Business Intelligence in the Hotel Industry

(1)

Master Degree

Business Intelligence in the Hotel Industry

Author: Rei Shahini

Supervisor: Behrooz Golshan

Examiner: Associate Professor Päivi Jokela Date: 2020-10-01

Course Code: 4IK50E, 15 credits Subject: Information Systems Level: Graduate

Department of Informatics

(2)

2

Abstract

Applications of artificial intelligence (AI) in hospitality and accommodation have taken an enormous percentage of service-provision, helping automate most of the processes involved such as booking and purchasing, improving the guest experience, tracking of guest preferences and interests, etc. The aim of the study is to understand the roles, benefits and issues with the improvement of business intelligence (BI) in hospitality. This research is purposed to discover the applications of BI in hotel booking and accommodation. The investigation focuses on hotel guest experience, business operations and guest satisfaction. The research also shows how acquiring proper BI is supported by implementing a dynamic technology framework integrated with AI and a big data resource. In such a system, the intensive collection of customer data combined with an improved technology standard is achievable using AI. The research employs a qualitative approach for data discovery and collection. A thematic analysis helps generate proper findings that indicate an improvement in the entire hospitality service delivery system as well as customer satisfaction. In this thesis, there are examined various subsets of BI in tourism. The assessment analyzes competition arising from the application of these technologies. The study also shows the importance and application of harnessing data to gather insights about guest interests and preferences through the establishment of well-developed BI. Insights enable the customization of hotel services and products for individual guests. There is a considerable improvement in guest services and guest information collection, which is achieved through the creation of guest profiles. The research performs a discussion on the incorporation of AI and big data among other sub-components in creating diversified BI and seeks to identify the need for current BI applications in the hotel industry.

Keywords

Business Intelligence, Hotel Industry, Big Data, Artificial Intelligence, Hospitality, Tourism

(3)

3

Acknowledgements

I would like to extend my deepest appreciation to my thesis supervisor Behrooz Golshan for his guidance, feedback and support throughout the process of writing this master thesis.

I want to show my gratitude to the module leader, Professor Anita Mirijamdotter and the examiner, Associate Professor Päivi Jokela, for their valuable feedback and advices throughout the seminars.

Furthermore, I wish to thank all the participants of my study for their time and commitment. This master thesis would not have been possible without their participation. I am thankful for their help and valuable contribution to this research study.

Finally, I must express my very profound gratitude to my family for providing me with unfailing support and continuous encouragement through the process of researching and writing this thesis. This accomplishment would not have been possible without them. Thank you.

(4)

4

List of Figures

Figure 1. Thesis Organization..………...………...11

Figure 2. The hierarchy relation of AI with machine learning and deep learning……….…15 Figure 3. Thematic Relationship Diagram………...………..29

List of Tables

Table 1. Interviewed Participants………..………24

List of Abbreviations

AI Artificial Intelligence BI Business Intelligence BIS Business Intelligence System DSS Decision Support System IT Information Technology IoT Internet of Things

(5)

5

1. Introduction

This section gives an introduction to the thesis topic themes followed by the research setting for this study. The purpose of the study is defined along with the research objective and research questions followed by scope and limitations. The section also introduces the thesis organization of main and sub-sections.

1.1. Introduction

Adopting industry-specific technologies in hospitality has continued to peak, starting from the early 1970s (Collins & Cobanoglu, 2008). This rise in the number of organizations embracing new technologies is as a result of the beneficial implications of these technologies, such as improving their service delivery. In tourism, the use of new technologies has resulted in the gradual development of a quality tourist experience. Upcoming digitization methods are causing large-scale transformations of various aspects of tourism, such as hotel service delivery and the value of individual business processes. The implications of having this digital revolution in tourism are raising questions and causing primary contests regarding the overall outcome of implementing new technologies (Gretzel et al., 2015).

The improvements in technology have enhanced most methods of conducting business operations that improve user interaction and service delivery. Firms in this sector have to venture into the digitization of all business processes and value chains related to hotel booking and other services in hospitality (Rus & Negrusa, 2014). The entire hotel accommodation framework is enhanced through this technological growth. The efficiency and effectiveness of business processes are pushed further by incorporating technological innovations such as artificial intelligence (AI), machine learning, deep learning, and big data, among other techniques. AI is the simulation of human intelligence in executing machine processes (Bughin et al., 2017).

Machine learning and deep learning are subsets of AI.

The need for quality business intelligence (BI) is a significant problem facing accommodation today (Geisler, 2018). BI can be explained as an assortment of all technologies, applications and activities required for the collection, analysis and visualization of data used in making strategic operational decisions. According to Duncan (2019), an enterprise must take advantage of any combination of technological innovations to improve data obtainability from both the internal and external hotel environment. This combination of technologies forms an organization’s BI (Duncan, 2019).

Jannach (2016) explains that most industrial applications and organizational platforms in the hotel tourism sector have a high demand for added BI features. There is no satisfactory BI even to an organization that uses top market applications (Jannach, 2016). The best intelligence implementation strategy is matching the technology abilities with their user types (Smith &

Lindsay, 2012). Williams (2016) defines capable BI as one that is scalable and flexible in a way that makes it capable of obtaining and adding consistent data to a big data resource (Williams, 2016).

According to Chugh and Grandhi (2013), a good BI is capable of identifying trends and patterns in the aggregate data, which provides critical support for the creation of new strategic business opportunities and predictions. Insights regarding new markets, product suitability, customer demands, market competition, and advertising impact are obtainable through implementing

(8)

8 capable BI (Chugh & Grandhi, 2013). Kimball and Ross (2016) explains the effectiveness of BI is achieved when external data coming from external sources, such as guest involvement in the market, is combined with data gathered from internal sources, for instance, booking data (Kimball & Ross, 2016). The aggregate data must be able to create an intel that is unobtainable from any single data set.

A brief representation of BI in hotel organizations and its articulation is provided by Mariani et al. (2018) showing all the primary fields and sub-components encompassed within intelligent systems in industrial BI (Mariani et al., 2018). These include:

i. Decision Support System: Computer-based information system used in supporting decision-making processes. Proper BI will collect useful insights and make selections and assortments required for strategic growth and development.

ii. Artificial Intelligence: AI operates as a logical system within the BI, which allows the AI to work by providing the ability in terms of extending software applications and specialized hardware (Shi, 2011). AI programs can learn and make decisions rationally, which makes it capable of creating expert business systems that utilize user devices and corporate databases to download and store information intelligently (Mest, 2018).

iii. Data Mining: This element entails the discovery of patterns and correlations within large data sets. AI elements such as machine learning are crucial for the mining process and mathematical modeling.

iv. Descriptive Analytics: This is another essential feature required in the hotel's BI. This type of analytics uses data aggregation and mining methods to give insights regarding the description of an event that has already happened.

v. Predictive Analytics: This feature provides forecasting and gives insights about events that have not occurred yet. This is easily misunderstood for prescriptive analysis, which is also an essential element of BI. The normative analysis gives insights about the strategies to achieve the best and optimal outcomes, possibly predicted by machine learning and modeling techniques.

vi. Big Data: describes large data sets that cannot be handled using a legacy data processing application. Big data requires advanced AI methods to extract and make meaningful texts from the big data resource. Big data has been examined by the research study as a significant component of BI in hospitality.

The intelligent application of named components above develops a high-quality BI that implements models capable of reasoning and cleverly analyzing information to provide problem- solving insights. Knowledge acquisition and knowledge representation are essential as a part of the decision making modeling processes (Sauter, 2011). BI includes system architecture, applications, databases, tools, and methodologies that aim at using data for supporting decision making. A powerful BI with decision support systems provides business process management, which helps improve planning and operations (Eom, 2020). The main idea of this thesis is to examine the use of sub-components mentioned above for various intelligent applications within BI in accommodation. The research must, therefore, explain the excellence of a BI by reviewing the features provided by each subcomponent.

According to Zohuri and Moghaddam (2020), good BI aims at streamlining the process of data collection, analysis and reporting (Zohuri & Moghaddam, 2020). AI helps BI to improve on data obtainability, quality and consistency, which has led to the growth in interest toward AI in

(9)

9 tourism and other industries such as medical healthcare, military, e-commerce, banking, etc.

(Zohuri & Moghaddam, 2020).

The study explores what it means to have robust BI and shows the role of data mining in such a system. By pooling leveraged data, a big data resource is created, which enables the BI to analyze relatable information using mining techniques such as pattern extraction. The gathered information is used to generate usable insights for organizational development and product improvement. This study explores the value aspect of data to the formation of competitive BI in hospitality. For organizations to benefit from data mining in big data repositories, critical data is collected and processed using powerful algorithms that help to search through this data (Bhasin, 2006).

Analytics in the hospitality industry is often used to classify guests using varying factors such as behavior and business trends. Descriptive analytics help to determine the hotel's most valuable guests, thereby focusing their loyalty to such guests. Predictive analytics has also played a significant role in changing the dependability on BI for decision making and business support (EyeforTravel, 2015). Predictive analytics involves different strategies and methods of information extraction and analysis to determine future events correctly. According to Wasserman (2019), BI analytics can compare historical, real-time, and predictive analysis report to develop a granular view of the hotel’s business performance (Wasserman, 2019).

Implementing a quality BI using big data is regarded as an innovative way to guide the growth of a business. It provides techniques through which users can willfully produce relevant data from their participation. The data is harnessed and stored in structured and unstructured forms within a big data repository (Williams, 2016). The study expresses the benefit of big data in improving BI and how a high quality BI is result of a better data collection and analysis. The study will largely perform an analysis centered around some hotels infrastructures that utilize new innovative technological methods to maintain an impressive fashion over other hospitality firms.

Considering the need for any hospitality complex to create dependable BI, AI and big data have proven to be quite useful and beneficial (Williams, 2016).

1.2. Research Setting

The thesis research pays attention to exploring the development of quality BI in the hotel booking and accommodation sector and the response from the individual firms. To achieve this, the research location had to meet predefined requirements listed below that would allow maximum content development from an intensified field. I developed these requirements based on research project objectives.

1. The location should experience a massive dependency on the tourism industry. This will help gather information from a region with the latest developments in tourism.

2. Data-intensive methods used in the selected hotel should include the innovative techniques in modern BI. The research is aimed at examining these components.

3. The region should experience a sharp competition in promoting their services. The competitive difference is observed by the research, which explores how BI helps sustain this variation.

4. Challenges related to tourism should not exist in the selected region i.e., problems such as social conflicts, monument degradation, environmental pollution, political unrest, and conflicts resulting from unequal economic development. The reason for this requirement

(10)

10 is to ensure a smooth time conducting the research. The state of tourism is also affected by these problems.

5. The hotel has considered obtaining customer consent to reveal some private information from a third party organization or obtain the data through intelligent third party applications. The claim will help comprehend how the advancement of a BI can reach for data from these third party sources.

6. The organization in prospect should provide corporate or management support to the investigative process.

7. There should be evidence of the strategic adoption of information technology to support the company's BI. A low application of technology will limit research development.

The targeted enterprise for conducting the research is the A, B, and C hotels in Albania, Balkan region. Due to ethical considerations, the real name of the hotels has been concealed. The location is heavily dependent on tourism based on its luxurious resources and natural endowments. The region also has popular seaside resorts and hotels that attract thousands of tourists internationally every year. Each resort in the sector competes for a more significant share of the market through innovative room designs and unique tastes of luxury (Vasileva, 2017).

Foreign citizens entering in Albania, come mainly from Kosovo (35 %), North Macedonia (11

%), Greece (9 %), Montenegro (6 %) and Italy (7 %) (INSTAT, 2019). The sustainability of tourism in this region is supported by the simple socio-economic structure of the industry serving tourists. The sector under research surrounds innovations in hotel accommodations, which plays a vital role in the tourism value chain within the Balkan region (Vasileva, 2017).

1.3. Purpose Statement and Research Questions 1.3.1. Purpose of the Study

After conducting a thorough survey of the existing literature on the involvement of BI analytics in tourism, a rationale for creating this article is realized. Current investigation studies have focused solely on the innovative development of AI and big data applications and their impact on the hospitality industry. There is a need to conduct research and gather information regarding the high or low quality of a hotel’s BI and how much benefits are added on to the existing system after the integration of intelligent analytics and technologies.

This study aims at identifying the need for supporting BI applications in the tourism sector, specifically, hotel booking and accommodation; to determine the underlying perception of the hotelier regarding influences on the tourism BI (Rus & Negrusa, 2014). The research aims at identifying specific applications of BI configured to support particular hotel booking and accommodation processes. The research problem must be solved by investigating a few hotel booking systems that utilize these new technologies to up their standards (Rus & Negrusa, 2014).

The studied hotels will provide empirical evidence for the following set of research questions.

1.3.2. Research Questions

We have three research questions focused on extracting the required information in this field of tourism:

1. How is BI applied in the hospitality industry?

2. What are some of the benefits achieved by providing competent solutions through the application of a high quality BI?

3. What are the business challenges of using low quality BI?

(11)

11 1.4. Topic Justification

It is necessary to develop meaningful and novel insights concerning the technological building blocks of hospitality. Performing this investigative study on BI in tourism provides more academic knowledge on the existing methods and upcoming innovations incorporated into the tourism industry to sustain the business operation. It is also important to investigate the benefits of providing competent solutions towards the improvement of service delivery in the hospitality industry. This encompasses all major benefits accrued from the application of a powerful BI as well as the negative impacts of the technological advances in the industry.

1.5. Scope and Limitations

The research development is inclined towards examining the growing trend towards the adoption of new types of services accomplished through the implementation of a quality BI. It is necessary to look into the technical support as implications of employing the new methods in business data analytics.

The standardization methods that allow interoperability between hospitality providers and other partakers in the tourism industry are critical to the service delivery of these firms. This has led to a broad spectrum of implementations that assist in service provision by acquiring and processing information from external sources. The research examines these solutions meant to accommodate technological changes. Privacy and security policies in selected research settings limit the ability of researchers to obtain personal and corporate data collected using these methods. The research seeks to identify the problems associated with poor BI analytics in this industry and other problems related to missing solutions provided by the new technologies.

The integration, aggregation, and analysis of big data repositories are an essential part of hospitality provision (Chugh & Grandhi, 2013). Adopting new technologies in this sector of tourism increases the volume of created and shared data, which enables the personalization of services, which is a significant contribution to the growth of hospitality. The study examines how vital data is gathered and accurately used to predict a guest's behavior or to provide a personalized experience. The systems handling this data must be capable of containing the incursion of large data sets from a variety of repositories. Such data would require expert security and privacy protection methods.

1.6. Thesis Organization

The basic thesis structure followed was developed at my university, Department of Informatics.

The outlined structure is obtained from the coursework showing an overall basic structure of a thesis as an argument. The structure provided the core argument for this thesis structure. In the figure 1 below is shown the thesis organization with all its chapters.

Figure 1. Thesis Organization (author’s work)

1. Introduction 2. Review of the

Literature 3. Methodology 4. Empirical

Findings 5. Discussion 6. Conclusion

(12)

12 The topic section gives the main idea of the field of study. The topic utilizes particular keywords and arrangement that allows the article to be retrieved easily through databases. An abstract section presents a brief summary of this thesis. Preliminary information about the study is also shown prior to the onset of the article. The first part of the thesis provides an introduction and brief overview of the topic of study and allows the research boundaries to be presented. This part is constructed from discussions regarding the formulation of this investigative study.

The general precept to investigate is defined in the first part as the adoption of new technologies i.e., decision support, AI, descriptive analytics, prescriptive analytics, big data and mining in hotel booking and accommodation sector in the tourism industry. This thesis section provides a general introduction of various BI sub-components and their contribution to the development of feasible BI. The section prepares the research by defining the setting of conducting the study and gives detail regarding the purpose of conducting the investigative study and the research questions through which the study is performed.

In the second part, a literature review is constructed using gathered research articles obtained.

These include research reports, conference reports, scholarly journals, research monographs, corporate journals and university dissertations as the primary sources. Secondary sources include open access journals, abstracts, web site articles, and textbooks. From these sources, information about sub-components of business intelligence is gathered in relevance to the research topic.

The research methodology is presented in the third section showing the research processes and methods used to collect and analyze findings. The selected methodology is able to cover research process details in great depth. The flexibility of the research is also shown in this section allowing the researcher to relate the research design and coverage with other investigative studies. This section provides data analysis for the gathered information. The analytical methods used to allow proper findings to be determined in support of the thesis.

Observed findings are presented and discussed in the sections that follow respectively to create a critical knowledge base. The limitations experienced during the investigative process are defined in this section allowing the research to express the data obtainability for the given topic. In the discussion section, the paper uses other primary and secondary sources to critic the findings by offering supporting data and contrasting information as well.

In a different section, the thesis article examines the contribution of the study towards available literature. The final part presents the conclusive remarks and makes suggestions on future advances and expected shortcomings regarding the research topic.

(13)

13

2. Review of the Literature

This section performs a literature review of corporate journals and scholarly articles related to the application of various innovative techniques within BI in the hospitality industry. A review is done for individual sub-components of BI to allow an in-depth understanding of various features and functions availed by a good quality business intelligence system (BIS) in hospitality.

This paper uses a definite research structure to examine related scholarly articles with useful information as follows critically. A systematic literature review approach is developed in this study. The method can be described as a reproducible and explicit method for defining essential works from other researchers. In general, the aim of using systematic review is to identify and combine the findings of all relevant individual studies. The research structure allows future developments and updates to be done on the thesis conclusions. The following steps define the approach relevant to the beginning of this study (Siddaway, 2014).

a) Definition of topic b) Search terms definition

c) Source filtration according to search terms d) Inclusion and exclusion criteria

According to Tranfield, Denyer and Smart (2003), the systematic literature review approach allows the optimal identification of relevant scientific works (Tranfield, Denyer & Smart, 2003).

The study has classified the identified articles into some subcategories which are related with BI in hospitality. A deep analysis of the articles was done based on the following features:

● Topic

● Concepts and theoretical definitions

● Data sources

● Data types and sizes

● Data collection and analysis methods

● Reporting and visualizations

As mentioned previously, the articles have been retrieved from large sources databases such as Research Gate, Science Direct, Springer, Taylor & Francis, Emerald and Wiley online library.

The selected databases have a comprehensive library of scholarly articles, and industrial research journals.

To obtain the papers, the following keywords were searched on the databases mentioned above.

The sets of keywords include “business intelligence”, “analytics in hotels”, “data mining”,

“decision support systems in hotel BI”, “hotel industry”, “BI in tourism”, “artificial intelligence”

and “big data” whose results were narrowed down further by matching the found cases with the field of interest i.e., hospitality. The keywords were therefore associated with “hospitality” and

“hotel industry” conjunctions to find related works. Some keywords were matched together to identify more relevant cases, for example “business intelligence” was searched in conjunction with “artificial intelligence” and “hotel industry”.

Results obtained provide a wide distribution of topics related closely to the search terms.

Specific articles used in the study were selected from the results leading to a significant selection of quality works with supportive content for the study. Sampled content from the research articles is also used to support findings in the discussion of the research outcome.

(14)

14 Search terms used to draw articles from the selected sources were developed from the topic definition and purpose of the study. The research focuses on the impact of BI techniques in hospitality. Various roles played by different subcomponents of the BI have been expounded on showing how each element supports the effectiveness of the BI.

The data gathering process began on March 2^nd, 2020, and filtered search results to match the period between 2000 and 2020. The study also includes conference papers and chapters in books consolidated with review and empirical articles. Out of searched items, a quick analysis using an inclusion and exclusion criteria is applied to review and dispose of unwanted documents.

Establishing an inclusion and exclusion criteria is necessary for a quality literature review. It helps sample the large number of relevant articles and easily determine the ones to choose for review as well as those to discard. Inclusion and exclusion of articles prevents biasness in selected sources since it is created based on the research study. The inclusive criteria includes peer reviewed sources as well as studies not earlier than year 2000. The study excludes studies that do not address the impact of BI in hotel industry and tourism.

2.1. Business Intelligence in Hotels

The current understanding of BI involves business processes viewed from historical, present, and predictive perspectives (Kimball et al., 2008). Functions within the BI are categorized into data collection, text and data mining, prescriptive analytics, business performance management, benchmarking, predictive analysis, and reporting behavior. The technologies supporting BI must, therefore, be capable of handling these huge chunks of structured and unstructured data (Rud, 2009). An in-depth literature review has been provided below.

Dzhandzhugazova et al. (2016) describes various ways through which different innovations are introduced and adopted into the hospitality industry. BI encompasses all aspects of the innovative types below, leading to a high adoption and integration process. Innovation in BI covers new supply methods, new products, and service delivery systems, the ability to exploit new markets, and new business organizations making up improved BI leading to better strategic decision-making processes as well as new methods of production. All these aspects form a hotel’s BI (Dzhandzhugazova et al., 2016).

Some newer applications of BI in the top-rated hotel tourism industries seem like something out of a fiction movie. The featured capabilities include facial identification, voice pattern recognition, task automation, conversation through chat bots, and real-time management of distributed applications (Bughin et al., 2016). An innovative trend is observable in the diverse technological advancements forming a large scale implementation of BI in some firms, the combination of particular analytical techniques gives them the ability to create new products and services.

2.1.1. Decision Support System

Decision support needs for advanced hotel booking and accommodation define the need for having an intelligent BI system (Bughin et al., 2017). Decision-makers and information users are the determinants for needed insights that aid in strategic planning. Information users provide demands for certain items or services which is comprehensible through proper analysis of business transactions (Burstein & Holsapple, 2008). Required data regarding the transactions are designed to fulfill operational requirements. Decision-makers formulate business goals that direct the information needs viable for the achievement of those goals (Garrett, 2012).

(15)

15 Burstein and Holsapple (2008) explain that decision-makers have no ultimate expectation for available information. BI is always capable of collecting more data (Burstein & Holsapple, 2008). According to Power (2014), some information needs are dynamic and change rapidly.

Other requirements are unconscious but can be observed from a user’s interaction with an artificially intelligent system (Power, 2014).

Decision support systems (DSSs) assist hotel organizations in obtaining insights and methods that align service delivery and product development with guest expectations. Sauter (2011) explains a DSS as an interactive decision making an application that provides supportive roles to decision-makers (Sauter, 2011). The new breed of DSS is integrated into BI which provides an impetus to a new breed of enhanced BIS. DSS, having begun in academics, has become adopted into business and, most recently, in hospitality. It is integrated with data analytics and other sub- components of BI such as AI and data mining to work more effectively.

Current uses of BI in the hospitality industry have been centered around four domains, i.e., customer applications, business operations, intelligence gathering and automation. DSS provide optimum guidance in enhancing guest experience based on analytical insights. These developments have been analyzed by Raphael (2013) in a study exploring BI as a continuum in tourism (Raphael, 2013). BI is capable of assessing complex tasks and creating a detailed description of events that assist make informed decisions that push business limits further.

Challenges that demand real-time decisions to be made, achieve it through DSSs (Geisler, 2018).

AI, deep learning methods, predictive analysis, and data analytics have ushered in a new era of BI. Decision support can pull data on practical insights and provides an in-depth analysis of complex data through unsupervised learning (Duncan, 2019). The DSS can be designed to monitor the entire system ordering specific data points to pass the required information. Having intelligent systems allows prediction of behavior, which assists in enhancing decision-making processes. For instance, a guest who misses their flight is recorded by the smart system that makes necessary adjustments, such as retaining the original guest room and extending their stay (Jiang, 2013). DSS is, therefore, transforming the way that BI works.

2.1.2. Artificial Intelligence

It is becoming a necessity for hotel corporations to invest in the growing level of big data to uphold service delivery. This growth in use for big data is raising the need to have an AI tool with deep learning and machine learning features (Zohuri & Rahmani, 2019). Figure 2 below shows the hierarchy relation of AI with these technologies.

Figure 2. The hierarchy relation of AI with machine learning and deep learning (adapted from Zohuri & Moghaddam, 2020)

Artificial Intelligence Machine Learning

Deep Learning

(16)

16 The first proposition for AI was made in 1955 by John McCarthy (Bughin et al., 2017). The target was to introduce machines capable of learning and similarly making intelligent decisions as human intelligence. Modern AI deployments within business information systems utilize complex deep learning and analysis algorithms that follow supervised and unsupervised learning (Geisler, 2018). Tsaih and Hsu (2018) show some commonly used algorithms using supervised learning to include simple neural networks, random forest, decision tree and support vector machines, while unsupervised algorithms comprise of Gaussian mixture model, hierarchical clustering, recommender system and K-means clustering (Tsaih & Hsu, 2018).

AI technology is making guests experience much better and reducing employee workload. It also sets the business growth on the most optimum path. In the review provided by Gupta (2018) explains that over 25% of tasks in hospitality within the U.S. will be fully automated by 2030 (Gupta, 2018). Hilton hotels have launched an AI technology that carries out the duties of a concierge. The robot, known as Connie, was put online in 2016 and exhibited features such as human interaction, self-diagnosis, repair and improvement, etc. (Hilton, 2016). The use of robots is an innovative approach that is currently taking up newer forms. AI can exist solely as software components distributed over multiple user devices and business management systems. The full spectrum of having such a deployment is observed to be more beneficial and efficient.

Bergeron, Raymond and Rivard (2004) imply that specific hotel information needs to shape the information technology (IT) architecture to be used by an AI system. The information needs are formed through the definition of business strategies, which are brought together from organized business aims (Bergeron, Raymond & Rivard, 2004). The IT structure adopted from these business aims, therefore, defines the organization’s BI (Zohuri & Moghaddam, 2020).

As the most excellent field solution, AI provides tailored methods of collecting data from users.

User-based applications gather data from responses and choices made by the user. AI can contribute to several fields within the hotel booking and accommodation sector. According to Russell and Norvig (2010), AI has been incorporated into (Russell & Norvig, 2010):

● Availability of hotel and travel information

● Modeling of user decisions

● Analyzing information usage statistics

● Transforming information into knowledge

● Data mining and knowledge extraction

● Voice and image recognition 2.1.3. Data Mining

According to Franks (2012), the explosion of data availability, even from hardware elements such as GPS, which collects essential location data, has drawn attention to data scientists who have realized the value of pooling such data (Franks, 2012). It has led to the adoption of new methods used in data warehousing and management, such as text mining and sentiment classification. Witten at al. (2016) explains the algorithms used for data mining as new methods adopted following the change in computing and interaction. Without these algorithms, it is difficult to analyze data in its original unstructured state. Some of these techniques are classified as machine learning. They are mainly used in data pattern extraction, correlation, and gathering knowledge from unstructured sources of data (Witten et al., 2016). Nave, Rita and Guerreiro

(17)

17 (2018) explain that data mining is needed when creating insights for decision-making support.

The article conducts text mining and sentimental analysis to structure online reviews that are presented on a decision support application (Nave, Rita & Guerreiro, 2018).

Big data methods have made a big difference in the operation of hotel accommodation and hospitality companies. It has created value for extensive data and improved the quality of BI (Gandomi & Haider 2015). Companies offering accommodation and other hospitality services have realized an improved ability to make better strategic and tactical decisions, which also reciprocated into added value in the general tourism industry (Verhoef, Kooge & Walk, 2016).

Kisilevich, Keim and Rokach (2013) propose a BI tool that collects strategic information about other hotels into a big data pool. The tool gathers data such as room rates and location properties, which can be used to discover profitable deals (Kisilevich, Keim & Rokach, 2013). Lu and Zhang (2015) propose machine learning methods such as meta-training and decision trees for performing the estimation of trip purposes for long-distance customers (Lu & Zhang, 2015).

In another work by Tseng and Won (2016), explores data mining and analysis using a force support system. Harmonized data collections methods help create a reliable big data resource containing unstructured and structured data (Tseng & Won, 2016). Mining techniques are necessary for quick information extraction using particular search parameters (Bhasin, 2006).

Analysis techniques make sense of queried information. It becomes easier to improve data quality by turning to the trails of millions of records left by individuals online when using the media sharing platforms. Most of these data is text from comments, reaction data via emoji, image/video/audio, real-time trends producing interest data, etc. (Tseng & Won, 2016).

The process of analyzing big data sources is gaining momentum through the tourism industry as a way to manage large data sets from gathered intelligence (Gandomi & Haider, 2015). Learning algorithms are widely incorporated into big data sources to discover data using patterns or based on rules. Witten et al. (2016) also explains both supervised and unsupervised algorithms help perform tasks in a generalized way without the need for explicit task programming for distinct tasks (Witten et al., 2016).

2.1.4. Descriptive Analytics

Cognitive engagement is the most beneficial feature provided by descriptive analytics. The machine code can process natural language and perform voice pattern recognitions as well as conversing using similar human voice patterns (Shi, 2011). Tsaih and Hsu (2018) argue about the availability of quality guest room services through cognitive engagement (Tsaih & Hsu, 2018). A BIS research indicates that the Starbucks hotel uses an AI product for serving personalized recommendations. The machine code is able to navigate customer conversations and determine their interests from what they say and do. Customers who perform daily running exercises will prefer being able to access a gym while visiting a limited area. Such information is difficult to obtain without proper BI.

Collectively, the generation of real-time knowledge in the hospitality industry requires significant improvement, as suggested by Kolas et al. (2015). This can be done using ubiquitous end-user applications and networks. The article also defines the empirical use of relevant data. In a tourist accommodation scenario, behavior analysis produces data that provides theories for handling similar contexts; for instance, the behavior of tourists staying over during winter

(18)

18 provides customer-based data, which will be used to influence strategic decisions for guests staying over during winter (Kolas at al., 2015).

Entirely, the main focus of harnessing data is to understand the guest better. Descriptive analytics gathers insights that describe guests and hotel positions. In a study of ALE (2018) based on a survey of global hotel operators shows the description and ranks for various pieces of information such as application usage, room automation, internet usage, building management, etc. Building guest profiles used for providing hotel insights about guests scores 74% ranking top (ALE, 2018). Many tourists use online travel agents who do not share their data with hotels. The hotels must, therefore, consider advanced deployments on their applications such as AI to track and collect essential insights that provide better descriptive analytics, thereby improving visibility and value of individual data (Poole and Mackworth, 2017). Hospitality industries are able to build a wealth of data regarding their clients. It becomes easier to describe the buyer persona from such data (Yang, Pan and Song, 2014).

2.1.5. Predictive Analytics

According to Zsarnoczky (2017), tourists are the most critical participators in activities that generate important data (Zsarnoczky, 2017). Intelligent applications harness this data to provide the ability to track tourist’s motives, demands, behaviors, decisions, satisfaction levels, moods, etc. (Jiang, 2013). The organization is at will whether to influence some of these aspects using the same intelligent applications which are done to produce a desired strategic outcome by using systematic methods such as improving the predictability of user behavior.

An example given by Duez (2018) shows the influential advantage of predictive analytics in the personalization of hospitality services. The illustration shows the hotel’s ability to predict what particular guests are interested in using analyzed insights to issue personalized guest experience.

For instance, guests waiting for a travel shuttle bus can obtain an online hotel room key after a pre-check-in is done before they travel. Upon arrival, the guests head straight to their rooms, and the hotel staff receive notifications regarding guest arrivals. The BI system welcomes the guest to the hotel room services and accessories. This is done while collecting crucial data regarding the visitor's stay, and preferences of music, entertainment, meals, air conditioning, bed warmth, travel requirements such as transportation means, etc. (Duez, 2018).

In another illustration, collective behavior can be predicted from past behavioral trends analyzed through predictive analytics. The hotel industry has attempted various forecasting models through the years. An example would be Micros Opera, a property management system built to report on different ongoing transactions. Using this system, revenue managers were able to gather plenty of data used for predicting and forecasting. Trends are analyzed and projected forward thereby creating an automatic forecast system (EyeforTravel, 2015).

In this industry, the customer is more informed than the hospitality provider. They keep up-to- date information regarding new services and provisions in the industry, which forms interests in them. They also maintain knowledge of what prices are currently being offered and which products are competing (Yang, Pan & Song, 2014). To acquire this educative data, new data sources are being utilized (Wood et al., 2013). These include:

● Search engine traffic: Google Analytics is often used to collect web traffic.

● Customer data: Information regarding customer transactions elsewhere such as purchases in a store in other industrial sectors.

(19)

19

● Review data: This is gathered from third-party sources such as TripAdvisor, Orbitz and Priceline.

● Social media: The platforms allow insight into what the general market says about a particular hotel or service as well as having clarity on competitor hotels. Spikes in traffic resulting from a campaign hosted by the organization can be analyzed against other competitor campaigns to identify the way in which hotels compete with each other.

Gathered insights from the sources mentioned above enhance the effectiveness of prediction analytics. A simple query from search engine traffic will show that an individual is interested in certain products (Wood et al., 2013).

The integration of predictive analytics, together with big data helps to form statistical prediction and fast decision making. The proper implementation of well-developed BI provides a prediction application with the ability to gather and analyze data from a big data resource with minimal effort (Geisler, 2018). The engine performs a comparison between current data and historical data through deep learning, thereby creating trend observations and predictions. This is achievable in real-time. Data stored in big data repositories can be accessed when needed to facilitate required insight on current affairs and forecasts.

2.1.6. Big Data

The realization of large data sets arising from continued content creation and sharing has led to the growth of BI (Duncan, 2019). As the extensive integration of different technologies continues to encourage the use of big data, the tools necessary for manipulating information in a big data resource are innovatively integrated into BI. The scholarly article by Koseoglu, Ross and Okumus (2016) describes BI as a popularization technique that is used to create many useful insights into the life of persons, markets and organizations (Koseoglu, Ross & Okumus, 2016).

The technologies have improved data proliferation through platforms that motivate user- generated content. An example would be online social networks that inspire users to create content accessed through user devices.

Hotels should examine three critical areas in the implementation of BI, i.e., the quality and size of available data, business technology requirements for creating the BI and the level of senior management involvement in the BI (Kimball et al., 2008). The quality and size of available data are particularly crucial to the implementation of successful BI (Chen, Chiang & Storey, 2012).

The hotel hospitality industry is considered to be data-rich due to the massive amounts of data generated from business processes and user interactions within this sector (Hsieh, 2009). The resulting big data resource has experienced a segmentation of collected data due to poor technological coordination between hotel organizations. Data used in running corporate networks in a hotel organization also lacks information coordination (Chugh & Grandhi, 2013). According to EyeforTravel (2015), Marco Benvenuti, a co-founder of Duetto Research Center, hotels require an approach for combining data to acquire an accurate picture using complete data. The need for correct data in real-time is critical to the formulation of a reaction from recommendations given by a prediction analysis setup (EyeforTravel, 2015).

Big data presents a challenge to its users due to its scalability and structure. To overcome this challenge, Chen, Chiang and Storey (2012) suggest the implementation of a new BI

(20)

20 infrastructure with tools such as data warehousing provider, business analytical environment and Hadoop cluster (Chen, Chiang & Storey, 2012).

2.2. Summary of the Literature Review

The review presents the significance and impact of intelligent applications in the hotel industry.

It explores how intelligent applications generate broader and deeper demands that promote changes in service deployment and strategic management methods. This eventually results in the new hospitality industry. The review focused on the value of intelligent applications in tourism, indicating that the development of BISs in tourism can promote the change of the industrial growth pattern.

Some newer applications of BI include DSSs that assist hotel organizations to obtain insights used to align strategic decision making. AI is also shown to be an upcoming application in BI. Its deployments within business information systems utilize complex deep learning and analysis algorithms. AI technology enhances guests experience while reducing employee workload.

Data mining as an application in BI is employed in data warehousing and management to perform tasks such as text mining and sentiment classification. Data mining makes it easier to manage millions of records and obtain specific text from big data sources. Descriptive analytics produce insights that describe guests and hotel processes while prescriptive analytics provides forecast predictions on particular business elements.

(21)

21

3. Methodology

This part describes the ascribed research methodology. The research worldview and philosophical assumptions applied in the empirical data collection are examined first, followed by approach and data collection method. This section also provides the validity, reliability and ethical considerations.

3.1. Methodological Tradition

A world view can be described as the belief and perception of the environment under investigation. Creswell (2009) defines four types of world views, i.e., participatory world view, social constructivism, post-positivism, and pragmatic world view (Creswell, 2009). This research utilizes the deterministic philosophy related to social constructivism. In this world view, the researcher has a purpose of understanding deeply about a specific research topic. Knowledge is obtained from interactions with the subject matter presented by relatable sources as well as experiences and considerations from a research participant. The research is highly dependent on the views of participants regarding the subject matter.

Philosophical assumptions show the choice of theories used to inform the work undertaken in the study. According to Creswell (2009), there are four types of philosophical assumptions, i.e., ontological, axiological, epistemological, and methodology (Creswell, 2009). The study follows an ontological and epistemological assumption where ontology defines the nature of reality and requires the researcher to embrace multiple assumptions regarding the nature and forms of those realities to analyze shreds of evidence-based on individuals’ perspectives. Epistemology points towards the nature of knowledge and how researchers understand what they know. It is necessary to review the ontological grounds of the study as well as the epistemological reflections for the given subject perspective. The quality of the research is governed by the ontological and epistemological assumptions over what exists and how we identify what exists.

The ontology of the research setting includes values, properties, events, minds, physical objects and abstract values such as numbers, data segments or sets. The ontology allows variations and reductions by others, e.g., physicalism will dictate that there are no minds, only brains;

nominalism will reduce numbers to symbols, etc. Epistemological claims indicate the means to knowledge obtainability by the researcher, i.e., intuition, sensation, perception and reason from abduction, induction and deductions performed by the researcher. These are the focus of the approach used to learn about reality as a social process and to record the acquired knowledge.

By getting close to selected participants, subjective evidence can be developed using the individual’s views on research topics. Both ontological and epistemological assumptions fall under the interpretive framework, which is a basic belief system that guides actions and understanding. Creswell (2009) defines interpretive frameworks as scientific theories in social sciences that help map the theoretical consideration in studies (Creswell, 2009).

3.2. Methodological Approach

The research process has been designed to follow an iterative approach of data discovery selected for the study. The strategy refers to a systematic and recursive procedure of conducting a qualitative analysis. Using an iterative model of research will help acquire the best results. A qualitative investigation is less focused on statistical data as much as collecting textual data that answers questions such as “why?” and “how?”. Conducting a qualitative analysis is more flexible and focuses on collecting people’s views. Some conventional methods used in a

(22)

22 qualitative study include observations, interviews, and questionnaires (Creswell, 2009).

Adopting a qualitative study approach will be suitable for the research. It will help the research gather more speculative data from the hospitality environment by allowing questions to be adapted to the research setting.

According to Orb, Eisenhauer and Wynaden (2001), qualitative studies are mostly done in settings that involve people’s participation while in their day to day environment. These interactions require an ethical consideration as well, which entails the appropriateness of a detailed investigation plan, source of funding, and within the methodological design (Orb, Eisenhauer & Wynaden, 2001). Using qualitative methods will allow the research to gather non- numerical data, which will be issued via interview transcripts and screen recordings.

3.3. Methods/Techniques for Data Collection

The research study focuses on conducting a qualitative assessment on specific hotel system applications of different innovative and intelligent systems to satisfy the aim of the study, which is to understand the roles, benefits and issues with the improvement of BI in hospitality. Some of the data used by the AI-enabled systems such as hotel booking data, property usage information logs, transaction data and guest experience analytics should be explained in the practical assessment. The study should prove the ability of a BI to address more complicated problems by having more sophisticated, accurate and timely data processing techniques necessary for the process. Data collected from appointed hotels helps to provide an analysis of undefined problems resulting from not having proper BI techniques involved.

Interviews will be a useful qualitative approach for examining underlying issues targeted by the study. Personal interaction during interviews allows the researcher to obtain a rich understanding through off-script discussions. It is necessary for the investigator to inquire and learn more from a specific answer. Interviews have also been considered a fast way of conducting in-depth assessments and understanding the perspective of an individual’s knowledge.

General scientific methods and special techniques such as thematic analysis and expert assessments have been used to draw conclusions based on a conducted study. The study must identify the role of BI methods and other innovative technologies used while examining the extent through which BI has become changed.

3.3.1. Qualitative Data Collection Methods

After the systematic literature review, the study examines the applications of BI in the infrastructure of lifestyle and boutique hotels. Boutique hotels are luxurious upscale environments that utilize varied architectural designs created for intimate settings and pleasurable accommodations. Lifestyle hotels are oriented towards providing various aspects of living lavishly, but normally. This study’s setting selection was determined to be the A, B and C hotels, which are specialized boutique and lifestyle hotels that maintain an exclusive thematic message and design. Inclusive hotels provide a high standard and private guest services and products.

This investigative study must gather all required information from selected hotel participants to produce feasible findings. The use of qualitative research methods will facilitate data discovery in the practical field of study. The study made use of interviews with open-ended questions as the selected data acquisition method. The purpose of using this interview method is to allow the

(23)

23 discovery of highly personalized information and underlying topics from the interview respondents.

Semi-structured interviews provide a comprehensive description of required subjects. The interviewer using this method does not follow a set of formalized questions strictly, but has a free will to blend in new ideas into the questions (Dzhandzhugazova et al., 2016). The approach is used to collect and analyze empirical data, which will be used to support the understanding of roles and benefits of implementing BI, and the consequences of not implementing the technologies.

3.3.2. Interviews

Semi-structured interviews are adopted for gathering data in qualitative research. Semi- structured interviews can be defined as an exploration method that makes use of themed open- ended questions that are designed around the topic of study. The approach is quite flexible, allowing information discovery for previously neglected data (Fielding, 2003). As a result of travel bans imposed to control the COVID-19 pandemic, access to respondents was acquired through online communication platforms such as Zoom Video Communications and Skype to provide real-time interview settings. The records were done via screen recording.

Invitations for interviews were sent to the different participants in A, B and C hotels depending on the type of information required. The questions to be used in interviews are organized in a fashion that investigates the contribution of BI in streamlining activities involved in hospitality services and how it complements hotel staff enhancing their service delivery. Knowledge will be gathered from the selected hotels in this phase and analyzed in the analysis section.

While planning for this research strategy, a preliminary test for the interview process is staged prior to conducting interviews on-site to identify valuable and relevant data. The trial interview process does not utilize a real hotel setting, but simulates the actual interview process to familiarize the researchers with the process. Preliminary testing creates an ethical background for the investigator since interviews are interactive sessions that require ethical considerations. From this process, limitations and problematic questions are identified, which allows better questions to be formulated and improved. The interviewer has a chance to grasp the interviewing process and to revisit best practices in interviews through the trial test method.

A. Interview Participants Career Demographics

The following section describes the career and individual demographics for the interviewees.

The empirical analysis consults data from the hotels’ managerial and staff departments by interviewing 10 participants from the three different hotels in the selected hotels—the interviewer scheduled the respective meetings. Selected interviewees were based on the following selection criteria: participant’s willingness, level of knowledge, availability for interview process and ethical background.

There were four managers from the A hotel, and another two operations managers belonged to the B hotel. Part of the interviewees was randomly appointed staff who belongs to office and administrative support occupations in C hotel. This includes two system administrators and two managers. Their names have been concealed due to ethical considerations.

Number of Interviewed Participants: 10

(24)

24 Table 1. Interviewed Participants (author’s work)

Alias Job Title Work Responsibilities Academic Training

Hotel

A1 Hotel

Operations Manager

Manage organizational goals, values and standards.

Hospitality and Business Management

Hotel A

A2 IT Manager Advice and supervise IT systems. Control data analysis.

Informatics Hotel A

A3 Sales and

Marketing Manager

Manage both the marketing and the sales staff and perform managerial duties to meet the company's

operations goals.

Business Economy

Hotel A

A4 Front

Office/Guest Services Manager

Supervise day to day operations.

Business Administration

Hotel A

B1 Operations

Manager

Ensure all operations are in alignment with optimal business performance standards.

Informatics and Finance

Hotel B

B2 Operations

Assistant Manager

Control operations for optimal business performance.

Business Management

Hotel B

C1 Database

Administrator

Monitor data requirements and usage of data resources.

Organize and analyze data.

Information Technology and Business Management

Hotel C

C2 IT Manager Manage BI and IT sectors and monitor theirs functionalities.

Computer Science

Hotel C

C3 Web

Administrator

Maintain cloud platforms and website functionalities.

Information Technology

Hotel C

C4 Hotel Manager Making sure that all areas of a hotel environment run

smoothly and work together successfully.

Hospitality Management

Hotel C

(25)

25 B. Interview Questions

This part provides qualitative research questions from the interview process. Extracted responses obtained are printed in the analysis section for each question accordingly. The questions have been provided in the Appendix B section. The interviews will address participants using several main questions written in an order that will reveal the deeper areas to be explored. The interviewer is at liberty to diverge from the matter placed to get more ideas. Each participant will receive an email about the consent to participate in the research study before the interview. The consent form is shown in Appendix A.

3.4. Methods/Techniques for Data Analysis

The qualitative research method selected was valid for the thematic analysis technique used for observed textual data. Thematic analysis can be adopted to reflect on the relevance of findings based on reviewed sources. There are a variety of ways to do a thematic analysis. The study adopts Clarke and Braun (2013) framework, which gives a six-step approach to performing a thematic analysis followed below (Clarke & Braun, 2013).

The analysis below is suitable for the qualitative identification of themes from primary and secondary data gathered in the empirical examinations. The selected framework for conducting the analysis will offer a clear and concise understanding of facts through proper data interpretation and summarization (Javadi & Zarea, 2016).

Thematic Analysis

It entails a constructive structure of assessing and categorizing an enormous amount of textual, audio, and video content and briefly discusses it in an essential way for this study. The method is able to identify patterns and relationships in particular themes and concepts for qualitative research. Inferences can be deducted from a proper evaluation of the content (Clarke & Braun, 2013). To perform the analysis, the collected data from semi-structured interviews and conference notes have been provided in the extract sections within each step. First, the data has to be broken down into meaningful units that exist as manageable theme categories by following the steps below to achieve the desirable results.

Step 1: Getting familiar with the collected data

This phase entails transcribing of information from interviews to extract relevant phrases used in creating initial codes. Information is read through and re-read through (Clarke & Braun, 2013).

During this phase of the qualitative analysis, gathered data in the interview transcripts is skimmed through multiple times. Empirical information collected is in the forms of transcript writings, interview session recordings in audio format, observation notes, and sideline comments. The analysis procedure has to thoroughly read and listen through the content, discovering meaning and ideas. This is necessary for the researcher to develop an in-depth understanding of the topics covered and the relations between responses.

The information relevant to the study was collected from the conducted interviews. From that data, extracts were made during the familiarization process.

Step 2: Creating initial codes

Business Intelligence in the Hotel Industry

Master Degree