of the objective function with respect to ξ, we have: Z ∂Π=∂ξ ¼ p0

C

f ðx−ξÞdx−kξ ¼ 0:

0

It is p0(F(C − ξ) − F(−ξ)) − kξ = 0, hence, we have ξ

FB

¼ arg f F ðC−ξÞ ¼ kξ=p0 þ F ð−ξÞg: ξ

■ A.2. Proof of Corollary 1 From F(C − ξFB) = kξFB/p0 + F(−ξFB), we have:     FB   FB k ∂ξ FB FB FB ∂ξ FB ∂ξ f C−ξ −f C−ξ ¼ −f −ξ ; p0 ∂C ∂C ∂C

  FB   FB k ∂ξ FB kξ FB FB ∂ξ FB ∂ξ ¼ − 2 − f −ξ ; −f C−ξ p0 ∂p0 ∂p0 ∂p0 p0

X. Guo et al. / Tourism Management Perspectives 12 (2014) 104–112

and   FB ξ FB ξ FB ∂ξ FB   FB FB ∂ξ FB ∂ξ −f C−ξ þ ¼ − f −ξ : p0 p0 ∂k ∂k ∂k Then, we can have Corollary 1 as follows from these three equations: FB   k     ∂ξ FB FB FB = −f −ξ N0; þ f C−ξ ¼ f C−ξ p0 ∂C

      ∂ξ FB FB 2 FB FB ¼ kξ = kp0 þ p0 f C−ξ − f −ξ N0; ∂p0 and       ∂ξ FB FB FB FB FB −f −ξ b0: ¼ −ξ = ξ þ p0 f C−ξ ∂k ■ A.3. Proof of Proposition 2 The first-order conditions of the OTA's objective function with respective of φ and ξ are:    Z C ∂πo Δs F ðx−ξÞdx C− ¼ − 1− φ  ∂φ  Z0 C Δs F ðx−ξÞdx þ ðω−φÞ 2 C− φ 0 ¼ 0; and  Z ∂πo Δs ¼ ðω−φÞ 1− φ ∂ξ

C

f ðx−ξÞdx−kξ ¼ 0:

0

Then we can obtain the optimal responses of the OTA as: 

φ ¼

  pffiffiffiffi pffiffiffiffiffiffi2 pffiffiffiffiffiffiffiffiffiffi  Δsω and ξ ¼ arg F ðC−ξÞ ¼ kξ= ω− Δs þ F ð−ξÞ : ξ

■ A.4. Proof of Corollary 2 From Proposition 2, we can get ∂φ*/∂ω N 0 directly as: ∂φ =∂ω ¼ pffiffiffiffiffiffiffiffiffiffiffiffi Δs=ω=2N0. pffiffiffiffi pffiffiffiffiffiffi2 



 Secondly, from F C−ξ ¼ kξ = ω− Δs þ F −ξ , we have: 

pffiffiffiffi pffiffiffiffiffiffi2 ∂ξ 





− f C−ξ −f −ξ ω− Δs þ F C−ξ −F −ξ ∂ω pffiffiffiffi pffiffiffiffiffiffi 1  ω− Δs pffiffiffiffi ω ∂ξ ¼k ; ∂ω pffiffiffiffiffiffiffiffiffi

   ð F ðC−ξ Þ− F ð−ξ ÞÞ 1− Δs=ω ¼ further, we can get ∂ξ pffiffiffi pffiffiffiffi 2 N 0 from the above   ∂ω ðkþ f ðC−ξ Þ− f ð−ξ ÞÞð ω− ΔsÞ equation. ■ References Anderson, C. (2011). Search, otas, and online booking: An expanded analysis of the billboard effect. Cornell Hospitality Report, 11(8) (http://www.hotelschool.cornell.edu/ research/chr/pubs/reports/abstract-15540.html). Bastakis, C., Buhalis, D., & Butler, R. (2004). The perception of small and medium sized tourism accommodation providers on the impacts of the tour operators' power in eastern mediterranean. Tourism Management, 25(2), 151–170.

111

Campo, S., & Yagüe, M. J. (2007). The formation of the tourist's loyalty to the tourism distribution channel: How does it affect price discounts? International Journal of Tourism Research, 9(6), 453–464. Casadesus-Masanell, R., & Yoffie, D. B. (2007). Wintel: Cooperation and conflict. Management Science, 53(4), 584–598. Chevalier, J. A., & Mayzlin, D. (2006). The effect of word of mouth on sales: Online book reviews. Journal of Marketing Research, 43(3), 345–354. Ding, W. (2011). Challenges and opportunities of hotel online booking in china. Design, user experience, and usability. Theory, methods, tools and practice (pp. 3–12). Springer. García, D., & Tugores, M. (2006). Optimal choice of quality in hotel services. Annals of Tourism Research, 33(2), 456–469. Göymen, K. (2000). Tourism and governance in turkey. Annals of Tourism Research, 27(4), 1025–1048. Gu, Z. (1997). Proposing a room pricing model for optimizing profitability. International Journal of Hospitality Management, 16(3), 273–277. Guo, X., & He, L. (2012). Tourism supply-chain coordination: The cooperation between tourism hotel and tour operator. Tourism Economics, 18(6), 1361–1376. Guo, X., Ling, L., Dong, Y., & Liang, L. (2013). Cooperation contract in tourism supply chains: The optimal pricing strategy of hotels for cooperative third party strategic websites. Annals of Tourism Research, 41, 20–41. Huang, Z., & Li, S. X. (2001). Co-op advertising models in manufacturer–retailer supply chains: A game theory approach. European Journal of Operational Research, 135(3), 527–544. Jaramillo, F., & Mulki, J. P. (2008). Sales effort: The intertwined roles of the leader, customers, and the salesperson. Journal of Personal Selling and Sales Management, 28(1), 37–51. Kim, J., Bojanic, D. C., & Warnick, R. B. (2009). Price bundling and travel product pricing practices used by online channels of distribution. Journal of Travel Research, 47(4), 403–412. Kim, Y. G., & Li, G. (2009). Customer satisfaction with and loyalty towards online travel products: A transaction cost economics perspective. Tourism Economics, 15(4), 825–846. Lai, K. -K., & Ng, W. -L. (2005). A stochastic approach to hotel revenue optimization. Computers & Operations Research, 32(5), 1059–1072. Ling, L., Guo, X., & He, L. (2012). Optimal pricing strategy of hotel for long-term stay. International Journal of Services, Technology and Management, 17(1), 72–86. Ling, L., Guo, X., & Liang, L. (2009). Optimal pricing strategy of hotel for cooperative travel agency. Computational intelligence and software engineering, 2009. CISE 2009. International conference on (pp. 1–5). Wuhan, China: IEEE Computer Society. Ling, L., Guo, X., & Liang, L. (2011). Optimal pricing strategy of a small or medium-sized hotel in cooperation with a web site. Journal of China Tourism Research, 7(1), 20–41. Little, J.D. C. (1979). Aggregate advertising models: The state of the art. Operations Research, 27(4), 629–667. Nash, J. F. (1950). The bargaining problem. Econometrica, 18(2), 155–162. Pan, C. -M. (2007). Market demand variations, room capacity, and optimal hotel room rates. International Journal of Hospitality Management, 26(3), 748–753. Schulz, C. (1994). Hotels and travel agents: The new partnership. The Cornell Hotel and Restaurant Administration Quarterly, 35(2), 45–50. Schwartz, Z. (2006). Advanced booking and revenue management: Room rates and the consumers' strategic zones. International Journal of Hospitality Management, 25(3), 447–462. Shapley, L. S. (1953). A value for n-person games. In H. W. K. a. A. W. Tucker (Ed.), Contributions to the theory of games, Vol. 2. (pp. 307–317). Princeton: Princeton University Press. Shin, H., & Benton, W. C. (2007). A quantity discount approach to supply chain coordination. European Journal of Operational Research, 180(2), 601–616. Shon, Z. -Y., Chen, F. -Y., & Chang, Y. -H. (2003). Airline e-commerce: The revolution in ticketing channels. Journal of Air Transport Management, 9(5), 325–331. Thomson, G. (2012). The new age of online hotel reservations. (In http://blog.hftp.org/thenew-age-of-online-hotel-reservations/ (Vol. 2013)). Toh, R. S., DeKay, C. F., & Raven, P. (2011). Travel planning searching for and booking hotels on the internet. Cornell Hospitality Quarterly, 52(4), 388–398. Toh, R. S., Raven, P., & DeKay, F. (2011). Selling rooms: Hotels vs. third-party websites. Cornell Hospitality Quarterly, 52(2), 181–189. van der Rest, J. -P. I., & Harris, P. J. (2008). Optimal imperfect pricing decision-making: Modifying and applying Nash's rule in a service sector context. International Journal of Hospitality Management, 27(2), 170–178. Wu, E. H., Law, R., & Jiang, B. (2013). Predicting browsers and purchasers of hotel websites: A weight-of-evidence grouping approach. Cornell Hospitality Quarterly, 54(1), 38–48. Yoon, M. G., Yoon, D. Y., & Yang, T. W. (2006). Impact of e-business on air travel markets: Distribution of airline tickets in Korea. Journal of Air Transport Management, 12(5), 253–260. Zhang, X., Song, H., & Huang, G. Q. (2009). Tourism supply chain management: A new research agenda. Tourism Management, 30(3), 345–358.

Xiaolong Guo received his PhD degree from the University of Science and Technology of China and currently a postdoc researcher of the School of Management, University of Science and Technology of China. His research interests include logistics and supply chain management, service operations research and hospitality & tourism management.

112

X. Guo et al. / Tourism Management Perspectives 12 (2014) 104–112 Xiabing Zheng received her PhD from the joint program between City University of Hong Kong Joint and University of Science and Technology of China (USTC-CityU Joint Advanced Research Center), and is currently a postdoc-researcher at the University of Science and Technology of China. Her research interests include information systems, supply chain management and e-commerce.

Liuyi Ling holds a PhD from the University of Science and Technology of China and is now a senior lecturer of Management Science at School of Management, University of Science and Technology of China. His research interests include tourism supply chain management, tourism economics, pricing and revenue management, and tourism strategic management.

Chenchen Yang received her doctoral degree of management science and engineering from the University of Science and Technology of China in 2012, and is now a lecturer of the School of Economics, Hefei University of Technology. Her research interests include decision science, DEA and performance evaluation.