Keywords

1 Introduction

A tourism destination is considered an essential unit of analysis for understanding the whole tourism system [1, 2]. Tourism destinations are focal points of tourism activity and an important subject of tourism research [3]. Although there are several ways how to define a tourism destination, the current research streams [4,5,6] highlight mainly the fact, that the tourism destination consists of a number of different components. From the supply-side perspective, there are stakeholders of different sizes and structures, while tourists represent the demand side. There are dynamic connections among these components that are many times nonlinear. Furthermore, due to the impact of the external environment, these relations are open and unpredictable [7]. From this point of view, tourism destinations are viewed as complex systems [8, 9].

The complexity and limited power to influence the number of stakeholders resulted in a network approach to the tourism destination and its governance [5, 6]. Destination governance focuses on the role of influential actors, their interests, affiliations, and the roles they play in destination development [10]. In the sense of network governance, destination is seen as a cluster of interrelated stakeholders embedded in a social network [12].

2 An Overview of Network Approach Focusing on Tourism Destination Governance

Based on the idea that the most relevant characteristics of the system are its components and the relationships between them, the network approach started to be applied to study the complex systems as tourism destinations. In the tourism research, the network approach has been neglected for a long time [13]. In the beginning of the 21st century, some scholars elaborated on the idea of network approach, although mainly from the qualitative perspective dealing with aggregated or dyadic relationships. Firstly, inter-organisational relationships in tourism were analysed through the lenses of sociology and management [14]. Then, the building of networks by public and private tourism sector organisations was examined [15].

More quantitative examination of destination networks started in the mid-2000s, focusing mainly on the supply-side networks. The aim was to examine the topological characteristics of destination networks and their peculiarities [12, 16]. One of the major contributions in this domain was the review of the methods of networks science with the application to the field of tourism studies [17]. Further, the applicability of the network approach to identify the most relevant actors in tourism was proven [18,19,20]. Some significant contributions in this domain were made by analysing the innovation potential and knowledge transfer in tourism destinations [21,22,23].

Advances in computer science and the use of information and communication technologies allowed to analyse real and virtual components in the digital ecosystems of destinations [24,25,26]. Until recently, tourist mobility has been examined using the network perspective [27, 28].

Although a conjecture of a universal model of the network structure of a tourism destination that can help in design, planning and governance activities was made [30], there is still a lack of research to prove the contribution of a network approach to tourism destination governance. Moreover, a lager sample of rigorous research is needed to better confirm the results [31].

3 Methodology

The aim of this study is to conceptualise the contribution of the network approach to tourism destination governance. To meet this aim and fill the research gap, two research questions were developed.

RQ1: In which areas can the network approach help tourism destination governance?

RQ2: What are the managerial implications of the application of network approach in tourism destination governance?

The study uses ten networks of mature tourism destinations in Central Europe, Slovakia. In order to meet the aim of the study, different types of networks were chosen:

  • three networks based on the cooperation of destination stakeholders in the High Tatras, Liptov and Central Slovakia [32];

  • four leadership networks based on financial ownership of resources in the High Tatras and Liptov [19];

  • one virtual network based on knowledge transfer in the High Tatras [26];

  • one network based on strategic visitor flows in the High Tatras [33];

  • time series of visitor arrivals mapped into a network [34].

Besides the graphical interpretation of the networks, several quantitative characteristics of a network analysis are used. The global metrics include density, clustering coefficient and modularity. From the local metrics centrality measures are used (degree, closeness, betweenness and eigenvector) as well as the measure of connected components. The analysis was carried out using UCINET 6.720 and Gephi 0.9.3 software.

4 Application of Network Approach in Tourism Destination Governance Research

4.1 Supply-Side Networks

Concerning supply-side networks in tourism, the most common way of exchange (or links) among stakeholders is cooperation. Networks created based on the cooperation of destination stakeholders in product development and marketing communication in three destinations are presented (Fig. 1). The cooperation is concentrated in tourism associations and DMOs (red circles), while the cooperation among individual stakeholders of the same type is not evident.

Fig. 1.
figure 1

(Source: Gajdošík T (2015) Network analysis of cooperation in tourism destinations. Czech J Tour 4:26–44)

Graphs of networks based on cooperation

Full size image

The graphical interpretation is supported by the examination of the quantitative characteristics. When calculating the quantitative characteristics, there is no reference value that indicates what is good or bad. Rather, it is recommended to compare the values among similar destinations in the same stage of destination life cycle or to compare the values with the random network. The random network has the same number of links and density as the original network, but the links are randomly distributed by the computer program using the Erdos Renyi model (Table 1).

Table 1. Quantitative characteristics of network based on cooperation
Full size table

The situation in the analysed destinations revealed that there is a very low density and the clustering of nodes is comparable with the randomly distributed networks. This indicates that stakeholders do not create tight groups with a high intensity of cooperation. In order to find the reasons for a low level of cooperation, the quantitative analysis of networks properties should be enriched with in-depth qualitative knowledge of the destination system. This analysis provides several implications. Firstly, the network approach is capable of identifying the weak connections and thus highlighting the problems in cooperation in the tourism destination that should be further qualitatively analysed. Further, based on the quantitative characteristics, it is possible to benchmark destinations and thus to find the competitive position and to define improvement goals.

Supply-side networks can also be created on the basis of leadership. Leadership can be identified as a key factor of tourism destination development, where leaders provide strategic directions to destinations; however, they need resources and power. Therefore, power and ownership of resources can be taken into account to create the network.

This kind of network was created in two Slovak destinations and the change between two time periods (1995 and 2015 = 20 years) was analysed. In 1995 the major structural changes in the offer of destinations occurred. These changes were present because of the new legislative and institutional framework, as well as because of creating more support for entrepreneurial activities within the business environment. In 2015 the situation was influenced by the existence of the Tourism Support Act that initiated top-down development of DMOs in Slovakia. The situation in 1995 in both destinations indicates that the relationships are centralised to local tourism associations (Tourism Association of the High Tatras, Tourism Association of Liptov–LTA) and municipalities. However, the situation in 2015 presents a significant change towards the corporate-based model as a private stakeholder (Tatry Mountain Resort, Inc. –TMR) is gaining more powerful position (Fig. 2).

The graphical interpretation is supported by discovering the leaders in the networks using centrality measures. These measures are important indicators to point to the privileged positions of some stakeholders compared to other members of the network (Table 2).

Based on these measures, it can be concluded that in the examined destinations, one private stakeholder (TMR) is gaining more power and changes the community-based structure of destinations. The analysis has several implications. The network approach is suitable for the identification of leaders in a destination and provides the baseline for destination leadership research. Moreover, the network approach is appropriate to determine the change in a destination structure and can be applied to measure the reengineering processes in traditional tourism destinations.

Fig. 2.
figure 2

(Source: Gajdošík T, Gajdošíková Z, Maráková V, Flagestad A (2017) Destination structure revisited in view of the community and corporate model. Tour Manag Perspect 24:54–63)

Graphs of networks based on leadership

Full size image
Table 2. Quantitative characteristics of networks based on based on leadership
Full size table

The third examined type of supply-side network is a network based on knowledge transfer. Successful destinations in a globalised, knowledge-based economy are those, where stakeholders, embedded in collaborative network, are engaged in knowledge sharing processes [35]. To identify the knowledge transfer, the technological virtual network was used as virtual networks can potentially mirror and represent the real networks are less costly to study [23]. The network based on website citations was created in the destination High Tatras using the Webometric Analyst application [36]. Due to the limits of search engine Google, where weighted direct link networks of up to 22 websites can be calculated for free, only the most important stakeholders were selected (Fig. 3).

Fig. 3.
figure 3

(Source: Gajdošík T (2022) Smart Tourism Destination Governance: Technology and Design-Based Approach. Routledge, London)

Graph of network based on knowledge transfer

Full size image

Following the graphical interpretation, the most important link is between the Tourism Association and the DMO. Knowledge transfer among other stakeholders is significantly lower. As the network is directed in this case, the centrality measures were slightly modified (Table 3).

For directed weighted networks, the previously used centrality measures have to be modified. To calculate the ‘prestige’ of a node, in-degree centrality counting the number of incoming ties is used. Closeness centrality is not well suited to directed data in fragmented networks. However, the network based on knowledge transfer in the High Tatras is not fragmented, therefore the use of closeness centrality is appropriate. Betweeness centrality can be applied to directed data without any modification.

Table 3. Quantitative characteristics of network based on knowledge transfer
Full size table

The eigenvector is similar to degree centrality, where the eigenvector centrality can be split to two concepts: the right eigenvector corresponding to out-degree and the left eigenvector corresponding to in-degree. In this case, the use of the left eigenvector is used to indicate the amount of direct and indirect potential influence of the node [37]. Based on the geometric mean of these centrality measures, the importance index was calculated, indicating that the DMO is a hub in a knowledge transfer.

There are implications from this analysis. The network approach to knowledge transfer is capable of identifying knowledge hubs. Knowledge hubs manage the flow of ideas, information and innovations. By finding this role, knowledge transfer in a destination can be supported as cooperation with hubs that have a link to external organisations can produce new knowledge [22].

4.2 Demand-Side Networks

The governance of tourism destinations is nowadays challenged by incorporating more demand-related issues related to visitor flows [38]. The following network is created based on passive mobile positioning data (Fig. 4). Together with the analysis of tourism supply, it is a starting point for finding the intersection of demand and supply, as the success of a destination depends on the action of stakeholders and their ability to attract visitor flows and create synergies for these flows.

Fig. 4.
figure 4

(Source: Gajdošík T, Gajdošíková Z (2021) DMOs as Data Mining Organizations? Reflection over the Role of DMOs in Smart Tourism Destinations. Lect Notes Networks Syst 228:290–299)

Graph of network based on visitor flows

Full size image

To explore the structural properties of the network of identified visitor flows in the destination High Tatras, the most relevant metrics of centrality (in-degree, out-degree and betweeness) were used (Table 4).

Table 4. Quantitative characteristics of network based on visitor flows
Full size table

The in-degree centrality counts the number of incoming flows to a place, therefore, it can be interpreted as a measure of attractiveness or popularity (places - Štrbské Pleso, Starý Smokovec, Tatranská Lomnica). Out-degree centrality counts the number of outgoing flows. The places having the highest outdegree centrality can be considered as departing points (places Štrbské Pleso, Starý Smokovec, Poprad). Betweeness centrality is calculated as a proportion of all the shortest paths from one place to the other through the focal one. In this context, it indicates the place that has the highest influence (places - Štrbské Pleso, Starý Smokovec, Tatranská Lomnica). The quantitative characteristic of strongly connected components indicated those nodes that are easily reachable, thus indicating the boundaries of a destination. Moreover, based on new boundaries, new destination stakeholders who are able to attract the visitor flows can be found using online geographic information systems (e.g., Google maps).

To provide implications from this analysis, it can be stated that the network approach is able to identify the roles played by places, which has a significant contribution to flow-based destination management. Moreover, it allows one to determine the boundaries of a destination and find new stakeholders, thus focusing more on the intersection between tourism demand and supply. This creates opportunities to engage new destination stakeholders and better serve the tourist needs.

To understand the complexity features of a tourism destination from the demand side, the network approach can also be used for time series analysis. In order to transform time series into networks, the horizontal visibility graph algorithm is a useful approach. It considers each point in the time series as a node connected with another node under the condition that it is possible to trace a horizontal line not intersecting another intermediate node. In this way, the time series of tourist arrivals to a destination in the High Tatras was mapped into a network (Fig. 5).

The analysed time series represent the daily visitation of the destination High Tatras from 24th December till 12th April 2020. The modularity index of the network Q = 0.781 indicates a good separation between different nodes. The analysis divided the time series into 10 different dynamic phases. These modules thus correspond to periods with similar dynamics or the same business cycle [39]. In this case, it is the same dynamics in tourist behaviour. The most interesting implications are revealed in the seventh and eighth cycles. The seventh phase corresponds to the spring holidays in the western Slovakia region. The eighth cycle grouped together the spring holidays of the eastern Slovakia region with the central Slovakia region. The visitor flows were divided according to geographical location. Based on geographic segmentation, visitors from western Slovakia form one flow, while visitors from central and eastern Slovakia form another flow. The network approach can identify turning points in the time series and is capable of better characterising the visitor flows.

Fig. 5.
figure 5

(Source: Gajdošík T, Valeri M (2022) Complexity of Tourism Destination Governance. In: Valeri M (ed) New Governance and Management in Touristic Destinations. IGI Global, Hershey, pp 119–132)

Mapping time series into a network

Full size image

5 Conclusion

It can be concluded that the network approach has valuable research implications to tourism destination governance. It can help to measure the strength of connections, thus determining the problems in cooperation; it is capable to find the leaders in a destination, thus supporting the theory of destination leadership, it is able to detect the knowledge hubs supporting innovations, it classifies places according to their roles and it is able to find out the turning points in visitor behaviour. These findings extend the previous knowledge on destination governance.

Thanks to the network approach, tourism destination managers can more easily benchmark their destination as quantitative characteristics can be compared with similar destinations. Changes in the structure of a destination can signal the need for a reengineering of the structures and processes in a destination. Finding the knowledge hubs helps to better diffuse the innovation and thus support the competitiveness of a destination. Moreover, as tourism destinations have blurred boundaries, the network approach can help to define them more precisely. The ability to analyse visitor behaviour helps to better understand the strategic visitor flows (Fig. 6).

Fig. 6.
figure 6

Contribution of network approach to tourism destination governance

Full size image

However, this approach also has several limitations. One of the main problems of the use of the network approach in the context of a tourism destination is the data collection. Traditional methods of collecting data (e.g. surveys, archival records) provide past static data and their collection is many times time-consuming and misleading (e.g. high nonresponse rate in network surveys). Moreover, some big data collection methods, e.g. data from mobile operators or sensors in a destination, are too costly. Another limitation is the difficulty of data analysis and interpretation. Knowledge is needed to correctly interpret the quantitative characteristics of network analysis; however, without a further in-depth knowledge of a destination and its characteristics it is not recommended to draw definitive conclusions [6]. Therefore, to analyse and interpret the data correctly, the researcher should cooperate with the destination manager.

Following the results of the presented research, several implications arise for further research. First, there is a challenge of shifting attention from small and big data to “smart” data, thus creating the opportunities to examine the real-time networks. In addition to static analysis, simulations and modelling of a destination system can shed more light on destination resilience.