Tourists’ activity patterns and variety seeking behavior: innovative tools
July 14, 2020
Tourism has been “locked down” everywhere and is still suffering dramatically from recent circumstances and the pandemia that has developed worldwide. To recover it needs a big effort and to start anew most effectively and efficiently, taking advantage of all the available resources and of the most advanced technologies.
Tourism is undergoing a radical transformation, driven by the diffusion of digital innovation and the spread and availability of “Big Data”, and among them, of “path data”. Path data describe the positions of consumers in space and time and their movements within the environment (Hui, Fader and Bradlow 2009) and are well suited to describe the processes that tourists carry out in their trips.
Indeed, in their trips, tourists engage in a sequence of activities. Before their trip, they collect information about their destination and about transportation and accommodation options, book transportation and accommodation, and so on; during their trip, they collect last minute and on the spot information, move towards and around the destination, enjoy food, take part into sport and cultural events, exploit entertainment opportunities, and so on; after their trip, they recall information about the destination and about their behavior, communicate and share their memories with friends, and so on.
In this way, they perform activities that bear upon their judgments, their feelings and the relations they establish with their suppliers and relevant stakeholders.
Among the most notable features of these activities is their variety. Indeed, “variety” is an important ingredient in our lives and empirical evidence suggests that consumers are often motivated by a positive bias towards variety and that variety is an important driver of consumer behaviour (Mc Alister 1984). The search for variety is a very common and well documented behavior in consumer markets across industries and it features prominently in tourism and in shaping tourist decisions and experience (Anton, Camarero and Laguna-Garcia 2018).
Research objective and methodology
The aim of our research is to understand tourists’ activity patterns and to analyze their variety seeking behavior, to develop tools to cope with it. Our research is based on the analysis of the sequence of activities, or segments of activities, tourists engage in. It is meant to measure variety in behavior and to cluster these sequences, to perform a behavior-based segmentation of tourists in their activity patterns.
Variety is a multidimensional construct and can be measured by drawing on several metrics such as the number of activities performed at least once, whether few or many (number of unique activities); the balance among activities, whether homogeneous and balanced or heterogeneous and un-balanced (Shannon entropy index); the amount of switching across activities, whether frequent or not (switching index).
Here we avoid technicalities, but just let us mention in passing that in our analysis we apply up-to-date concepts and techniques, like Levenshtein distance (a Sequence Alignment Method), to measure the distance between sequences, and Fuzzy C-Medoids (a FCMd algorithm), to “fuzzy” cluster the sequences and extract “medoids” or observed prototypes of these sequences.
The research setting
We study skiers’ behavior, and we analyze their daily ski trips, i.e. the sequence of ski lift chosen by skiers in their daily trips. Data for our analysis are drawn from the Dolomiti Superski dataset, which reports ordered sequences of seasonal Radio-Frequency Identification (RFID) cards swiping, or “bips” as coded in the dataset. Each record of the database identify the lift visited by a skier. Given this data, we can reconstruct the daily path of each skier for a long time period, and we sample 330 skiers observed during a week vacation.
A number of useful and interesting lessons can be learned. The research shows, first of all, that activity patterns across tourists are heterogeneous, and to account for such heterogeneity we need, in our rather small sample, a large number (14) of medoids or representative skiers. Therefore, service providers should be prepared to deal with, and to accommodate for, a lot of heterogeneity in behavior in their markets.
But our research shows also that heavy and light users do not differ in their variety seeking behavior. Usage-rate is not related to variety in activity patterns, in any of its forms, and there is no significant relationship between the total number of lifts used and the indices we are using to measure variety in behavior. In choosing their targets and defining their positioning, therefore, service providers should be ready to address a broad range of behaviors, from the more diverse to the more homogeneous.
Finally, as expected, indicators of variety are intertwined to each other to a large extent, so that skiers display different profiles of variety seeking consistent across dimensions.
In conclusion, we remember that we have focused our research on tourists’ activities, investigating their variety seeking behavior, a traditional and key concept in marketing. We have gathered path data, that are better suited to understand processes, and we have applied new and improved tools to our analysis. Our analysis confirms, by and large, previous findings on variety seeking, although it suggests also that some of the findings do not hold in our context, and most notably that there is no association between purchase or consumption quantities and variety in behavior.