Diffusion of elBulli’s innovation: Rate of adoption in Allrecipes and Epicurious

Abstract

Eating habits provide valuable input to the high-cuisine industry seeking to develop effective strategies to attract new clientele to its restaurants. It is noticeable that many professional chefs working in high-end restaurants have adopted elBulli’s innovation (ideas, practices, cooking style); however, this innovation trend is not so clear among amateur cooks, whose recipes could give us a more accurate idea of people’s food preferences. In this paper, we aim to check whether or not amateur cooks in online recipe communities are also adopting elBulli’s innovation, and if so, at what rate. In particular, we focused on the use of ingredients and cooking techniques. Our methodology is based on machine learning and statistical analysis: firstly, we developed several classifiers that learned the characteristics of creativity in elBulli’s cuisine; secondly, we applied the classifiers to two sets of recipes from Allrecipes and Epicurious, respectively; and lastly, we ran hypothesis tests and compared the results in order to reach an agreement on the rate of adoption of elBulli’s innovation on both platforms with statistical significance. Our key findings are that the recipes on these communities are comparable to the most creative recipes of elBulli in the time spans studied (for which self-presentation may play an important role) but the rate of adoption over time is very slow (we conjecture that it can be due to the lack of rush for innovation on these communities where competitivity is not a real need). We finally propose an extension of the study with the inclusion of demographic data that might be more influential than time on the rate of innovation adoption.