LATENT PROFILE ANALYSIS OF BEER CONSUMERS WITH THE APPLICATION OF R SOFTWARE

Purpose:The goal of the paper is to analyze beer consumer preferences by applying latent profiles analysis.Design/methodology/approach: Latent profile analysis (LPA) is used to identify the latent (unknown) profiles that are present.The main difference between latent profile analysis and well-known decompositional approach (e.g.conjoint analysis, discrete choice methods, etc.) is that LPA estimates the latent profiles, while in decompositional approach, profiles are prior known, and evaluated by customers.Findings: The results identified two latent clusters, each with distinct preferences.Cluster 2 exhibited higher means for alcohol content, additional flavors, pasteurization, and filtration, whereas Cluster 1 showed stronger preferences for packaging, serving size, and beer color.The best model was determined based on the Bayesian Information Criterion (BIC), selecting the VVI model, where variances vary within and between classes, while covariances are set to zero.Research limitations/implications: Latent profile analysis provides the information on latent profiles, but the determination of optimal number of profiles is challenging.We must rely in statistical criteria (e.g.AIC, BIC).Including too many latent profiles may lead to overfitting, capturing noise, rather than meaningful profiles.Also LPA assumes that, given the latent profile, observed variables are independent of each other.LCA is also sensitive to sample size.Many papers suggest to use at least 500 observations.Despite these limitations, LPA remains a valuable tool when applied carefully, with proper model validation and robustness checks.Practical implications: The findings of this study provide valuable insights for breweries and marketers seeking to refine their product offerings and promotional strategies.By understanding the distinguishing characteristics of the identified clusters, businesses can develop targeted marketing campaigns and optimize their product portfolios to align with consumer preferences.Originality/value: The paper uses a latent profile method that is not widely-known in Poland, and it is not a common method for preference analysis compared to decompositional approach (e.g.conjoint analysis, discrete choice methods, etc.).