An integrative method for identifying different types of apple lovers : who and why?

A fine knowledge of the drivers impacting product acceptability and underlying consumers segmentation implies an integrated approach taking simultaneously into account the consumer likings, their habits and attitudes as well as the sensory attributes of the products. A new approach based on the Clustering around Latent Variables analysis with L-structured data, called L-CLV [1], is applied on data collected during a large study on apples (COSI-VEG project, 2009-2013). A conventional sensory profile of 31 apple cultivars was carried out by 15 trained panellists, while 224 consumers were asked to score their liking on the same products using a 9-point hedonic scale. The consumers also answered to a questionnaire on consumption habits and attitudes. Three clusters were identified by the L-CLV approach. The first and second segments (37% and 43% of the respondents respectively) were similar regarding their sensory drivers (they appreciated sweet apples with high aroma intensity) but were differentiated regarding the age category, purchase criteria and apple knowledge. The consumers of the third segment (20% of the panel) had a very specific pattern of sensory drivers (they did not reject a fondant texture, appreciated the rustic aromas and rejected acid fruits). The analyses were also performed on the same data sets using a more classical way to proceed which consists to work in two steps: first, the external preference mapping was performed and led to segments of consumers having similar drivers of preference; then the segments were described using the consumer attributes. It was shown that, in the case study under consideration, by directing the segmentation towards the sensory drivers, no further relevant information was gained from the additional information collected from the consumers. For a better investigation of consumer preferences, from a more marketing viewpoint, it seems more relevant to introduce all the available information simultaneously and that L-CLV is suitable for this purpose.