Đại lý du lịch cần làm gì để giảm thiểu sự không hài lòng của khách hàng?

Nội dung text: Đại lý du lịch cần làm gì để giảm thiểu sự không hài lòng của khách hàng?

1. HOW MUCH EFFORT SHOULD A TRAVEL AGENT MAKE TO REDUCE TRAVELING DISPUTES? ĐẠI LÝ DU LỊCH CẦN LÀM GÌ ĐỂ GIẢM THIỂU SỰ KHÔNG HÀI LÒNG CỦA KHÁCH HÀNG? Chien-wei Wu - National Chi Nan University Chih-chi Ni - Lunghwa University of Science and Technology Yih-ming Lin＊ - National Chiayi University Abstract The purpose of this paper is to investigate how much effort a travel agent should make to avoid customer’s dissatisfaction. When a customer is looking for a travel package, the customer needs the travel agent providing the information related to the travel package. On the other hand, the travel agent does not know the customer preference. Therefore, the problem between a customer and a travel agent is a matching problem, in which the travel agent and the customer are both with incomplete information. In this study, we employed cognition model to analyze this problem. We characterize the Nash equilibrium. We find that the travel agent’s effort and the customer effort are strategic substitutes. The results indicate that the customer and the travel agent both choose less effort than social optimal level. Keywords: matching, Nash equilibrium, cognition Tóm tắt Nghiên cứu nhằm mục đích điều tra các nỗ lực mà đại lý dịch cần thực hiện nhằm giảm thiểu sự không hài lòng của khách hàng. Khi một khách hàng sử dụng dịch vụ du lịch trọn gói, họ cần đại lý du lịch cung cấp cho mình các thông tin cần thiết về chuyến đi. Tuy nhiên, đại lý du lịch lại không hiểu rõ về thị hiếu của khách hàng. Do đó vấn đề giữa khách hàng và đại lý du lịch là vấn đề về việc hợp nhất, trong đó đại lý du lịch và khách hàng đều không có thông tin đầy đủ về đối tác. Trong nghiên cứu này, chúng tôi sử dụng mô hình tri nhận để phân tích vấn đề. Chúng tôi sử dụng phương trình cân bằng Nash. Kết quả nghiên cứu cho thấy các đối lực của đại lý du lịch và khách hàng có tính thay thế. Kết quả này chỉ ra rằng cả khách hàng lẫn đại lý du lịch đều chưa nỗ lực đến mức tối ưu. Từ khoá: hợp nhất, cân bằng Nash, tri nhận 1. Introduction Travel package is a typical example of incomplete contract since contingencies always happen. The traveling disputes or customer’s dissatisfaction always occurs when contingencies or accident events happen. For example, an airliner on the scheduled has been cancelled due to bad weather condition or the scheduled hotel is not available due to ＊Corresponding author, associate professor, Department of Applied Economics, National Chiayi University, Chiayi, 600, Taiwan. E-mail: yxL173@mail.ncyu.edu.tw; Tel: 886-5-2732860. Fax: 886-5-2732853. Residual errors are ours alone. 135
2. special local festival. Those contingencies will make customers unhappy and dissatisfied unless the travel agent makes some effect to recover it. Otherwise, it is easy that the travelling disputes could happen. The purpose of this paper is to investigate how much effort a travel agent should make to avoid customer’s dissatisfaction. When a customer is looking for a travel package, the customer needs the travel agent providing the information related to the travel package. 2. Model Consider an economy with a customer and a travel agent. The customer is looking for a travel package and needs the travel agent providing the information related to the travel package. Suppose the travel agent has or design a travel plan A for the customer whilst the travel plan costs C0 with a price P. At first, they can help themselves of common-knowledge for the travel plan. Due to the property of travel plan, the travel agent will insist on a contract before the travel package completed. The contingencies of this travel plan occur with probability ρ. The contingencies or accident events could be an airliner on the scheduled has been cancelled due to bad weather condition or the scheduled hotel is not available due to special local festival. The travel plan is completed without any contingency with the probability 1-ρ and the customer receives utility G. Since there is probability with contingencies, the seller will prepare a backup plan, A’, when the contingency or accident event actually happens. In this paper, we assume that all travel disputes happen when the contingencies occur since the original travel plan is usually a standard travel plan without disputes. In practice, most of the travelling disputes occur when the travel contingencies happen. In order to reduce travelling disputes, the travel agent will communicate with the customer for the backup travel plan. For example, when the original travel plan cannot be completed due to the weather condition, such as heavy raining, the travel agent changes the travel plan to go to the museum instead of elphone show. However, some customers do not like the backup arrangement and dispute occurs. Before the travel plan is delivered, the travel agent will make effort to reduce the travel disputes or dissatisfaction. Suppose that the travel agent make his effort s with communication cost Cs(s). It is assumed that Cs(.) is increasing in s and convex in s, (.)>0 and (.)>0. Similarly, the communication cost the buyer Cb(b) where the b the effort the customer make to communicate with the travel agent. It is assumed that Cs(.) is increasing in b and convex in b, (.)>0 and (.)>0. Furthermore, we assume that Cb(0)=0 and Cb(1)>(1-ρ)G. When the contingencies happen and the backup plan is delivered, two results could happen. One is that the customer is satisfied with this adjustment for the contingency. The adjustment costs the travel agent C1. We assume that the buyer is satisfied with the probability F(s,b). It is assumed that the probability is positively related to travel agent effort (s) and the buyer effort (b). For simplicity, we In other word, Fs(.) and Fb(.) are both positive. On the other hand, the customer could be dissatisfied and travelling dispute occur with the probability, 1-F(s,b). When the dissatisfaction or travelling dispute occurs, the customer is very unhappy which means he receive the utility 0 instead of G. Since the 136
3. dispute happens, it raise that there is a negative impact on seller’s reputation, L. Furthermore, we assume that Cs(0)=0 and Cs(1)>(1-ρ)L. The payoff function for the buyer can be specified as Πb(b,s)=[ρ+(1-ρ)F(b,s)]G-P- Cb(b), (1) Equation (1) indicates that the buyer will get utility G if there is no contingency (with probability ρ) or there is contingency during the travelling and the buyer is satisfied the adjustment. Of course, the buyer has to pay the price P for the travel package and communication cost Cb(b). Furthermore the payoff function for the seller can be written as Πs(b,s)=P-(1-ρ)[1-F(b,s)]L-C0-(1-ρ)C1, (2) For simplicity, we let F(b,s)=1-(1-b)(1-s) where 0≤b≤1 and 0≤s≤1. 3. The Non-cooperative Equilibrium 3.1 Nash equilibrium In the Nash equilibrium, the buyer chooses how much effort he makes to communicate with the travel agent according the seller’s effort. Taking the derivative of equation (1) with respect to b, we can obtain the following first order conditions: (1-ρ)(1-s)G-2bβ=0. (3) The best response function of the buyer can be derived as Rb(s)=b= (4) Similarly, the travel agent chooses how much effort he makes to communicate with the customer. Taking the derivative of equation (2) with respect to s, we can obtain the following first order conditions: (1-ρ)(1-s)L-2sα=0. (5) The best response function of the seller can be derived as Rs(b)=s= . (6) According to the best response function of the buyer, the best response of the buyer is 0 if s . By the assumption of Cb(1)>(1-ρ)G, we can get that the slope of Rb(s) is negative and smaller than -1/2. Moreover, we also can find that the slope of the reaction function of seller, Rs(b) is also negative but bigger than -1/2. We can conclude that the seller’s reaction function is steeper than the buyer’s reaction function. The best response functions of the buyer and the seller are depicted in Figure 1. As shown in Figure 1, the 137
4. best response function of the buyer is whilst the best response function of the seller is . In figure 1, we can find that there are three intersection points, (1,0), (0,1) and E, between the two best response functions, and . Since Cb(1)>(1-ρ)G, the best strategy of the buyer cannot be 1. Therefore, we can rule out the point (1,0) as a Nash equilibrium. It is similarly that the best strategy of the seller cannot be 1. The point (0,1) can be eliminated. Therefore, we can find that there is only a unique Nash equilibrium E in Figure 1. [insert Figure 1 about here] Furthermore, according to equations (4) and (6), we can solve the Nash equilibrium as (b*, s*)= . (7) Since Cb(1)>(1-ρ)G, we can get . Similarly, we also can get . It is shown in Figure 1 that the buyer strategy b* is less than 1/2 and the seller strategy s* is also smaller than 1/2 in the Nash equilibrium. 3.2 Comparative static In this subsection, we will investigate how the Nash equilibrium changes if other exogeneous variables change which is so-called comparative static analysis. We can examine the relation using Figure 1. As shown in equation (7), we can obtain the Nash equilibrium is affected byρ, G, L, αandβ. As shown in Figure 1, if the reputation loss increases, we can find that becomes flatter whilst remains unchanged. We can find that the optimal strategy (b*,s*) shift to the right, which means the seller will increase his effort but the buyer will decrease his effort. 3.3 Social Optimum To investigate whether the efforts of buyer and seller are efficient from the view of whole economy, we consider a factitious social planner’s problem. In this setup, the social planner is aiming for maximize the total social welfare which is defined as sum of the seller and the buyer payoff function in this model. Thus, the social welfare function can be obtained by combining Equations (1) and (2), as follows: W(b,s)= ρG+(1-ρ)F(b,s)u-(1-ρ)[1-F(b,s)]L-C0- Cb(b)- Cs(s), (8) In order to maximize social welfare, the central planner can choose the two variables, b and s to maximize equation (8). Taking the derivative with respect to b, we can obtain the following first order condition: 138
5. =(1-ρ)(1-s)(u+L)-2βb=0. (9) By rearranging equation (9), we can obtain b= . (10) If we compare equation (10) with equation (4), we can find that equation (4) and equation (10) have an intersection point at (s,b)=(1, 0). It is worthwhile to mention that equation (10) is always on the top of equation (4). The intuition behind this phenomenon is easy: the buyer’s effort of cognation possesses positive externality on seller’s welfare, but buyer won’t take it into account when she chooses the effort level of cognition. So buyer’s Nash equilibrium effort level is smaller comparing to the level of social optimum. Similarly, we will see the same phenomenon in seller’s side. Taking the derivative with respect to s, we obtain the other first order condition: =(1-ρ)(1-b)(u+L)-2αs=0. (11) By rearranging equation (11), we can obtain s= . (12) Similar to equation (10), we can find that equation (12) and equation (6) have an intersection point at (0,1). Since the seller’s effort of cognation possesses positive externality on buyer’s welfare, but seller won’t take it into account when she chooses the effort level of cognition. Therefore, equation (12) is also on the top of equation (6). Equation (10) and (12) can be illustrated in Figure 3 and the social optimal solution can be depicted as X in Figure 3. [insert Figure 3 about here] Furthermore, the social optimal buyer and seller effort level can be solved by equation (10) and (12) simultaneously. X=(ss,bs)= . (13) Using the social efficient outcome X as the benchmark point, We can compare the Nash equilibrium outcome with it in Figure 3. As shown in Figure 3, there are three cases. Case I is that the seller’s effort and buyer’s effort in Nash equilibrium are both smaller than social efficient optimal level (i.e. sn is smaller than ss and bn is also smaller than bs). As expected, social planner tends to have both players to choose higher effert level of cognition because their choices possess positive externality on each other. Therefore, as we 139
6. pointed out in proposition I, buyer and seller’s choices are strategic substitutes. The phenomenon that both buyer and seller provide too less efforts on cognition is called underprovident. However, there are also two cases appear with one-side overprovidence problems, Case II and Case III which associated with two rather extreme circumstances. First, according to Case II of figure 3, we can find that if the difference of two endpoints of equation (4) and (10), , is small enough, then it can be shown that seller’s Nash equilibrium effort level of cognition sn is smaller than seller’s social optimal effort level ss but buyer’s Nash equilibrium effort level bn is larger than buyer’s social optimal effert level bs. The intuition behind this phenomenon is in the following: We first distinguish the situation into two cases: L is sufficiently small or is sufficiently large. In the cases that L is sufficiently small, firm might choose a very small ss. Since s and b are strategic substitute, a smaller s induce the buyer to choose higher effort in Nash equilibrium. In some extreme cases, is greater than the effort level in social optimal . In the cases that is sufficiently large, since the buyer’s cognition cost is much higher, it will be more efficient that seller(buyer) chooses higher (lower) level of cognition from the views of society. Second, according to Case III of figure 3, we can find that if the difference of two endpoints of equation (6) and (12), , is small enough, then it can be shown that buyer’s Nash equilibrium effort level of cognition bn is smaller than seller’s social optimal effort level bs but seller’s Nash equilibrium effert level sn is larger than buyer’s social optimal effert level ss. In the cases that is sufficiently small, buyer might choose a very small bn and then induce the buyer to choose higher effort sn in Nash equilibrium. In some extreme cases, it might be higher than the effort level ss in social optimal. In the cases that α is sufficiently large, since the seller’s cognition cost is much higher, it will be more efficient that buyer(seller) choose higher (lower) level of cognition from the views of society. Therefore, we have the following proposition. Proposition 3: The seller’s effort (sn) in Nash equilibrium is smaller than the seller’s effort (ss) in social optimum, or the buyer’s effort (bn) in Nash equilibrium is less than the buyer’s effort (bs) in social optimum, or both. 140
7. In this paper, we assume that traveling disputes occur because the customers are dissatisfied with the backup plan after the travel contingencies happen. To avoid the disputes, it depend on how much efforts have been put in by both seller and buyer. In other words, given the effort levels of seller and buyer, the probability of traveling disputes happen is which we define it to be . In the case that and (case I of Figure 2), this probability of disputes in Nash equilibrium is clearly higher than (i.e., ) )). In the cases that or (case II and case III of Figure 2), this cases associate respectively with the conditions that and is sufficiently small. It is not clear whether ) is greater or smaller than ). Though we can’t characterize this problem completely, we answer it partially by providing a sufficient condition for ) in Proposition 4. For ease to state and prove Proposition 4, we first present two facts, Lemma 1 and Lemma 2, and prove them in the appendix. Lemma 1: If sn (1-ρ)L and Cb(1) > (1-ρ)u, then (1-ss)(1-bs) (1-ρ)L , we have s < . Similarly, bn < . === Using Lemma 1, we can get < and < . Furthermore, by Lemma 2, = < = 1. Hence, we conclude that (1- ss)(1-bs)<(1-sn)(1-bn). Therefore, we have ). 141
8. The meanings of the inequality ) is that seller and buyer provide less effort in decentralize circumstance (Nash equilibrium) than the social optimal level. This inequality holds obviously in the case I of Figure 2. Though this inequality is not necessary hold in two rather extreme cases that A or B are small enough (case2 and case3 of Figure 2), the sufficient condition for the inequality to be true is that A and B are smaller than X. This sufficient condition coincides partially with the precondition of Case 2 and 3 of Figure 2. With the results above, we can conclude that: In Nash equilibrium, it often happens that seller and buyer provide less effort of cognition to avoid travel disputes (compare to social optimal effort level). 4. Concluding Remarking In this study, we have developed a simple theoretical model of cognition game between consumers and travel agency, in which the consumer complaining behavior is incorporated, to capture the relation between the travel agent and the customer. Since tourism products have several characteristics, such as intangibility, simultaneous production and consumption, heterogeneity, and perishability, it makes that the contents of travel service are difficult to standardize and the quality of the travel service is unpredictable. Moreover, the contingencies always happen in travel service or tourism industry. Travel service is a typical example of incomplete contract since contingencies always happen. We follow and modify Tirole (2009) to develop a cognition model between the seller and the customer in travel business. The existence of Nash equilibrium in this cognition game is confirmed and characterized. It is shown that the travel agency effort and the customers’ effort are strategic substitute, which means the best response of seller’s effort is negatively related to the buyer’s effort, and vice versa. Secondly, the comparative static analysis for the Nash equilibrium is investigated. It is shown that the travel agency will increase his effort and the customer will decrease his effort in cognition if the possible travel agent’s reputation loss increases. If the utility gained from the travel (u) increases, the travel agent will decrease his effort whilst the customer will increase his effort. If the probability of contingency increases, the efforts of the seller and the buyer both decrease. In addition, we also find that we find the sufficient condition that the probability of occurrence of travel disputes in Nash equilibrium is bigger than in social optimal level. REFERENCES Alegre, J. and Garau, J. (2009), “Tourist satisfaction and dissatisfaction,” Annals of Tourism Research, 37, 52–73. Bearden, W. O., & Teel, J. E. (1983), “Selected determinants of consumer satisfaction and complaint reports,” Journal of Marketing Research, 20, 21-28. Burk, James F., and Resnick, B. (2000), Marketing and Selling the Travel Product, 2nd edition, Delmar Thomas Learning Publishers Inc. Albany, N.Y 142
9. Day, R. L. (1980). “Research perspectives on consumer complaining behavior,” In Theoretical Developments in Marketing, Carles Bamb and Patrick Dunne, Chicago: American Marketing Association, 211-215. Day, R. L. and Landon, E. L. (1977), “Collecting Comprehensive Consumer Complaint Data by Survey Research,” Advances in Consumer Research, 3, 263-269. Del Bosque, I., and San Martin, H. (2008), “Tourist satisfaction: A cognitive-affective model,” Annals of Tourist Research, 35, 551-573. Ismail, Joseph A. and Mills, Juline E. (2001), “Contract Disputes in Travel and Tourism: When the Online Deal Goes Bad,” Journal of Travel & Tourism Marketing, 11, 63- 82. Li, Tzu-Hui (2004), The Case Study of Travel Disputes, Yang-Chih Book Co. Ltd. (in chinese) Singh, J.(1988), “Consumer Complaint Intentions and Behavior: Definitional and Taxonomical Issues,” Journal of Marketing, Vol,52, pp.93-107. Tirole, J. (2009), “Cognition and incomplete contracts,” American Economic Review, 99, 265-294. Appendix: Lemma 1: If sn < then < and if bn < , then < . Proof: By equation (13), we can solve ss= = , where A denotes and B denotes . Furthermore, we can rewrite ss = = = = . 143
10. In addition, based on equation (13), we can get sn = = = . Therefore, if sn < = = , then ss < = sn or < . Similarly, it is easy to show that if < , then < , which completes the proof of Lemma 1. Lemma 2: = Proof: Since ( ) satisfies that = and = , we can derive = . Therefore, we can get = = . (A1) Moreover, since ( ) satisfies that = and = we can get = . (A2) According to equations (A1) and (A2), we get = (A3) Rearranging (A3), we can get = , which completes the proof of Lemma 2. 144
11. Table 1. Classification of Travel Disputes (1990-2004) Classification Number of Case percentage persons Cancellation before departure 1,085 17.50% 4,059 Schedule problem 787 12.69% 9,396 flight problem 761 12.27% 3,358 others 746 12.03% 2697 inadequate service 615 9.92% 3753 travel document problem 534 8.61% 1,319 Hotel changed 364 5.87% 2,766 schedule changed 190 3.06% 1,405 down payment 168 2.71% 959 shopping 127 2.05% 371 force majeure incidents 112 1.81% 691 accident 100 1.61% 345 overpaid 95 1.53% 399 Items cancelled 93 1.50% 640 Luggage lose 92 1.48% 185 compensatory 89 1.44% 8.283 flight delayed 67 1.08% 783 medical care 65 1.05% 188 Tips 42 0.68% 75 refuse to pay remainder payment 35 0.56% 600 Stay oversea 24 0.39% 124 Schedule changed due to VISA loss 10 0.16% 49 Total 6,201 100% 42,445 Data resource: Li (2004) Take No Action Disputes or Dissatisfaction Seek Redress Occurs Directly From Business Take Some Public Action Legal Action Action Complain to Public or Private Agencies Boycott Seller or Private Manufacturer Action Warn Friends and Relatives Figure1 Day and Landon’s (1977) Classification of Consumer Complaint Behavior 145
12. b 251661312 1 B (1-ρ)u E 2β s (1-ρ)L B 1S 2α Figure 2 The Nash equilibrium b 251661312 1 (1-ρ)(u+ L) 2β X (1-ρ)u 2β E s (1-ρ)L (1-ρ)(u+ L) 1 2α 2α Case I 146
13. 251661312 b X E s Case II b 251661312 X E s Case III Figure 3 The social optimum and the Nash equilibrium 147