This paper aims to investigate how large-scale, online wine reviews from Vivino contribute to understanding sensory descriptors, consumer behaviour and linguistic trends in wine markets. Specifically, it analyses how novices evolve their tasting vocabulary through digital prompting and explores regional flavour dynamics not captured by traditional expert reviews.
Using 60 million user-generated Vivino wine reviews over a decade the analysis used descriptive statistics, network analysis and structural topic modelling to examine trends. The impact of Vivino 2019s standardized flavour prompts on reviewer’s descriptive language is evaluated in a regression framework.
Results demonstrate that large-scale online reviews reveal dynamic trends in wine flavour preferences and regional reputations. Regression analysis shows that Vivino’s flavour prompts significantly accelerated the shift from novice (“Pleb”) to expert (“Pro”) vocabulary, with engaged reviewers adopting flavour descriptors at more than three times their baseline rate.
Wine producers, retailers and online platforms can leverage insights from large-scale textual data to refine marketing strategies, identify emerging consumer trends and improve consumer engagement through targeted linguistic interventions.
This research offers a novel approach by using computational linguistic methods at an unprecedented scale. It highlights how consumer-driven digital data can effectively complement and potentially substitute for, traditional expert evaluations. In doing so, these alternative datasets can provide unique insights into the global wine market.
