Product extrinsic cues as quality predictors in New Zealand Pinot Noir: Vintage outperforms price

New Zealand Pinot Noir exports have grown significantly due to strong market acceptance. Novice consumers often rely on extrinsic cues (e.g., region, vintage, price) when selecting wines; however, previous studies investigating the relationship between these cues and sensory attributes have been limited by small sample sizes and outdated vintages. This study analyzed a representative set of NZ Pinot Noir wines for common wine public buyer spanning five regions, ten vintages (2011–2020), and varied price points to examine the relationships among extrinsic cues, sensory attributes, and chemical parameters using diverse machine learning methods. The objective was to explore the relationships among extrinsic cues, sensory attributes, and chemical parameters using diverse machine learning methods, with a primary focus on sensory attributes and product extrinsic cues. While Principal Component Analysis (PCA) effectively characterized wines based on chemical parameters and product extrinsic cues, it was less effective for sensory attributes. To address this, feature selection methods—including decision trees and ensemble tree-based methods—were applied to identify key sensory variables, which were further visualized using PCA and scatter plots. Results highlighted sensory similarities between wines from Central Otago and Marlborough, and between Martinborough and North Canterbury. Older vintages were positively associated with quality indicators, and both older vintages and higher-priced wines showed strong correlations with total tannin levels.