36 DATA-DRIVEN MARKET SEGMENTATION ANALYSIS – Contributions by Sara Dolnicar
36 DATA-DRIVEN MARKET SEGMENTATION ANALYSIS – Contributions by Sara Dolnicar
For 20 years, since the beginning of my PhD under the supervision of the legendary Josef Mazanec (Dolnicar, 2014) at the Vienna University of Economics and Business, I have studied market segmentation methodology. I was often asked why I was so interested in market segmentation methodology and why I felt my work mattered. It matters because market segmentation analysis – and the data analytic methods used in market segmentation analysis – are heavily relied upon by tourism industry to gain market insights and by academic tourism researchers to develop theoretical knowledge (Dolnicar, 2004). A review by Zins (2008) suggests that about five percent of academic articles published in tourism utilise market segmentation in some way (e.g., most recently Mauri & Nava, 2021; and the role of segmentation in choice modelling, Kemperman, 2021). When broadening the review scope to methods of data analysis (rather than merely the creation of market segments as the outcome of the study) this proportion increases further: ten percent of studies published in tourism journals use cluster analysis (Mazanec, Ring, Stangl & Teichmann 2010). Uptake of market segmentation in industry is also high, with most national and tourism organisations specifying the target segments they focus their efforts on.
Because of the wide uptake of market segmentation in tourism industry and academia, any weakness in market segmentation methodology translates directly into a suboptimal market segmentation solution used as the basis for marketing action by the tourism industry or incorrect conclusions by academic researchers using segmentation methods to push the boundaries of theoretical knowledge. Preventing incorrect conclusions motivated my colleagues – most notably Bettina Grün and Friedrich Leisch – and I to work on improving market segmentation methods for two decades, culminating ultimately in our book titled “Market Segmentation Analysis: Understanding It, Doing It, and Making It Useful” (Dolnicar, Grün & Leisch, 2018; freely accessible at https://link.springer.com/book/10.1007/978-981-10-8818-6) and the free Massive Open Online Course (MOOC) on market segmentation analysis accompanying the book (https://www.edx.org/course/market-segmentation).
What do we view as our most important contributions?
Offering the first sample-size recommendation
Running a posteriori (Mazanec, 2000), post-hoc (Myers & Tauber, 1977) or data-driven (Dolnicar, 2004) segmentation studies with a large number of segmentation variables (such as 30 vacation activities or 25 benefits sought) but only a small sample size leads – unbeknownst to the user of the segmentation solution – to random segments that are meaningless at best and dangerous at worst. Because a sample size recommendation was historically not available, many segmentation studies in tourism and beyond worked with too small samples but were unaware of the consequences