Hedonic ratings and individual definitions of an ideal apple were used to explain repeated choices during one‐week experiment. Participants (77 women, 31 men, mean age 45, range 19–65) rated the importance of six sensory attributes (red, green, sweet, sour, firm, crispy) in apples. Based on ratings, three clusters (CL) were identified: CL1 preferred sweet and soft; CL2 sour and firm; and CL3 medium sour and medium sweet apples. Liking of four cultivars with distinct sensory properties was measured three times (Monday–Wednesday–Friday) using a 7‐point scale (like not at all – like very much). Following each tasting, three fruits from one cultivar were chosen to take home. Liking and choice reflected CL to some extent. Logistic regression models showed that both the liking of the cultivar chosen and disliking the nonchosen alternatives affected the choice. Liking rated just before a choice was a better predictor than the first (initial) liking. Furthermore, CL and hedonic flexibility (the largest difference between the liking scores of the cultivars selected) significantly predicted some choices. Hence, along with liking a product, disliking other options, inclination to also accept less preferred alternatives, and individual preference for certain qualities in apples play a role in the choice. Practical Applications In this study, we show that initial liking does not alone explain later liking. When products are offered side by side, such as in a real shopping situation, the properties of less‐preferred options also affect the choice. Descriptions of an ideal product and the clusters constructed thereof moderately reflect liking. Such definition of an ideal product can be a useful tool in revealing consumer preferences as long as the rated attributes are relevant. The results emphasize the importance of variety of fruits available to satisfy individual, fluctuating consumer preferences.

No abstract