Sunteți pe pagina 1din 10

Tourism Management 62 (2017) 302e311

Contents lists available at ScienceDirect

Tourism Management
journal homepage:

Pricing strategies of tour operator and online travel agency based on

cooperation to achieve O2O model
Yong Long, Pingping Shi*
School of Economics and Business Administration, Chongqing University, Chongqing 400030, PR China

h i g h l i g h t s

 Study the pricing games based on cooperation between the TO and OTA to achieve O2O model.
 Find the optimal pricing policy for the TO and OTA to maximize the revenue.
 Compare the optimal pricings and revenues in the Stackelberg and Bertrand game.
 Give suggestions to TOs and OTAs on how to cooperate.

a r t i c l e i n f o a b s t r a c t

Article history: This paper studies the optimal pricing strategies of a tour operator (TO) and an online travel agency
Received 23 December 2016 (OTA) when they achieve the O2O model through online sale and offline service cooperation. By con-
Received in revised form structing a competition model, cooperation conditions, pricing strategies and revenues are analyzed and
1 May 2017
compared in the Stackelberg and Bertrand game. Results indicate that service level, unit sale commission,
Accepted 3 May 2017
Available online 15 May 2017
service cost coefficient and unit service compensation coefficient have different influences on the TO's
and OTA's pricing decisions. When the unit sale commission is greater than the threshold, the TO's and
OTA's pricing in the Bertrand game are higher than in the Stackelberg game. Being a leader is the
dominant strategy for the TO. In addition, the revenues of TO and OTA in sale and service cooperation are
Pricing strategies analyzed by numerical examples and some suggestions for establishing cooperation contract are
O2O model provided.
Tour operator © 2017 Elsevier Ltd. All rights reserved.
Online travel agency

1. Introduction development. However, the key strength of TOs is their ability to

provide professional information and personalized advice to trav-
In tourism supply chain, TOs assemble tourism attractions, elers on a continuous basis (Walle, 1996). A few large OTAs, which
restaurants, hotels, transportations and its own special services to accumulated experience and know-how of e-commence operation,
form packages with competitive price, which are convenient when strengthen their competitive advantages in the market and build
travelers purchase (Clerides, Nearchou, & Pashardes, 2008; Yang, up higher entry barriers for new-comers (Huang, 2006). The
Huang, Song, & Liang, 2009). With the rapid development of advice-offering capability of TOs and the attributes of OTAs such as
internet and information technology, more and more TOs are of- security, ease use are the important factors which will affect the
fering a point of contact via the World Wide Web, which enables travelers decisions in choosing their packages (Bennett & Lai, 2005;
travelers to search for appropriate travel products and services Kim et al., 2007; Law, Leung, & Wong, 2004). In China, the coop-
(Kim, Kim, & Han, 2007). As a matter of fact, due to the restrictions eration between TOs and OTAs has occurred, such as ZhongXin TO
of online marketing base, technology, human and other resources, and Uzai (, JingJiang TO and Lvmama
TOs are difficult to attract travelers and achieve online profits (Yao, ( Through establishing online sale and
Ma, & Li, 2014). Meanwhile, OTAs are unable to provide the cor- offline service cooperation, the TOs and OTAs can play resource
responding store service for packages, which inhibits their further advantages simultaneously to achieve O2O model. The O2O model
includes traditional and e-commerce distribution channels which
can supplement each other to provide travelers with the greatest
* Corresponding author. satisfaction (Law et al., 2004; Lu & Liu, 2016), and provide a
E-mail address: (P. Shi).
0261-5177/© 2017 Elsevier Ltd. All rights reserved.
Y. Long, P. Shi / Tourism Management 62 (2017) 302e311 303

business opportunity for the further development of both TO and the retailer provides the same level of service in both channels,
OTA. Therefore, what is the cooperation condition for the TO and Yang and Zhang (2014) pose a service cooperation incentive
OTA? How do the TO and OTA design the cooperation contract? mechanism in a dual-channel supply chain under service
How do cooperation parameters impact on the TO's and OTA's differentiation.
pricing decisions, demands and revenues? The purpose of this However, it is hard for small or medium-sized or even some
paper is to discuss these issues. large tourism firms to run their own direct online channels inde-
The O2O model, as a new e-commerce business model, pendently because of the low popularity, lack of e-commence
combining both online trading and offline experience, has become operating experience and so on. For example, China Youth Travel
an important strategy for development of enterprises in recent Service opened up its online channel AoYou (http://www.aoyou.
years (Lu & Liu, 2016). As the major characteristic feature of tourism com/) which had been continuously unprofitable for 8 years.
products is that production and consumption happen simulta- Online channels play a crucial role in the service and hospi-
neously, the cooperation between OTAs and hotels or airlines tality industry. According to Wu, Law, and Jiang (2013), one-third
makes the tourism O2O model rapid development. Quite a few of the outbound travelers in HongKong search for hotel informa-
studies have examined the cooperation problem between hotels or tion online, and among them approximately 50% make room
airlines and OTAs and have given suggestions on how to achieve reservations through the websites from which they obtain the
cooperation (Koo, Mantin, & O'Connor, 2011; Guo, Ling, Dong, & information. Furthermore, from a survey of 249 leisure travelers,
Liang, 2013; Ling, Guo, & Yang, 2014; Dong & Ling, 2015; Guo, Toh, Dekay, and Raven (2011) find that 80 percent of the travelers
Zheng, Ling, & Yang, 2014; Xu, He, & Hua, 2014). Although the search for hotel information using web tools, with more than half
importance of TO's advice-offering and OTA's attributes in travelers' making their booking through hotels' host websites or third-party
booking has been demonstrated by some scientific researchers, websites (i.e., OTAs). Yoon, Toon, and Yang (2006) study the
little literature in the hospitality and tourism fields has studied the impact of e-business on the distribution of airline tickets in Korea
pricing problem of cooperation between TOs and OTAs to play and point out that consumers may be disinterested in the Website
advantages of both and achieve O2O model. of an individual service firm because of its small scale, especially
To fill this gap and provide some suggestions for TOs and OTAs in the hospitality industry, and suggest that airlines pay more
managers on establishing the O2O model through cooperation, in attention to cooperation with third-party websites. Some re-
this paper a competition model is proposed to describe decision searchers have studied the cooperation problems between hotels
interactions consisting of a TO and an OTA. The TO establishes or airlines and the third party website (Koo et al., 2011; Guo et al.,
online sale cooperation with the OTA, and the OTA establishes 2013; Ling et al., 2014; Dong & Ling, 2015; Guo et al., 2014; Xu
offline service cooperation with the TO at the same time. By et al., 2014). Koo et al. (2011) point out that the airlines are less
constructing a competition model based on O2O model, the likely to use OTA platforms if the airlines have a large loyal con-
cooperation conditions, pricing strategies and revenues of TO and sumer base or if the OTA platform is highly competitive. Guo et al.
OTA are analyzed and compared in the Stackelberg and Bertrand (2013) and Ling et al. (2014) study the optimal pricing strategy for
game. tourism hotels when they operate their online channel by coop-
The rest of this paper is organized as follows. Section 2 reviews erating with a third party website. Pricing is not only a key stra-
the related literature. Section 3 describes the non-cooperation and tegic lever deployed by hotels to manage revenue (Kimes & Chase,
cooperation between a TO and an OTA and obtains the model 1998) but also an important tool for building and enhancing
equilibriums in the Stackelberg and Bertrand game. Section 4 an- cooperation.
alyzes the model equilibriums and presents the numerical analyses Owing to the rapid development of economy and the innovation
results. Section 5 concludes this paper by summarizing some of the of information technology, travelers' consumption demand has
managerial implications obtained. become more diverse and OTAs cannot rely only on lowering the
price to attract consumers. By gathering and organizing informa-
2. Literature review tion, offline channels can provide scene experience, personal and
professional information and advice to meet travelers’ demands on
The O2O model involves online channel and offline service, a continuous basis (Kim et al., 2007; Walle, 1996). Despite the
which improves the value and feeling of customer experience as impact of technology and the advent of online bookings, this
well as improving the operational efficiency of enterprise value observation still stands, Bitner and Booms (1982) observe that
chain and utilization efficiency of social resource (Lu & Liu, 2016). A offline channels play a pivotal role in the tourism supply chain.
tourism online channel is a powerful platform for travelers to Moreover, opening traditional channels is also difficult for OTAs
search for a wide variety of travel-related information, and pur- due to the capital resource, professionals and so on. For example,
chase products or services conveniently (Dennis, Sandhu, & Harris, Uzai ( discontinued its offline channel, and
2002). Offline channels provide much more customized advice, formed an alliance with ZhongXin TO.
detailed product information by face-to-face counseling. Facing with fierce competition in tourism market, the TOs and
Considering the preferences of customers and the advantages OTAs should develop O2O model to provide new profitability op-
of offline and online channels, most of firms run dual-channel portunity and improve tourists' satisfaction. As a matter of fact, due
simultaneously. Multichannel coordination problems and sugges- to restriction of resources, the TOs and OTAs are difficult to achieve
tions have been studied. Online channel explored by manufactures O2O model separately. The OTAs have accumulated great number of
can reduce the retailer's market prices (Hendershott & Zhang, tourists, e-commence operating experience, know-how and other
2006), increase consumer welfare and expand sales (Chiang, resources while the TOs have many stores with experienced
Chhajed, & Hess, 2003), improve supply chain performance (Park salesmen who can provide professional information and advice to
& Keh, 2003), but pose a threat to retailers. In order to avoid travelers. Considering the resource advantages of both, the coop-
channel conflict and improve service efficiency, Xiao, Dan, and eration to achieve O2O model between the TOs and OTAs can
Zhang (2009 & 2010) study cooperation strategies between improve their whole competition advantage and weaken compe-
manufactures and retailers, in which the retailers finish all online tition. However, no study has addressed the problem of TOs' and
channels' order or manufactures outsource internet channels' OTAs’ O2O model by cooperating with each other. In this situation,
service to retailers. In consideration of the question that whether TOs and OTAs have little information and experience regarding the
304 Y. Long, P. Shi / Tourism Management 62 (2017) 302e311

cooperation and know little about the cooperation conditions. To According to the above description and notations, the TO's
enrich the scientific literature and provide vital suggestions to TOs revenue under the non-cooperation condition can be described by
and OTAs on how to achieve cooperation, this paper studies the the following:
optimal pricing strategies through the analysis of a competition !
model composed of a TO and an OTA. hT s21
PTNC ¼ p1  c 1  d1  cT dN
1: (3)
3. The model
The OTA's revenue under the non-cooperation condition can be
3.1. Model description and model constructing described by the following:
In this paper, the participants consist of a TO who provides his hO s22
PONC ¼ p2  c2  dN
2: (4)
package 1 at a unit cost c1 and an OTA who provides his package 2
2 at a unit cost c2 to the same tourist destination. Here the
subscript “i”(i ¼ 1,2) represents the kinds of packages. Before
achieving O2O model, travelers can purchase package 1 only at 3.1.2. Cooperation description
store, and enjoy the scene experience, personalized and profes- In the cooperation state, the TO and OTA establish online sale
sional travel advice, etc., and the services of package 1 is denoted and store service cooperation, that is, the TO commissions the OTA
by s1. In practice, a strictly convex service function c(s) is used to to sell package 1 and pays k (0 < k  cT) for unit package 1, where k
depict its unit cost of the TO service. That is, the cost for providing denotes unit sale commission. Therefore, the sale commission of
the services has properties of dcðsi Þ=dsi > 0 and d2 cðsi Þ=ds2i > 0. One OTA is PO N
1 ¼ kd1 . The OTA commissions the TO to provide services
form commonly adopted in previous literature is given as: s2 for the package 2. According to Xiao, Dan, and Zhang (2010), the
cðsi Þ ¼ hT s2i 2, where hT is the TO's service cost coefficient and
OTA pays ws22 (hT/2 < w  hO/2) for unit package 2, where w denotes
vcðs1 Þ= is the TO's marginal service cost for providing s1.
vs1 unit service compensation coefficient. Therefore, the service reve-
Whereas the travelers can only purchase package 2 online, but
nue of TO is PTs2 ¼ ðws22  cðs2 ÞÞdN
cannot enjoy related services provided in-store. Facing with fierce
competition in tourism market, the TO and OTA decide to develop The TO's revenue under the cooperation condition is
O2O model, which is that the two packages can be bought online ! !
and provided service si. P ¼ p1  c1 
hT s21 2 =
d1  kdN
1 þ ws22  hT s22 2
2: (5)
The parameter a represents the base level of packages demands
of tourist destination. Let m (0 < m < 1) represent the ratio of online
demand. Correspondingly, 1-m represents the ratio of offline de- The OTA's revenue under the cooperation condition is
mand. Let gi (0 < gi < 1) denote the ratio of package i online de-  
mand. (1-m)a represents the offline demand base level of package 1, PO ¼ p2  c2  ws22 dN2 þ kdN1 : (6)
mgia represents the online demand base level of package i. Let dF1
denote the offline demand of package 1, dN i denote the online de-
Here the superscript “T”, “O”, “N”, “F”, “TS”, “B” represents the pa-
mand of package i, respectively. d1 (d1 ¼ dF1 þ dN 1 ) denotes total
rameters corresponding to the TO, the OTA, the online, the offline,
demand of package 1. bp and bs measure the responsiveness of the Stackelberg game and the Bertrand game. The subscript “NC”
packages market demand to its own price and service, respectively. represents the non-cooperation condition.
bp and bs are measures of the intensity of competition between the
two packages with regards to pricing and service, respectively, 3.2. Model equilibrium
where bp > bp > 0, bs > bs > 0.
Linear demand functions have been adopted in Lu, Tsao, and According to the revenues PTNC and PO NC under the non-
Charoensiriwath (2011) and Tsay and Agrawal (2000), and many cooperation condition and the revenues PT and PO under the
others. The corresponding demand functions to package i are cooperation condition, the non-cooperative and cooperative
described as follows. optimal pricing equilibriums are obtained in the Stackelberg and
Offline demand of package 1 is Bertrand game.

dF1 ¼ ð1  mÞa  bp p1 þ bp ðp2  p1 Þ þ bs s1  bs ðs2  s1 Þ: (1) 3.2.1. Stackelberg game

The problem is analyzed as a Stackelberg game where the TO
Online demand of package i is
acts as the leader and the OTA as the follower. As a first mover, the
    TO announces his pricing policy and based on it the OTA decides its
i ¼ gi ma  bp pi þ bp pj  pi þ bs si  bs sj  si :
dN (2) pricing decisions simultaneously. The TO optimizes its pricing
policy considering the rational behavior of OTA and maximizes its
where a > 0, 0 < m < 1, 0 < gi < 1, bp > bp > 0, bs > bs > 0, i ¼ 1,2 and revenue, while the objective of OTA is to maximize its own revenue
j ¼ 3-i. corresponding to the pricing policy of TO. The equilibrium point
that specifies the pricing policy of TO and OTA is found out by the
3.1.1. Non-cooperation description solution of a two-stage non-cooperative game.
In the non-cooperation state, the TO and OTA develop O2O Proposition 1 In the Stackelberg game,
model separately, in which the TO runs his online channel and costs
cT for unit online demand and the OTA opens stores to provide (i) the optimal pricing decisions of TO and OTA under the non-
offline service s2. Considering the service cost function is similar to cooperation condition are
TO's, the OTA will cost hO s22 2 for proving service s2, where hO is the

OTA's service cost coefficient. Because the TO has the advantage of A1 hT s21 þ 2A2 s1 þ A4 hO s22 þ 2A4 s2 þ 2A5 þ A1 cT
being close to travelers, the TO's cost of providing the same service
1N ¼ ; (7)
is lower than OTA's, that is hO > hT.
Y. Long, P. Shi / Tourism Management 62 (2017) 302e311 305

A1 hT bp s21 þ B5 s22 þ 2B2 s1 þ 2B3 s2 þ A1 bp 2c1 þ cT þ 2B5 c2 þ 2B6
2N ¼
  : (8)
8A1 bp þ bp

(ii) the optimal pricing decisions of TO and OTA under the

cooperation condition are

2A1 hT s21 þ A3 ð6w  hT Þs22 þ 2 A1 þ 2b2p k þ 4ðA2 s1 þ A4 s2 þ A5 Þ
1 ¼ ; (9)

2C1 hT s21 þ C2 s22 þ 2A2 s1 þ 2A4 s2 þ 2C3 k þ 2C4

1 ¼ 2
; (13)
2A1 hT bp s21 þ B1 s22 þ 2 5A1 þ 2b2p bp k þ 4ðB2 s1 þ B3 s2 þ B4 Þ 8b2p þ 16bp bp þ 6bp
2 ¼
16A1 bp þ bp
2A3 hT s21 þ D1 s22 þ 4A4 s1 þ 4A2 s2 þ 10A3 k þ 2D2
2 ¼ : (14)
(10) 16b2p þ 32bp bp þ 12bp

where, A1 ¼ 2b2p þ 4bp bp þ b2p , A2 ¼ 2bp bs þ 2bp bs þ 2bs bp þ bp bs ,

where, C1 ¼ b2p þ 2bp bp þ b2p , C2 ¼ 4bp bp w þ 4b2p w  bp bp hT b2p hT ,
A3 ¼ bp ðbp þ bp Þ,
C3 ¼ b2p þ 2bp bp þ 2b2p ,
A4 ¼ bs bp  2bp bs  bp bs ,
C4 ¼ 2ðbp þ bp Þ2 c1 þ A3 c2 þ bp ð1  m þ g1 mÞa þ bp a,
A5 ¼ bp ð1  m þ g1 mÞa þ bp a þ A1 c1 þ A3 c2 ,
D1 ¼ 8b2p w þ 16bp bp w þ 10b2p w  b2p hT ,
B1 ¼ ð8b2p hT þ 32bp bp w þ 14b2p w  b2p hT Þðbp þ bp Þ;
D2 ¼ 2A3 c1 þ 4C1 c2 þ ð4bp g2 m þ bp þ 3bp g2 mÞa.
B2 ¼ 2bp bs bp  6bp bp bs þ 2bs b2p  4b2p bs  b2p bs ;
B3 ¼ 4b2p bs þ 8bp bs bp þ 3bs b2p þ 4b2p bs þ 6bp bp bs þ b2p bs ,
B4 ¼ 2A1 ððbs þ bp Þc2 þ g2 maÞ þ bp A5 , 4. Equilibrium and numerical analysis
B5 ¼ 3b3p þ 4b3p þ 11bp b3p þ 12b3p bp
B6 ¼ A3 a þ ð2A1  bp bp Þg2 ma: 4.1. Cooperation condition

Because the parameters are positive in section 3.1, such as bp > Proposition 3: In the Stackelberg game, when the ratio of
bp > 0, bs > bs > 0, the values of A1, A2, A3, A5, B1, B3, B4, B5, B6 must package 1 online demand g1 meets gTS 1min  g1  g1max , the TO and

be positive. However, the values of A4, B2 may be positive or OTA can cooperate. In the Bertrand game, when g1 meets
negative. gB1min  g1  gB1max , the cooperation can also be achieved.

2A1 hT s1 þA3 ðhT þ2wÞs2 þ4A2 s1 þM1 s2 þM2 kþM3

2 2
3.2.2. Bertrand game gTS
¼ 4ð3b þ2b Þma
p p
In the Bertrand game, the procedure is as follows: the TO de- 2A1 bp hT s21 þN1 s22 þN2 s1 þN3 s2 þN4 kþN5 þN6
termines package 1's price, so as to maximize its revenue. The OTA gTS
1max ¼ ;
4ð4b2p þ7bp bp þ2b ma

decides the price package 2's price to maximize its revenue without 2
M1 ¼ 8bs bp þ 4bs bp þ 4bs bp , M2 ¼ 2ð2b2p þ 4bp bp  bp Þ,
knowing the price of package 1.
Proposition 2 In the Bertrand game, M3 ¼ 4ð4bp bp 2b2p  b2p Þc1 þ4A3 c2 þ4ðmbp ðbp þ bp Þþ2mbp Þa,
N1 ¼ 16wb3p  2hT b3p  2wb3p  48wb2p bp  34wbp b2p  hT bp b2p
(i) the optimal pricing decisions of TO and OTA under the non-
cooperation condition are N2 ¼ 16b2p bs  4b2p bs  12bp bp bs þ 8bs b2p þ 8bs bp bp ,
N3 ¼ 4b2p ðbs þ 3bs Þ þ 16ðbs þ bs Þb2p þ 4bp bp ð3bs þ 8bs Þ,
2C1 hT s21 þ A3 hO s22 þ 2A2 s1 þ 2A4 s2 þ 2C4 þ 2C1 cT N4 ¼ 24bp b2p  12b2p bp  2b3p ,
1N ¼ 2
8b2p þ 16bp bp þ 6bp N5 ¼ ð16bp b2p þ 8b2p bp þ 2b3p Þc1  ð48b2p bp þ 36bp b2p þ 4b3p
(11) þ8b3p Þc2 ,
N6 ¼ 4b2p a þ 4ð2b2p þ 4b2p þ 7bp bp Þma þ 4bp bp a
A3 hT s21 þ 2C1 hO s22 þ 2A4 s1 þ 2A2 s2 þ D2 þ A3 cT
2N ¼ : (12)
8b2p þ 16bp bp þ 6b2p In addition, the gB1min and gB1max are not list.
Because the parameters are positive in section 3.1, such as bp >
bp > 0, bs > bs > 0, the values of M1, N3, N6 must be positive and that
(ii) he optimal pricing decisions of TO and OTA under the of M2, N1, N4 must be negative. However, the values of M3, N2, N5
cooperation condition are may be positive or negative.
306 Y. Long, P. Shi / Tourism Management 62 (2017) 302e311

When g1 meets a certain condition, the online demand of higher, the TO faces service cost pressure with s1 increasing, which
package i is non-negative. The OTA can obtain sale commission and leads to demand of package 2 increase, while at the same time s2
the TO can gain service revenue, in which case, the cooperation remain unchanged, so PTs2 increases. It also reveals that when
model makes economic sense and the TO and OTA would like to
marginal service cost is high, PTs2 can be added and d1 cannot be
increased by improving service s1
Corollary 2 In the Stackelberg game, when0 < k  kTS , vk1  0;
4.2. Analysis and comparison equilibriums in two games whenk > kTS , vk1 < 0.
In the Bertrand game, when 0 < k  kB , vk
 0; when
In the case of meeting the cooperation condition, the proposi- vP
k > kB , vk1
< 0.
tions 4e7 and corollaries 1e2 can be obtained.
vpi vpi vpi
Proposition 4 (i) vsi > 0; vk
> 0; vw > 0.
G1 s21 þG2 s22 þG3 s1 þG4 s2 þG5 þG6
kTS ¼ ,
4ð2b2p þ4bp bp bp Þ
TS*ðB*Þ 2
vpi vpj vpB* vpB*
(ii) vsi > vsi ; 1
> vk2 ; vw < vw .
G1 ¼ 2b2p w  12b2p  6b2p  24bp bp ,
Proposition 4 (i) shows that the package price pi increases with G2 ¼ 3ðb2p þ bp bp Þ þ 2bp bp w,
increasing service si, unit sale commission k and unit service G3 ¼ 8bs ðbp þ bp Þ þ 4bs ðbp þ 2bp Þ, G4 ¼ 4ðbs  bs Þbp  8bs bp ,
compensation coefficient w. This is straightforward because an
G5 ¼ 4ðbp bp þ b2p Þc2  20b2p  40b2p  80bp bp ,
increase in service and unit service compensation coefficient
intuitively means an increase in unit services cost, so pi will in- G6 ¼ 4ð2bp þ bp Þma þ 4ð2bp þ 3bp Þg1 ma  4ðbp þ bp Þa.
crease. That means even if the TO and OTA are in service compe-
tition, travelers cannot enjoy free service. In addition, an increase in In addition, the kB is not list.
the unit sale commission means an increase in sales cost, so the TO Corollary 2 shows that if the unit sale commission k is less than a
can gain revenue only by increasing price. Proposition 4 (ii) in- threshold, the sale commission PO
1 increases with the k increasing.
dicates that the rate of change of p1 with respect to s1 and k is larger Correspondingly, if k is more than the threshold, then PO
1 decreases
than that of p2, and the rate of change of p2 with respect to s2 and w with the k increasing. When k is large, the rises in price lead to a
is larger than that of p1. This means the TO and OTA will take the
1 . Owing to the negative effect of d1 on P1
decrease in demand dN N
same price strategy, but the degree of price change is different. For
example, when s1 increases, the TO must take a higher pricing more than positive effect of k, PO
1 decreases with k increasing. In the
strategy to maintain revenue, meanwhile, the OTA also increases p2 sale cooperation, the OTA cannot add PO
1 by asking for a higher k.
to compete with the TO, the rate of change of p2 with respect to s1 is vp
Proposition 5 When s1  s2 , vi h > 0.
smaller than that of p1. Similarly, the rate of change of pi with T
Proposition 5 shows that when s1  s2 , the optimal pricing de-
respect to s2, k, w can also be analyzed. cisions of both TO and OTA are positively related with the TO's
From proposition 4, the corollaries1-2 can be obtained. service cost coefficient. In order to keep lower price, the OTA should
Corollary 1. In the Stackelberg game, when B2 < 0 and select the TO with lower service cost coefficient as the partner. The
b2p bs þ bp bp bs < bp bs bp þ bs b2p , the relationships can be obtained by TO also improves package 1's price as unit service cost increases.
the following. Proposition 6 In the Stackelberg game, when B2 < 0 and
vPs 0 < vcðs1Þ
(i) if 0 < vcðs 1Þ
vs1 <  A b ,  0; in the Bertrand game, when A4 < 0 and
B2 vd1 2

vs1 <  A b , vs1 > 0 and vs1 < 0;

1 p vs1
1 p
(ii) if ABb2  vcðs1Þ A2 vd1 0 < vcðs1Þ
 AA43 , vs21  0.
vs1  A1 ,
1 p vs1 0 and vs12  0; vs1

T Proposition 6 shows that the price of package 2 is not only
vcðs1 Þ A2 vd1
vs1 > A1 , vs1 < 0 and > 0.
(iii) if vs1 influenced by s1, but also affected by the marginal service cost of TO.
When TO's marginal service cost is low, the TO has an obvious
In the Bertrand game, when A4 < 0. service cost advantage when s1 increases. The OTA only take lower
price strategy to compete with the TO under s2 unchanged.

(i) if 0 < vcðs 1Þ A4

vs1 <  A3 ,
vs1 > 0 and vs1 < 0;
Proposition 7 There is a threshold of the unit sale commission
vPs k0 , and the relationships can be obtained by the following.

(ii) if AA43  vcðs 1Þ A2

vs1  A1 ,
vs1 0 and vs12  0;
vcðs1 Þ A2 vd1 (i) pTS* < pB* TS* ¼ pB* , if k ¼ k ; pTS* > pB* , if k < k ;
(iii) if vs1 > A1 , vs1 < 0 and > 0. i , if k > k0 ; pi
vs1 i i 0 i i 0
TS* B* TS* B*
(ii) dTS* B* N
1 > d1 and d2 < dN TS* B* N
2 , if k > k0 ; d1 ¼ d1 and d2 ¼ dN
2 ,
Corollary 1 (i) shows that if the marginal service cost is positive if k ¼ k0 ; dTS* < dB*and dN
> dN
1 1 2 2 , if k < k0 ;
and less than a threshold, the total demand of package 1 increases TS*
* TS* B *

and the TO's service revenue decreases with the service s1 (iii) PT >P , if ksk0 ; PT ¼ PT , if k ¼ k0
increasing. (ii) indicates that if the marginal service cost is in a
certain range, the total demand of package 1 and the TO's service where,
revenue increase with the service s1 increasing. (iii) shows that if
T1 s21 þT2 s22 þT3 s1 þT4 s2 þT5
the marginal service cost is higher than the threshold, then the total k0 ¼ ,
2bp ðbp 4bp bp 2b2p Þ

demand of package 1 decreases and the TO's service revenue in-

creases with the service s1 increasing. This is straightforward T1 ¼ 4hT bp b2p þ 8hT bp b2p þ 2hT b3p ,
2 2
because an increase in s1 means an increase in package 1's price T2 ¼ 12ð2w  hT Þbp b2p  9hT bp bp þ 14bp bp w þ 4ð2w  hT Þb3p
(see Proposition 4), the promotion effect of s1 on d1 is greater than
ð2w þ hT Þb3p ,
inhibitory effect of p1 on d1, so the total demand of package 1 de-
creases. In addition, when the marginal service cost of package 1 is
Y. Long, P. Shi / Tourism Management 62 (2017) 302e311 307

T3 ¼ 8bp bs bp  8bp bs bp  4bs b2p  8bs b2p , 4.3. Numerical examples

T4 ¼ 8bp bs bp þ 4ðbs  bs Þ b2p ,

  In this section, the numerical examples are to illustrate the
T5 ¼ 4 4bp b2p þ b3p þ 2bp b2p c1 þ 4bp bp ma  4b2p a  4 OTA's preference game and the TO's and OTA's revenues changes
  under cooperation. Moreover, the numerical results whether
b3p þ bp b2p c2  4bp bp a  4bp bp g1 a. influenced by the base level of demand are also checked. Consid-
ering that the references on cooperation between the TO and OTA
Proposition 7 (i) compares the equilibrium prices in the Stack- are less and numerical examples are usually hypothetical, the
elberg and Bertrand game. It shows that if k is greater than k0, the values of parameters are assigned according to Xiao et al. (2010)
pricing is lower in the Stackelberg game than that in the Bertrand and Wu (2012) and also meet cooperation condition of Proposi-
game. Conversely, the pricing is higher if k is less than k0. When k is tion 3. A selected set of parameters is as follows: m ¼ 0.6, bp ¼ 4,
higher, the TO and OTA select the Stackelberg game in which they bp ¼ 2, bs ¼ 3, bs ¼ 1, c1 ¼ 5, c2 ¼ 4, k ¼ 6, s1 ¼ 2, s2 ¼ 1, hT ¼ 3, w ¼ 3,
can keep lower price. Proposition7 (ii) compares the equilibrium g1 ¼ 0.4, g2 ¼ 0.6. The results of numerical examples are summa-
total demand in the Stackelberg and Bertrand game. It shows that if rized in Figs. 1e8.
k is greater than k0, the demand of package 2 is higher in the Ber-
trand game than that in Stackelberg game, and the demand of
package 1 is lower in the Bertrand game than that in Stackelberg 4.3.1. Unit sale commission and the OTA's revenue
game. If k is less than k0, the finding is the opposite. Proposition7 Before analyzing the influence of unit sale commission on OTA's
(iii) shows that the TO as the leader in the Stackelberg game can revenue, the ranges of admissible value about a and k are given.
always benefit from k except the critical point k0. Compared with When the other parameters are constant, according to cooperation
TO's revenue in the Bertrand game, the TO can gain more in the condition of Proposition 3 in different competition games, the
Stackelberg game when k is not equal to k0. Thus, being a leader is formulas are obtained in Table 1.
the dominant strategy for the TO. On the basis of admissible values in Table 1, two graphical re-
gions are illustrated in two games as shown in Fig. 1. When the base
level of demand a increases, the unit sale commission k also in-
creases. This is because an increase in a means the TO can afford
more sale commission to open online market.
According to Table 1 and Fig. 1, when a ¼ 300, a ¼ 600 or
a ¼ 1200, the impacts of unit sale commission in the admissible
range on the OTA's revenue are shown in Fig. 2. Among Fig.2 (a)
~(c), the overall change trends are similar in the admissible
range, and the changes of a just affect the values of PO. When the
unit sale commission k increases, PO increases firstly and then
decreases. This is straightforward because an increase in k means
sale commission increase at first and then decrease (from Cor-
ollary 2), which leads to the similar change of PO. When k < km
Fig. 1. Admissible value regions of a and k in different competition games. (km is the intersection), the OTA gains more revenues in the

Fig. 2. Impacts of k on OTA's revenue PO.

Fig. 3. Admissible value regions of g1 and k.

308 Y. Long, P. Shi / Tourism Management 62 (2017) 302e311

Fig. 4. Impacts of g1 and k on TO's revenue PT.

Fig. 5. Impacts of g1 and k on OTA's revenue PO.

Fig. 6. Admissible value regions of s2 and w.

Fig. 7. Impacts of w and s2 on TO's revenue PT.

Stackelberg game than in the Bertrand game, so that the OTA 4.3.2. Revenue analysis in sale cooperation
prefers the Stackelberg game. When k > km, the OTA's revenue is When the other parameters are constant, according to the
larger in the Bertrand game than that in the Stackelberg game. cooperation condition of Proposition 3 in the Stackelberg game, the
When k ¼ km, the OTA's revenue is equal in two games. formulas of g1 and k are obtained in Table 2.
Y. Long, P. Shi / Tourism Management 62 (2017) 302e311 309

Fig. 8. Impacts of w and s2 on OTA's revenue PO.

Table 1 formulas of s2 and w are obtained in Table 3.

Ranges of admissible value about a and k in different competition games. Considering moral hazard and service ability, the TO will not
The competition games Stackelberg game Bertrand game provide higher service for the OTA than its own. On the basis of
Ranges of admissible value 0 < k  19.5 þ 0.1*a 0 < k  19.8 þ 0.1*a
admissible value in Table 3 and w  hT/2 ¼ 1.5, the shadow areas of
0 < k  18.3 þ 0.3*a 0 < k  18.2 þ 0.3*a Fig. 6 are the region for the range of admissible values of s2 and w.
The Districts 1e3 and s2 2½0; 1 which belong to the shadow areas
are used to numerical examples in this paper.
On the basis of ranges in Table 2, the shadow areas of Fig. 3 are When s2 and w are in the Districts 1e3 and s2 2½0; 1, the TO's
the region of admissible values about g1 and k. The Areas 1e3 and OTA's revenues are obtained as shown in Figs. 7 and 8.
which belong to the shadow area are used to numerical examples in Among Fig.7 (a)~(c) or Fig.8 (a)~(c), the changes of the base level
this paper. of demand a do not affect the overall trend and only affect the
When g1 and k are in the Areas 1e3, the TO's and OTA's revenues values of PT or PO. In Fig. 7, the overall trend of s2 on PT is
are obtained as shown in Figs. 4 and 5 respectively. increasing with the increase of w. However, when w and s2 are
Among Fig.4 (a)~(c) or Fig.5 (a)~(c), the overall change trends are lager, PT decreases slightly. In Fig. 8, the overall trend of s2 on PO
also similar in the admissible range. However, with the increase of increases at first and then decreases with the increase of w. How-
a, the range of k increases, and PT or PO change in varying degree. ever, when w and s2 are lager, PO increases slightly. When s2 in-
From the Fig. 4, the effect trend of k on PT changes from U-trend to creases within a small range near zero, the OTA can benefit from the
decrease trend with the increase of g1, and the decrease trend be- service competition and the revenue curve increase to a lesser
comes more significant. In Figs. 5 and 2, the effect trend of k on PO extent. When s2 has a larger increase, the service competition is
changes from inverted U-trend to increase trend with the increase enhanced between package 1 and 2. The service cost of TO is lower
of g1. By using the same analysis process of Fig. 2, the TO's and OTA's than that of OTA, which leads to the TO's revenue increase and the
revenues have a similar trend with sale commission when k in- OTA's revenue decrease. When s2 is near the maximum, the lager s2
creases. Meanwhile, the increase of g1 leads to the increase of and w lead to the faster price growth of OTA than that of TO. The
proportion with sale commission to the TO's and OTA's revenues, rapid increasing of OTA price can compensate for the loss of service
which causes the change trend more and more obviously. Though competition. By analyzing the change trend of revenues, the OTA
analyzing the change trend of revenues, the TO should establish should establish the contract of low service and low unit service
online sale cooperation with the low unit sale commission when compensation coefficient with the TO.
the ratio of package 1 online demand is larger. The OTA requires the The revenues in the Stackelberg game are discussed by nu-
unit sale commission which should be related to the ratio of merical examples and the similar conclusions in the Bertrand
package 1 online demand, and should not request the TO to pay game can also be obtained. Due to the length of this paper, it is
unreasonable high unit sale commission. omitted.

4.3.3. Revenue analysis in service cooperation 5. Conclusions and implications

When the other parameters are constant, according to the
cooperation condition of Proposition 3 in the Stackelberg game, the In this paper, a framework is developed to study online sale and

Table 2
Ranges of admissible value about g1 and k in different base level of demand.

Base level of demand a ¼ 300 a ¼ 600 a ¼ 1200

Ranges of admissible value 0 < k  31.5 þ 96*g1 0 < k  43.5 þ 192*g1 0 < k  67.5 þ 384*g1
0 < k  117.6 e 117.5*g1 0 < k  253.6 e 235.1*g1 0 < k  525.4 e 470.2*g1

Table 3
Ranges of admissible value about s2 and w in different base level of demand.

Base level of demand a ¼ 300 a ¼ 600 a ¼ 1200

Ranges of admissible value w 0.7/s22 þ0.7/s2 -1.5 w 132.7/s22 þ0.7/s2 -1.5 w  396.7/s22 þ0.7/s2 -1.5
w 18.5/s22 þ0.7/s2 -0.02 w 40.9/s22 þ0.7/s2 -0.02 w  85.6/s22 þ0.7/s2 -0.02
310 Y. Long, P. Shi / Tourism Management 62 (2017) 302e311

offline service cooperation to achieve the O2O model between a TO Proof of proposition 4
and an OTA. The results indicate that the price decisions are very
important for both TO and OTA. The optimal decisions for prices In the Stackelberg game,
and revenues of TO and OTA are examined in the Stackelberg game

vpTS* h s A v pTS*  p TS* 2bp þ bp A1 hT s1 þ I1
ðiÞ 1
¼ T 1 þ 2 > 0; ðiiÞ 1 2
¼   > 0:
vs1 2 2A1 vs1 4A1 bp þ bp

and Bertrand game, and the impacts of cooperation parameters on where, I1 ¼ 4b2p bs þ 8b2p bs þ 6bp bs bp þ 12bp bp bs þ 2bs b2p þ 3b2p bs .
cooperation condition, pricing decisions, demands and revenues The relationships of pi and k, pi and w in two games can be
are analyzed and compared. Some suggestions for establishing proofed by using the same method.
cooperation contract are provided.
The results show that cooperation can be achieved when the Proof of proposition 5
ratio of package 1 online demand meets a certain range. The
service level, unit sale commission, service cost coefficient and In the Stackelberg game,
unit service compensation coefficient have a great effect on TO and  
OTA price decision, however, the influences are different. The TO's
vpTS* 2A1 s21  A3 s22 vpTS* A1 bp s21 þ 4 bp þ bp b2p s22
service level and unit sale commission have a greater impact on 1
¼ ; 2
¼   :
TO's pricing. The OTA's service level and unit service compensa- vhT 8A1 vhT 4A bp þ b 1 p
tion coefficient have a greater influence on OTA's pricing. When
vpTS* vpTS*
the marginal service cost is high, the TO can improve service when s1  s2 , vh1 > 0, vh2 > 0.
revenue and cannot increase demand by improving TO's service. If In the Bertrand game, we can prove in the same way.
the OTA's service level is less than the TO's, the OTA's pricing is
positively associated with the TO's service cost. When the unit sale Proof of proposition 6
commission is greater than a threshold, TO can achieve more
revenues in the Stackelberg game, in which TO's and OTA's pricing In the Stackelberg game,
are lower. In addition, from the numerical studies, the TO is in- vpTS* A1 bp hT s1 þB2 vcðs1 Þ
Let vs1 ¼ 4A1 ðbp þbp Þ
 0, we can get hT s1  ABb2 , vs1  ABb2 .
clined to sign a low unit sale commission contract with the OTA 1 p 1 p
vcðs1 Þ A4
when the ratio of package 1 online demand is larger, and the OTA vs1  A3 can also be obtained in the Bertrand game by using
should sign a low service and low unit service compensation co- the same method.
efficient contract with the TO. The research results can provide
theoretical supports for TO and OTA on deciding pricing and Proof of proposition 7
establishing contract.
The results from this study can provide useful guidelines for (i) Let k ¼ k0 þ t, plugging k to pTS* B* TS* B*
1 , p1 , p2 ,p2 , we can get
cooperation to achieve O2O model. This study also demonstrates  
2 2
that the TO and OTA should fully consider the impact of various  2b2p þ 4bp bp  bp bp t
factors on pricing in the process of cooperation. It also reveals the pTS* B*
1  p1 ¼    ;
importance for the OTA to choose the TO with lower service cost as 4 2b2p þ 4bp bp þ bp 2bp þ bp 2bp þ 3bp
partner if the OTA wants to pay low service compensation and keep
lower price. Moreover, being a leader is the dominant strategy for  
 2b2p þ4bp bp  b2p b3p t
the TO, in which the TO and OTA also can take lower pricing stra-     :
pTS* B*
2 p2 ¼
tegic in a lager unit sale commission. 8 bp þ bp 2b2p þ4bp bp þ b2p 2bp þ bp 2bp þ3bp

Obviously, Proposition 7's (i) can be obtained.

TS* B*
(ii) Let k ¼ k0 þ t, plugging k to dTS* B* N N
1 , d1 , d2 , d2 , we can get
This work was supported by the Research Fund for the Doctoral   
Program of Higher Education of China (No. 201101191110031), 2b2p þ4bp bp þ b2p 2b2p þ4bp bp  b2p b2p t
Graduate and Innovation Foundation of Chongqing, China (No. dTS* B*
1 d1 ¼
CYB16004). 4 2b2p þ4bp bp þ bp bp þ bp 2bp þ bp 2bp þ3bp

2 3
Appendix TS* B*
 2b2p þ 4bp bp  bp bp t
2  dN
2 ¼    :
8 2b2p þ 4bp bp þ bp 2bp þ bp 2bp þ 3bp
Proof of proposition 3
Obviously, Proposition 7's (ii) can be obtained.
Plugging pTS* TS* back to dN , let dN  0,dN  0, we can get
i ,pj i i 2
g1min  g1 ,g1  g1max . In the same way, we can get gB1min ; gB1max in
(iii) Let k ¼ k0 þ t, According to equilibrium results and propo-
the Bertrand game. sition 7's (i) and (ii), we can get
Y. Long, P. Shi / Tourism Management 62 (2017) 302e311 311

2 2 4
* * 2b2p þ 4bp bp  bp bp t 2
P P ¼    2  2  0:
4 2b2p þ 4bp bp þ bp bp þ bp 2bp þ bp 2bp þ 3bp

TS* B*
Whenksk0 , obviouslyPT > PT . Walle, A. H. (1996). Tourism and the internet opportunities for direct marketing.
Journal of Travel Research, 35(1), 72e77.
Wu, C. H. (2012). Price and service competition between new and remanufactured
References products in a two-echelon supply chain. International Journal of Production
Economics, 140(1), 496e507.
Bennett, M. M., & Lai, C. K. (2005). The impact of the Internet on travel agencies in Wu, E. H., Law, R., & Jiang, B. (2013). Predicting browsers and purchasers of hotel
Taiwan. Tourism & Hospitality Research, 6(1), 8e23. websites: A weight-of-evidence grouping approach. Cornell Hospitality Quar-
Bitner, M. J., & Booms, B. H. (1982). Trends in travel and tourism marketing: The terly, 54(1), 38e48.
changing structure of distribution channels. Journal of Travel Research, 20(4), Xiao, J., Dan, B., & Zhang, X. M. (2009). Study on cooperation strategy between
39e44. electronic channels and retailers in dual-channel supply chain. Journal of Sys-
Chiang, W. K., Chhajed, D., & Hess, J. D. (2003). Direct marketing, indirect profits: A tems Engineering, 24(6), 673e679.
strategic analysis of dual-channel supply chain design. Management Science, Xiao, J., Dan, B., & Zhang, X. M. (2010). Service cooperation pricing strategy between
49(1), 1e20. manufactures and retailers in dual-channel supply chain. Systems Engineering-
Clerides, S., Nearchou, P., & Pashardes, P. (2008). Intermediaries as quality assessors: Theory & Practice, 30(12), 2203e2211.
TOs in the travel industry. International Journal of Industrial Organization, 26(1), Xu, L., He, P., & Hua, Z. (2014). A new form for a hotel to collaborate with a third-
372e392. party website: Setting online-exclusive-rooms. Asia Pacific Journal of Tourism
Dennis, C., Sandhu, B., & Harris, I. (2002). From bricks to clicks: Understanding the Research, 20(6), 635e655.
e-consumer. Qualitative Market Research, 5(4), 281e290. Yang, S., Huang, G. Q., Song, H., & Liang, L. (2009). Game-theoretic approach to
Dong, Y., & Ling, L. (2015). Hotel overbooking and cooperation with third-party competition dynamics in tourism supply chains. Journal of Travel Research,
websites. Sustainability, 7, 11696e11712. 47(4), 425e439.
Guo, X., Ling, L., Dong, Y., & Liang, L. (2013). Cooperation contract in tourism supply Yang, Q., & Zhang, M. (2014). Service cooperation incentive mechanism in a dual-
chains: The optimal pricing strategy of hotels for cooperative third party stra- channel supply chain under service differentiation. American Journal of Indus-
tegic websites. Annals of Tourism Research, 41, 20e41. trial & Business Management, 4, 284e294.
Guo, X., Zheng, X., Ling, L., & Yang, C. (2014). Online competition between hotels and Yao, Y. B., Ma, R. J., & Li, J. L. (2014). A retrospective review of information technology
online travel agencies: From the perspective of cash back after stay. Tourism applications by foreign travel agencies from 2004 to 2013. Tourism Science,
Management Perspectives, 12, 104e112. 28(2), 83e94.
Hendershott, T., & Zhang, J. (2006). A model of direct and intermediated sales. Yoon, M. G., Yoon, D. Y., & Yang, T. W. (2006). Impact of e-business on air travel
Journal of Economics & Management Strategy, 15(2), 279e316. markets: Distribution of airline tickets in Korea. Journal of Air Transport Man-
Huang, L. (2006). Building up a B2B e-commerce strategic alliance model under an agement, 12(5), 253e260.
uncertain environment for Taiwan's travel agencies. Tourism Management,
27(6), 1308e1320.
Yong Long received his PhD degree from Chongqing Uni-
Kimes, S. E., & Chase, R. B. (1998). The strategic levers of yield management. Journal
versity of China and now is a professor of school of eco-
of Service Research, 1(2), 156e166.
nomics and business administration, Chongqing University
Kim, D. J., Kim, W. G., & Han, J. S. (2007). A perceptual mapping of online travel
of China. His research interests include tourism management
agencies and preference attributes. Tourism Management, 28(2), 591e603.
& tourism business model innovation.
Koo, B., Mantin, B., & O'Connor, P. (2011). Online distribution of airline tickets:
Should airlines adopt a single or a multi-channel approach? Tourism Manage-
ment, 32(1), 69e74.
Law, R., Leung, K., & Wong, J. (2004). The impact of the Internet on travel agencies.
International Journal of Contemporary Hospitality Management, 16(2), 100e107.
Ling, L., Guo, X., & Yang, C. (2014). Opening the online marketplace: An examination
of hotel pricing and travel agency on-line distribution of rooms. Tourism
Management, 45, 234e243. Pingping Shi received the B.S. and M.S. degrees from the
Lu, C., & Liu, S. (2016). Cultural tourism O2O business model innovation-A case Huazhong Agriculture University and Chongqing University,
study of CTrip. Journal of Electronic Commerce in Organizations, 14(2), 16e31. China, in 2009 and 2013 respectively. Currently, she is working
Lu, J. C., Tsao, Y. C., & Charoensiriwath, C. (2011). Competition under manufacturer toward the Ph.D. degree in school of economics and business
service and retail price. Economic Modelling, 28(3), 1256e1264. administration at the Chongqing University, Chongqing, China.
Park, S. Y., & Keh, H. T. (2003). Modelling hybrid distribution channels: A game- Her research interests include tourism cooperation, tourism
theoretic analysis. Journal of Retailing and Consumer Services, 10(3), 155e167. supply chain and tourism business model innovation.
Toh, R. S., Dekay, C. F., & Raven, P. (2011). Travel planning: Searching for and booking
hotels on the internet. Cornell Hospitality Quarterly, 52(4), 388e398.
Tsay, A., & Agrawal, N. (2000). Channel dynamics under price and service compe-
tition. Manufacturing & Service Operations Management, 2(4), 372e391.