Required Sample Sizes for Data-Driven Market Segmentation Analyses in Tourism

Abstract

Data analysts in industry and academia make heavy use of market segmentation analysis to develop tourism knowledge and select commercially attractive target segments. Within academic research alone, approximately 5% of published articles use market segmentation. However, the validity of data-driven market segmentation analyses depends on having available a sample of adequate size. Moreover, no guidance exists for determining what an adequate sample size is. In the present simulation study using artificial data of known structure, the impact of the difficulty of the segmentation task on the required sample size is analyzed in dependence of the number of variables in the segmentation base. Under all simulated data circumstances, a sample size of 70 times the number of variables proves to be adequate. This finding is of substantial practical importance because it will provide guidance to data analysts in academia and industry who wish to conduct reliable and valid segmentation studies.

Season-Oriented Destination Positioning for Visitors in the Mediterranean

“…Firstly, the obtained seasonal based samples cannot be regarded as representative because of the lack of seasonal/monthly data of demographics and travel related characteristics. Secondly, Dolnicar, Grün, Leisch, & Schmidt (2014) proposed to use a conservative rule for sample size at least 70 times the number of variables for clustering, but due to financial and time constraints, the study samples could not be larger. Finally, issues for future research are dynamic drivers of tourist benefits over time and across tourism destinations' development.…”

Step 4: Exploring Data

“…As can be seen, NOT EXCEEDING THE PLANNED BUDGET (represented by the arrow pointing in the top slightly left direction) is a travel motive that is quite unique, whereas, for example, Sometimes principal components analysis is used for the purpose of reducing the number of segmentation variables before extracting market segments from consumer data. This idea is appealing because more variables mean that the dimensionality of the problem the segment extraction technique needs to manage increases, thus making extraction more difficult and increasing sample size requirements (Dolnicar et al 2014(Dolnicar et al , 2016. Reducing dimensionality by selecting only a limited number of principal components has also been recommended in the early segmentation literature (Beane and Ennis 1987;Tynan and Drayton 1987), but has been since shown to be highly problematic (Sheppard 1996;Dolnicar and Grün 2008).…”

Step 3: Collecting Data

“…According to Qiu and Joe (2015), the sample size should -in the simple case of equal cluster sizes -be at least ten times the number of segmentation variables times the number of segments in the data (10 · p · k where p represents the number of segmentation variables and k represents the number of segments). If segments are unequally sized, the smallest segment should contain a sample of at least 10 · p. Dolnicar et al (2014) conducted extensive simulation studies with artificial data modelled after typical data sets used in applied tourism segmentation studies. Knowing the true structure of the data sets, they tested sample size requirement for algorithms to correctly identify the true segments.…”

“…5.2, a sample size of at least 60 · p is recommended. For a more difficult artificial data scenario Dolnicar et al (2014) recommend using a sample size of at least 70 · p; no substantial improvements in identifying the correct segments were identified beyond this point. Dolnicar et al (2016) extended this line of research to account for key features of typical survey data sets, making it more difficult for segmentation algorithms to identify correct segmentation solutions.…”