Given the explosive growth of the online food market since the COVID-19 outbreak, this study aimed to examine search attributes affecting food sales in online channels compared to offline channels, focusing on a sensory attribute – visibility of the package. The results affirm a difference in the food sales model of on/offline channels before and after adding a sensory attribute. This study utilized POS (point of sale) data from a multi-channel retailer in the Republic of Korea. All products were classified according to five search attributes, and the effect of each attribute on sales in a channel was analyzed. Compared to the model that included only non-sensory attributes, the model that added a sensory attribute was more sophisticated in the online channel, while the offline channel did not have a significant change after adding a sensory attribute. Comparing on and offline channels with attribute effects, product category (frozen food, ambient food, vs. produce), brand type (private vs. national), and visibility of package (non-sensory vs. sensory line) showed significant differences. This study provides evidence that perceived invisible risk, which is a barrier to online purchase intentions, can be affected by package visibility based on empirical sales data. The study presents theoretical and practical insights for the multi-channel retailer and consumer research literature.