旅宿業跨文化顧客網路抱怨行為之研究：內容分析法與資料探勘演算法之應用

本研究之主旨在探討旅宿業跨文化顧客網路抱怨行為，藉由多元的研究方法收集實證資料，並以內容分析法與關聯規則演算法分析與建構旅宿業跨文化顧客網路抱怨行為之模式。研究目的有四: (1) 進行旅宿業服務品質相關文獻回顧，了解跨文化旅客與旅館類別對服務品質之影響以及研究缺口、 (2) 調查旅客網路負面口碑，分析顧客抱怨類型與造成顧客不滿意的屬性、 (3) 分析預測引發不同類別旅館之顧客網路抱怨的因素、 (4) 建立不同類別旅館之顧客網路抱怨類別間之關聯規則。四個連續性之研究包含: (研究一) 旅宿業服務品質相關文獻回顧: 透過110篇選自於18 本SSCI期刊，年度: 2004 ~ 2018之旅宿業服務品質相關文獻探討，運用系統化文獻分析，了解跨文化旅客與旅館類別對服務品質之影響以及研究缺口； (研究二) 顧客抱怨類型與造成顧客不滿意原因的屬性分析: 透過網路平台TripAdvisor節錄顧客負面口碑，共計採用2,020篇網路評價，來自於353 間旅館之顧客，其國籍遍及5大洲 63國。研究結果顯示顧客文化背景影響顧客抱怨類型，同時住宿不同類別旅館的顧客抱怨類型也有差異； (研究三) 預測引發不同類別旅館之顧客網路抱怨的因素分析: 透過質性內容分析法分析顧客網路評價，確認11種網路抱怨類型與 65個抱怨項目，其中「網路評價online review」相較於相關文獻，為一種新的抱怨類型。而不同類別旅館之顧客網路抱怨中，「服務經驗service experience」則是跨文化旅客中均重視的因素。 (研究四) 不同類別旅館之顧客網路抱怨類別間之關聯規則建立: 透過決策樹分析方法，本研究提出四種常見旅宿業服務失誤類型: 「服務接觸service encounter」、「清潔度 cleanliness」、「空間Space」、「安全Security」。同時佐以資料探勘演算法分析，本研究提出前10大顧客網路抱怨類別間之關聯規則。本研究以多元分析方法分析實證資料並提出顧客抱怨模式，可供業界參考，加強旅宿業服務品質，且導入決策樹與關聯規則演算分析方法，除建構旅宿業跨文化顧客網路抱怨行為模式外，更可建立大數據分析模式，強化消費者行為模式之研究。關鍵字：網路口碑、網路抱怨行為、文化背景、旅館類別、服務品質、內容分析、決策樹分析、關聯規則演算法

關鍵字

網路口碑；網路抱怨行為；文化背景；旅館類別；服務品質；內容分析；決策樹分析；關聯規則演算法

並列摘要

Introduction: Literature has demonstrated that online reviews from tourism-specific websites and other electronic channels (e.g., TripAdvisor) do influence tourists’ decisions in all aspects. At this global era, online complaints have become increasingly influential on the purchasing behaviour of customers in recent years. In an effort analyze large quantities of textual complaints and detail the various aspects of them, Analysis of Textual Data was looked to as an ideal framework to take on the task. Purpose: The main purpose of this study is i) to enrich literature on service quality in the hospitality and tourism industry by examining variation in online complaining behaviours across cultures of origin and classes of hotel; ii) to investigate which online complaint attribute categories are influencing customers' overall dissatisfaction; iii) to predict which hotel items/attributes will receive significantly different attention in the online complaining behavior of guests staying at different classes of hotels; and iv) to identify associations or relationships among the different attributes receiving attention in the online complaining behavior of guests staying at different classes of hotels. Design/methodology/approach: The current research applies both qualitative and quantitative analysis approaches. For instance, Systematic Review Approach, T-Test, ANOVA, Frequency Analysis, Content Analysis, Decision Tree Algorithm, and Association Rules Algorithm. The analyses of this mix of qualitative and quantitative methods reinforce the process of data triangulation. The population or sample: The targeted objects of review by the subjects in this study were TripAdvisor-listed British hotels. A total of 404 hotels were randomly selected from a population of 1086 hotels located in the United Kingdom (TripAdvisor, 2018). The hotels ranged from 2 to 5 stars according to the British hotel rating system. Consequently, 51 hotels were omitted from the study on account of the fact that the related information did not meet the necessary criteria. A manual approach was applied to conduct the initial data collection, meaning that the researchers went through every single review to code it subjectively. In total, 353 hotels with 2,020 usable individual complaining reviews were collected for the analysis. The total sample of 2,020 complaining online reviews represents 5 continents and 63 nationalities. The reviews were classified as belonging to either Asian or non-Asian travellers based the regional markings (UN, 2019). Of the samples, 51% (N = 32) were provided by non-Asian travellers, and Asian travellers provided 49% (N = 31) of the samples. The topics or theories: Two types of theories were employed in this study, there are: Theory of Expectancy Disconfirmation (Oliver, 1980), and Hofstede's Cultural Dimensions Theory (1980). Findings: By conducting a systematic review on eWOM literature, the results reveal that research trends on eWOM has dramatically increased since 2004, with the most dramatic growth (79%) occurring between 2014 and 2017. Results indicate that online complaining behavior is influenced by cultural background. Asian and non-Asian travelers appear to place a similar emphasis on Value for money. However, Asian guests are more likely to complain about Service, while non-Asian guests are more likely to complaint about Cleanliness, Room, Sleep Quality, and Location. Additionally, online complaining behavior varies between different classes of hotel. Guests of high-class hotels are more likely to complain about Service and Value, while guests of low-class hotels are more likely to complain about Cleanliness, Room, Sleep Quality, and Location. The results from the word frequency analysis reveal that both Asian and non-Asian travelers tend to put more emphasis on Booking and Reviews when posting complaints online. Based on a manual qualitative content analysis, 11 different major online complaint categories and 65 sub-categories were identified. Among its important findings, results of this study show that non-Asian guests frequently make complaints which are longer and more detailed than Asian customers. The sensitivity analysis indicated that the most important online complaining attributes are Service Encounter, Space, Security, and Cleanliness, respectively. Based on the CHAID results, the findings also revealed that about 70% of the complaints made about Service Encounter, Space and Security came from luxury hotel guests, while approximately 71% of the online complaints about Cleanliness were from budget hotel guests. The Apriori algorithm was employed to identify the top 10 rules associated with complaining behaviors of hotel guests at luxury and budget hotels. Significance of the study: This study offers the first attempt to analyze and compare online complaining behavior trends seen among guests of different cultural backgrounds and guests of different classes of hotel. Practical implications: Decision trees are used build rules with only a single conclusion, whereas association rules attempt to find many rules, each of which may lead to a different conclusion. By the new approach in the data analysis, more profound findings can be utilized to reinforce the strengths of hotel operation to meet the expectations and needs of target guests. Originality/value: One of the main contributions of this study lies in the fact that it is a novel attempt to predict the online complaining behaviour of guests from different classes of hotels by utilizing Data Mining algorithms. Delimitations: First, this research only concentrated on hotels in the UK. For future studies on cultural differences, it would be worth expanding the analysis to other international tourism destinations, such as Hong Kong, Singapore, Paris, or New York. Second, this study analysed online complaining reviews, which are written in English by a very small segment of the tourist population. Therefore, it might be bias for guests who are not be able to communicate in English. Finally, this study's model accuracy (i.e. Decision Tree Models) is in the range of acceptable due to the data limitation. In this study, dataset was built from the manual content analysis which is restrained on time-consuming and labour intensive. Further research is recommended to increase volumes of data and performed the comparisons. Keywords: eWOM studies, online complaining behaviour, cultural background, hotel class, service quality, Content Analysis, Decision Tree, Association Rules

並列關鍵字

eWOM studies ； online complaining behaviour ； cultural background ； hotel class ； service quality ； Content Analysis ； Decision Tree ； Association Rules

參考文獻

Abubakar, A. M., & Ilkan, M. (2016). Impact of online WOM on destination trust and intention to travel: A medical tourism perspective. Journal of Destination Marketing & Management, 5(3), 192-201.

Google Scholar

Adelaide, G. M.-K., Ishmael, M., & Edem Kwesi, A. (2015). Service Failures in 3- to 5-Star Hotels in Accra, Ghana. J. of Tourism and Hospitality Management, 3(2). doi:10.17265/2328-2169/2015.04.003

Google Scholar

Ady, M., & Quadri-Felitti, D. (2015). Consumer research identifies how to present travel review content for more bookings. Available at: http://www.trustyou. com/press/study-reveals-travelers-prefer-summarized-review-content-full-text-reviews, Accessed 11-Apr-15.

Google Scholar

Agrawal, R., & Srikant, R. (1994). Fast Algorithms for Mining Association Rules. IBM Research Report RJ9839, June 1994, IBM Almaden Research Center, San Jose, Calif..

Google Scholar

Ahmad, W., & Sun, J. (2018). Modelling consumer distrust of online hotel reviews. International Journal of Hospitality Management, 71, 77-90.

Google Scholar

主題瀏覽