Business Analytics in Tourism: Uncovering Knowledge from Crowds.

AutorMarcolin, Carla

Introduction

The decision-making process, in different managerial environments, faces many challenges as increasingly larger pools of data are produced every day. Companies are progressively pressed to access these data using analytics tools to support their decisions (Ransbotham & Kiron, 2017). Indeed, the concept and practice of Business Analytics had significant growth in the last decade, attracting the attention of researchers and managers from different areas (Mortenson, Doherty, & Robinson, 2015).

Business Analytics allows leveraging value from data, thus being an important tool for the decision-making process (Acito & Khatri, 2014). Business Analytics helps the analysis of large amounts of data and integrates different data sources, making it possible to improve a company's performance and identify business opportunities (Bayrak, 2015). However, in recent years, the presence of data in different formats poses extra challenges. Companies, in addition to dealing with large volumes of data, now need to handle data types such as voice, text, log files, images and videos (Davenport & Dyche, 2013).

In such a diverse context, textual data has been drawing organizational attention, as millions of people express themselves daily using text in many applications and tools available. If appropriately processed, textual data represent a perception sensor about customer experiences that is not only useful but vital for business analysis and decisions (Zhao, 2013). In this sense, the tourism and hospitality industry has been interested in customer perceptions for improving operations in the services industry (Han, Mankad, Gavirneni, & Verma, 2016). Nowadays, travel platforms like TripAdvisor collect volunteered information, openly distributing user reviews to firms and managers, challenging them with a great volume of data (Yoo, Sigala, & Gretzel, 2016) and building reputational economies for the tourism and hospitality industry (Langley & Leyshon, 2017).

Since platforms that facilitate experience sharing have become more and more popular, customers are willing to rely on electronic word-of-mouth (eWOM) as an important step before making a destination decision (Sparks & Browning, 2011). As most data are available in text format, finding effective ways to analyze and transform them into valuable information is one of the challenges that connects this industry to Business Analytics (Tang & Guo, 2015). As eWOM provides genuine information about customers, their opinions about tourism and hospitality services, expressed in natural language, form an important source of information for hotel managers (Carrasco & Villar, 2012).

However, despite the relevance of customer data in textual format to support the decision making of hotel managers, its use is not as frequent as it would be expected, due to the difficulty of analyzing and interpreting large amounts of data in that format, making it hard to acquire useful information for hotel strategy (He et al., 2017). In this way, capturing an accurate and complete picture of the customer experience is a most challenging task for hotel managers in recent years (Han et al., 2016).

In addition, there is evidence of concern about how to develop strategies to respond to such issues, given that, for most companies in the tourism and hospitality industry, there has been a change in managerial logic, pushed by the market and with intense use of eWOM platforms (Del Vecchio, Mele, Ndou, & Secundo, 2017). Given that online content is being produced faster than the capacity to analyze it (Ferreira, 2019) and developing strategies to adequately respond to customer's needs is an urgent need in this industry (Del Vecchio et al., 2017), the present article aims to answer the following research question: how can hotel managers analyze a large volume of data on guest reviews, so as to gain information and develop strategic actions in line with market trends? More specifically, our main objective is to identify the key evaluation topics presented in online guest reviews, revealing growing or falling trends through the years. This contributes to practices of hotel managers by developing and demonstrating the applicability of text mining tools, based on open-source solutions, and providing insights from the data and assisting in their strategic decision-making process. In addition, this article contributes with theory by demonstrating how to combine unsupervised learning and longitudinal analysis to make market trends evident, using publicly available customer textual data.

Although the evidence supporting the importance of reviews for hotel managers has already been explored, our approach is different from previous research in three aspects. First, rather than conducting experiments or focus group sessions (Horner & Swarbrooke, 2016; Sparks & Browning, 2011), this study analyzes real-world data extracted directly from websites that provide open access to information, like TripAdvisor. The data come from texts written directly by customers, representing the Voice of the Customer (VOC) itself and making it possible to understand what the customers are sharing about the organizations (Spangler & Kreulen, 2007). Second, the use of unsupervised learning tools such as Latent Semantic Analysis (LSA) allows for an objective analysis (Ashton, Evangelopoulos, & Prybutok, 2014), since the emergent categories are not provided by the analyst or taken from platforms (Xu, 2018), neither taken from any keywords framework or pre-existing ontologies (Thomaz, Bizb, Bettonic, Mendes-Filho, & Buhalise, 2017), rather emerging from the text, given the latent semantic relation between reviews and words. Finally, in order to provide strategically useful information, we analyzed topic trends through the years, allowing for the identification of growing or falling aspects of interest in the customer's view, instead of delivering a photograph of an instant in time, as usually done in other studies (Xu, 2018; Xu, Wang, Li, & Haghighi, 2017).

The article is organized as follows: in second section, we discuss some aspects related to the customer's review presented in the literature, as well as descriptions of the model applied; in third section, we present the methodological procedures adopted in the study; in fourth section, we discuss the results; and, in fifth section, we present final remarks and conclusions.

Conceptual Background

As this work identifies the main topics of online guest reviews by revealing the evolution throughout the years, we explored textual data in a Business Analytics framework in the tourism and hospitality industry. Our conceptual background includes several papers that have explored guest reviews regarding the decision-making process of choosing hotels or destinations, thus demonstrating the power of eWOM and review ratings in social media. Additionally, we analyze previous works on LSA, which was chosen to demonstrate the value of text analysis in decision support.

Guest reviews

Accommodation and hotel services have high impact in tourism development (Vieira, Hoffmann, & Alberton, 2018). From the customer perspective, the importance of previous reviews for their decision-making process regarding hospitality has been extensively demonstrated in the literature (Sparks & Browning, 2011; Ye, Law, & Gu, 2009). Even without knowing the other users behind the screen, one important step in planning a travel, and thus deciding a place to stay, is to access a review from well-known websites and take that information in consideration. Social media and customer review websites, like TripAdvisor, have changed the tourism and hospitality industry and the practices of hotel managers (Molinillo, Sandoval, Morales, & Stefaniak, 2016). In the tourism literature, studies about eWOM have developed quickly over the last years with the increased popularity of customers' online booking and online review behavior (Xu, 2018).

Another important aspect is the strong predictive power of the so-called social media review rating and hotel performance metrics. Kim and Park (2017) compared traditional customer satisfaction of a hotel with the same data from four different websites. They discovered that not only social media ratings were better predictors for metrics like average daily rate and percentage of occupancy, but also that data from TripAdvisor had the closest correlation. Thus, social media rating is a significant predictor to explain hotel performance metrics like percentage of occupancy and room revenue (Kim & Park, 2017). Besides that, eWOM is associated with customer retention and loyalty, as online reputation comparison is facilitated through travel platforms (Cantallops & Salvi, 2014).

Previous experience from other customers has high importance before booking a hotel room online. Positive online reviews can significantly increase hotels booking rates. Besides, the polarity of reviews has a negative impact on reservations. Indeed, the tourism and hospitality industry should strongly consider online reviews, especially those posted in external portals apart from the organization's website (Ye et al., 2009). The review itself also tends to have more importance for customer perception, conveying more impact than ratings alone (Sparks & Browning, 2011).

Yen and Tang (2015) analyzed the motivations for posting hotel experiences with the online media chosen and identified whose eWOM motivations are affected by hotel attribute performance. The choice between TripAdvisor and Facebook, for example, is correlated with different motivations. TripAdvisor is associated with altruism and platform assistance, while Facebook is positively associated with extraversion, social benefits, and dissonance reduction. The findings suggest that motivations are not universally equal and eWOM behaviors are correlated with different motivations.

In this sense, recent advances in computer science, especially in Natural Language...

Para continuar a ler

PEÇA SUA AVALIAÇÃO

VLEX uses login cookies to provide you with a better browsing experience. If you click on 'Accept' or continue browsing this site we consider that you accept our cookie policy. ACCEPT