Tourism is a sector that has a substantial economic impact on countries. With the advancement of technology, personalized solutions have become essential in the tourism sector. Building a user profile is the most common method to personalize a system. Personalization in the travel industry aims to provide travelers with the same level of support that they would receive from an assistant who knows them best. This can be achieved by analyzing the traveler’s previous activities, demographic information, and characteristics as a traveler. Various new technologies have started to be used in tourism, and one of them is blockchain, which can be a critical element in the tourism industry in the following years. Travelers must present their identification at various points throughout their trip. Blockchain-based identification technology allows users to go paperless throughout their journey. Digital transactions can enhance the payment experience for visitors at tourism destinations. Destinations that offer digital payment services are viewed as smart tourist hotspots and tend to be more successful in attracting visitors. In this perspective, we added a digital payment system to our system. In this project, as a sub-part of the Celtics-Next project, SmarTravel, we planned to develop the personalized tourism system of the future, in which the customer is the focus, with three institutions: Setur Tourism Agency, Turkcell Technology, and Paycell. We hope that our project will set an example of how payment processes and blockchain-based identification can be integrated into customer-oriented smart systems.
1 Introduction
Tourism is one of the largest industries in the world, accounting for almost 10% of global GDP in 2019, according to World Travel and Tourism Council [24]. It is a sector that has a substantial economic impact on countries [13]. In recent years, with the advancement of technology, personalized solutions have become essential in the tourism sector [8]. Personalization in the travel industry aims to provide travelers with the same level of support that they would receive from an assistant who knows them best. This can be achieved by analyzing the traveler’s previous activities, demographic information, and characteristics as a traveler [23]. Building a user profile is the most common method to personalize a system [9]. User profiles have been created mainly by analyzing the items preferred by the users [3]. This information is then used to create personalized recommendations for tourists, such as where to travel and stay. Personalized smart tourism can offer benefits for both tourists and tourism destinations. For tourists, it can help them to have a more enjoyable and memorable travel experience. For tourism destinations, it can attract more tourists and increase tourism revenue.
The analysis of user actions enables the configuration of users’ habits, subsequently facilitating the construction of a comprehensive user profile [6]. Sellers employ both implicit and explicit information extraction processes to acquire knowledge about users based on the known attributes of the items and the information users provide [7]. Primary data about users can be obtained from multiple sources, including user information, users’ sales history, and their activities on the company website.
Anzeige
In recent years, various types of new technologies have started to be used in tourism, and one of them is blockchain, which can be a critical element in the tourism industry in the following years [22]. One of the primary advantages of blockchain technology is its ability to store data securely and transparently [26]. Travelers must present their identification at various points throughout their trip. Blockchain technology allows users to go paperless throughout their journey. Additionally, the elimination of identity-based checks for tourists ensures a smooth and hassle-free experience [2]. Another new technology in tourism is digital payment, which has been recognized for its potential to enhance the payment experience for visitors at various destinations [12]. Destinations offering digital payment services are often perceived as smart tourist hotspots, resulting in increased visitor attraction [19]. From these perspectives, we added a digital payment system and blockchain-based identification to our system.
In this project, as a sub-part of the Celtic-Next (C2020/2-3) and TUBITAK (9220043) project, we planned to achieve three main objectives: 1) Collaborating with Setur Tourism Agency, Turkcell Technology, and Paycell to design and develop the SmarTravel system, which incorporates digital payment and blockchain-based identification; 2) To develop a customer-centric personalized tourism system that utilizes the user’s digital payment history, purchased product features, and travel personality; 3) To demonstrate the advantages of integrating blockchain-based identification and payment systems into customer-oriented smart tourism services. We will analyze user activity logs (views, clicks, add to cart, purchase) to evaluate the success of this system. We will try to show that analyzing users’ traveler type, purchasing habits, and features of the purchased product is crucial in the personalization of tourism-specific systems.
2 Methodology
The methodology for profiling in tourism applications is a multi-step process that involves collecting, analyzing, and interpreting data from multiple sources. The methodology aims to create a comprehensive user profile that can be used to provide personalized recommendations and improve the user experience through the collaborative efforts of three distinct companies (Fig. 1).
Fig. 1.
Project structure
×
2.1 Identification
Travel industry increasingly relies on technology and digitization to create a safe and seamless passenger experience [17]. Passengers today can book their flights and check-in online, have their boarding passes on their smartphones, go through automated clearance gates, and even validate their boarding passes electronically to board planes [10]. As presented in the 2017 report on Digital Borders [21], enabling a secure, seamless, and personalized journey and incorporating new technologies into the process will dramatically reshape how the industry and governments manage the secure cross-border movement of people.
Anzeige
A digital identity that includes biometric and biographic data enables the traveler to authorize entities in their journey to access selected information about them to allow for risk-rating, verification, and access. Digital Identity App is a digital wallet that authenticates the user’s own identity using the credentials that have been issued. Self-Sovereign Identity (SSI) solution enables people to manage all of their digital credentials from the safety of their own phone [18]. A self-sovereign identity is “owned” by the individual. As an owner, the individual has access to, can refer to, and share components of this identity at their discretion. While certain components of the identity are established by issuing authorities (passport number, bank details), the individual must consent to sharing their identities and any related data. This is achieved by individuals securely storing their own identity data on their own personal devices and providing it efficiently to those who need to validate it without relying on a central repository of identity data.
Self-sovereign identity systems use blockchains - distributed ledgers - so that decentralized identifiers can be looked up without involving a central directory. That allows people to prove things about themselves using decentralized, verifiable credentials just as they do offline. They provide a transparent, immutable, reliable, and auditable way to address the seamless and secure exchange of cryptographic keys. However, smart identity cards can be used for registration and validation in a secure and confident way instead of an application. These ID cards involve written personal information on the front and backside, a chip that stores personal information, and an NFC area on the backside. Figure 2 shows a flow for registration to a system with the information taken from the card. The first three steps are the same as the registration flow in the Smart card readers’ verification of the customer’s identity process, and it continues: 1) Reading personal info from the card after verification, 2) Personal information is taken from the card inserted into the system.
Fig. 2.
SmartCard registration flow
×
NFC Technologies. There is an NFC area on the backside of ID cards. An NFC reader can scan the area to receive the information. The information is used for registration/validation similar to smart card reader processes. The customer chooses the identification types (Digital Identity App/ Smart Card Usage) before registration/verification processes, and which type preferred is asked to the customer by a dialog box on a system or SMS.
As the ecosystem evolves the number of parties grows, and the number of data sources, protocols/standards grow exponentially to create new value streams. The future connected ecosystem must be resilient to interoperability challenges facing data holders, consumers, and owners. This system will require an economical means to connect many organizations, services, and people together to create value - this will require a number of evolutions of authorization technologies to support data interoperability, identity interoperability, and local governance supporting connected communities that travelers will interact with. The future state of authorization will have some (but not limited to) of the following capabilities: 1) Interoperability with legacy and future identity paradigms (Traditional, federation, and SSI), 2) Ability to govern and interoperate with services and wallets authorized as agents in the ecosystem, 3) Integration pattern that supports non-direct, dynamically established connectivity between data holders and consumers, 4) Interoperability across regions/jurisdictions (locally relevant, internationally compatible), 5) Aligned to and supporting the internationally accredited trust frameworks.
Identification data with NFC is not stored in a static database. These data have been developed to be processed within the system by generating digital data created within the scope of General Data Protection Regulation (GDPR) [1]. The requirement for the use of NFC technology is aimed at the efficiency, accuracy, and speed brought by the technology. For this reason, it is ensured that the identities can be read quickly and their accuracy is determined. Information containing personal sensitive data (e.g., blood type and gender) that may be included in GDPR will be included in the blockchain system given below.
Blockchain. technology ensures the security and integrity of data through cryptographic techniques. Transactions are grouped into blocks and linked to previous blocks to form a blockchain. This connection is achieved through cryptographic hash functions that generate unique digital fingerprints for each block. Changing a block requires recalculating its hash and subsequent blocks, making it almost impossible to change [25].
The Camenisch-Lysyanskaya signature scheme enables secure and anonymous authentication. It allows users to prove their qualifications without revealing their true identity. These signatures securely encrypt identity data and store it on the blockchain. This method prevents the direct sharing of personal information, creating anonymous attributes that are verified against their peers in the blockchain [20]. Signatures are based on mathematical algorithms and cryptographic protocols and ensure user identity privacy. They offer authentication and authorization, preventing data breaches and unauthorized access [20]. Using blockchain and cryptographic methods such as Camenisch-Lysyanskaya signatures, the SmarTravel system provides a secure identification process. Users can interact with confidence knowing that their data is protected. In Fig. 3, the flow diagram shows how user information undergoes cryptographic transformation and how the corresponding signature equivalents are stored on the blockchain. This allows users to verify their credentials at any time, while preventing information from being stolen or copied.
Fig. 3.
Self-Identification architecture using cryptographic signatures
×
The “Proof of Signature Submission” step is executed by the CL Signatures system to verify user identity, Fig. 3. “UTVerkey” refers to the Verification Key, which is part of the user’s unique key pair. If there is a prior connection, the key is retrieved from the wallet; otherwise, new keys are generated. “keyValue Digital ID Attributes” contains the requested user credentials, where “key” refers to attributes such as first name, last name, and ID, while “value” represents the corresponding values. All this information forms the attributes of the credential.
2.2 Payment
The SmarTravel app transcends traditional payment paradigms by implementing a digital wallet API. This visionary solution safeguards customers’ card information with the utmost precision, simultaneously expediting transactions and enhancing the user experience. Powered by state-of-the-art card tokenization techniques, the digital wallet API replaces sensitive card data with impenetrable tokens. This approach mitigates data breach risks, rendering the tokens unusable to malicious actors. With this innovative layer of security, customers can confidently engage in seamless transactions, their private information shielded from harm.
The Paycell payment APIs integrate into various platforms (web, mobile applications) offered by member businesses, providing tourists with a frictionless payment experience through Paycell-defined mobile payment methods. These APIs empower member establishments to proficiently manage payment methods within Paycell, encompassing functionalities such as addition, update, mobile payment activation, and removal of payment methods. In this framework, member establishments ensure payment transactions through Paycell, leveraging Virtual POS information from banks tailored specifically for mobile payment methods. Furthermore, member businesses can incorporate mobile payment services, allowing travelers to conveniently make payments that effortlessly sync with their invoices. The payment process is simplified by selecting a Paycell-defined mobile payment method. Travelers can input their mobile payment phone number, followed by verifying the one-time password (OTP) sent to them. Notably, this integration design is exclusively dedicated to mobile payment methods, entirely excluding card-based transactions.
It is imperative to highlight that the functions do not incorporate a customer verification flow within their workflows. It is assumed that member establishments’ applications consistently verify customers through the transmitted MSISDN (phone number) during the integration process. Transactions encapsulated within the ambit of mobile payment method management include: 1) Payment Method Inquiry: Enabling the listing of customer mobile payment method information defined in Paycell; 2) Enabling Mobile Payment: Facilitating the activation of the mobile payment service for customers.
Transactions facilitated within the domain of mobile payment transactions encompass: 1) Payment: Seamlessly facilitating customers’ payments by allowing them to enter their mobile payment phone number and verifying the OTP. This functionality is exclusively tailored to mobile payment methods; 2) Process Result Inquiry: Empowering the inquiry into the status of executed operations; 3) Cancellation: Enabling the cancellation of mobile payment transactions on the same day. 4) Refund: Providing the ability to initiate payment cancellations or partial refunds for mobile payment transactions from the subsequent day onwards; 5) Reconciliation: Ensuring the synchronization of transaction details between member establishments and Paycell, aligning quantities and amounts within the realm of mobile payment transactions.
The integrated SmarTravel and Paycell dataset offers valuable insights into tourist behavior and spending habits, enabling informed strategic decision-making. Key features include payment date/time, amount/category, payment type, location, demographics, and exchange rate/currency. This rich dataset allows for in-depth analysis of tourist activity, popular categories, payment methods, regional trends, and demographic-based preferences, making it an essential tool for understanding the tourism industry.
In the context of the SmarTravel project, an in-depth examination was conducted on the financial data sets collected under the name Paycell. Specifically, a segmentation study was carried out based on the payment behaviors of the customers and their levels of activity. The resultant segments obtained through the analysis are delineated as follows: 1) Highly Active Customers: Individuals within this segment have conducted more than ten payment transactions per week. Their transactions primarily consist of daily small-scale amounts and frequent purchasing tendencies. 2) Moderately Active Customers: Users within this group have engaged in 5 to 10 transactions per week. They exhibit a balanced payment pattern, encompassing both online and physical store transactions. 3) Low Activity Customers: Individuals within this segment have executed between 1 to 4 payment transactions per week. Notably, substantial-value purchases or monthly payments are distinct characteristics of this group. 4) Quiet Customers: Users within this segment have conducted sporadic transactions every few weeks, displaying low-frequency activity patterns. 5) Dormant Customers: Individuals in this group have refrained from any payment transactions within the past month. They potentially constitute a retrievable customer segment. Listed segmentation outcomes collectively provide an extensive overview of customers’ payment tendencies.
The outcomes derived from our segmentation model provide a foundational basis for comprehending tourists’ behaviors, preferences, and needs. Utilizing these results, Paycell possesses the potential to offer more personalized experiences for tourists, such as personalized marketing campaigns, product and service customization, strategic business decisions, and recommendation systems. In conclusion, the outcomes of the segmentation process are key to providing tourists with personalized, effective, and memorable experiences. This step is critical for enhancing tourist satisfaction and optimizing Paycell’s business model and strategic approach.
Paycell implemented a different payment method called ‘Charge to My Invoice,’ enabling users to charge the cost of their desired product to their mobile phone bills. This option diverges among various mobile payment methods, particularly in permitting customers to opt for a card-independent payment approach. Lastly, Paycell’s member businesses integrate with all the services within the payment infrastructure separately. The aim is to enhance this process through a plugin to achieve further efficiency.
2.3 Profile Card
We create the user’s profile card with two main features: previous activities and demographic information. Previous activities have the following features:
Purchasing: Paycell extracts segmentation of the users according to payment activities.
Hotel: Setur extracts purchased hotel information of the user from previous activities, such as hotel type, hotel location, room count, and room type.
Campaign: Setur extracts the campaign information that the user used.
Holidays: Setur extracts which public holiday they purchased the hotel and stayed.
In marketing activities, business units utilize profile cards to send campaigns to customers. If users frequently make purchases on public holidays, targeted promotions are sent before these holidays. To retain regular customers, campaigns offer additional incentives on their preferred products. Furthermore, if users demonstrate specific hotel preferences, the campaign team suggests similar options from other regions.
The specifications extracted for the hotel were collected under four groups:
1.
Hotel Type: Grown-Up, Family Friendly, Disabled Friendly, Pet Friendly, Luxury, Holiday Hotel, Town Hotel
2.
Hotel Services: Room Service, Transfer Service, Tour Service, Cleaning
3.
Hotel Features: Car Park, Restaurant, Bar, Shopping, TV, Internet, Night Club, Casino, Aqua Park, Indoor Pool, Public Beach, Outdoor Pool, SPA, Golf, Riding, Hobby, Adventure Sport, Sport, Activity
4.
Hotel Location: Historical Destination, Nature, Seafront, Near the sea
The features extracted for the public holidays include public holiday names (such as Ramadan), preferences (1: if public holiday preferences exceed 30% of total purchases), and total count of purchases taken during public holidays. We extracted the features for the campaign that include the total count of early reservation discount usage and special discounts. In addition, Setur adds the average number of days a person stays, the average number of days in advance the purchase is made, the date of the last sale, the last check-out date, status (“Active” if the last release date is within the last two years, otherwise “Passive”), and frequency information (If the customer makes purchases within an average of 460 days (\(\approx \)15 months), it is considered “Frequent”, otherwise it is considered “Infrequent”. If the customer has a total of one purchase, it is directly labeled as “Infrequent”.).
2.4 Hybrid Personalization Model
The categorization of customers into distinct types is a crucial aspect of customer-oriented systems. This is particularly significant in the context of understanding traveler behaviors, where the classification of travelers in existing literature is taken into account. Researchers propose that a user’s travel history can be perceived as a portfolio, which effectively portrays their travel persona [4, 11, 16]. Consequently, by examining the characteristics of their travels, it is possible to gain further insight into a user’s travel behavior [5, 14, 16].
Based on the works of Park et al. published in 2008 [14] and 2010 [15], twenty travel personalities were identified, as depicted in Fig. 4. Participants were asked to select three of the twenty travel personalities listed in Fig. 4 in order to provide a more comprehensive understanding of their travel preferences and behaviors.
Fig. 4.
Travel personality question
×
We analyzed the user’s purchase history with defining features of items (hotels). We extracted the features of the hotels from the provided information from the hotel owners, as mentioned in Sect. 2.3. We developed a hybrid model using the nomenclature in the literature and the user’s previous purchase history. The system will feed and become more accurate as it is used.
Upon user registration for the application, a questionnaire is presented as an optional task. Users have the option to solve the questionnaire at any time, but they can win a discount for solving it, especially the users who have at least three purchase records. We pick their three most recent and most frequented sales records. We match the specifications of the hotel the user traveled to with the travel personalities we obtained from the questionnaire.
The hotel recommendation system takes the user’s traveler type choices and its matched activities. The system selects the hotels with the relevant activity under the traveler types from the database; shared activities between traveler types will have priority. The system ranks the received hotels from most to least according to the total sales count. The recommendation system presents the top 5 hotels as suggestions to the user based on the latest ranking, as illustrated in Fig. 5.
Fig. 5.
Hotel recommendation example in sliding window
×
In tourism applications, the systematic collection and analysis of user activity logs are instrumental in optimizing and evaluating the system. We recorded product list views, product list clicks, add to cart, and purchase activities (logs) of users by personalized recommendation. Our application is currently open to a limited group of users in a closed testing environment. We showed users a total of 3253 distinct hotels with the personalized recommendation system. Twenty-six hotels have more than ten thousand views in total. The click-through rate of these views was measured as 2.3%. In total, the rate of these clicks being added to the cart was measured as 5%. The purchase rate of the products added to the cart was measured as 60% in total. While the number of views of the best-selling hotel list is 25.39% more than the number of views of the recommendation system, the number of clicks of the recommended hotels is 35.28% more than the number of clicks of the best-selling hotel list.
The platform of the SmarTravel project is used by all partners communicating via RESTful APIs, as illustrated in Fig. 6. Each partner processes the confidential information on their local servers and shares only algorithmically necessary information and algorithm results with other partners via APIs. Similarly, services such as payment infrastructure, user authorization, and customer segmentation will be used jointly by partners.
Fig. 6.
SmarTravel SuperApp illustration
×
3 Conclusion and Future Work
Digitalization of identification systems, digital payment, and personalization can enhance tourist satisfaction and brand loyalty. Gaining insights into tourists’ behaviors and preferences can provide a distinctive competitive advantage in the increasingly competitive tourism market. The segmentation results attained will open doors to providing personalized experiences that improve tourists’ experience in countries. The segments generated from customer purchase history can serve as a robust foundation for augmenting the efficacy of marketing campaigns, pinpointing new product and service opportunities, and making informed strategic decisions based on customer preferences. This approach can contribute not only to short-term sales increases but also to strengthening long-term customer relationships.
In summary, this methodology revolves around the synergistic efforts of Setur, Turkcell, and Paycell, whose combined expertise enables us to create a highly customized and user-centric tourism application. Through the examination of user identification, payment history, and personal preferences, we seek to deliver the most relevant and appealing hotel recommendations to our users. User activity results showed that the personalized recommendation has a 35% higher click-through rate than the best-selling hotel listing.
Open Access This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.
The images or other third party material in this chapter are included in the chapter's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
