Made-to-Order: Targeted Marketing in Fast-Food Using Collaborative Filtering

To target marketing campaigns most efficiently to a receptive audience and improve order conversion and revenue per transaction, one must first be able to predict the actions of consumers. This is even more true in the quick-service food industry, where new products are launched regularly and individual brands live and die through effective advertising. Rather than focusing solely on demographics and a brand's most popular products, this solution predicts the date range and constituent items of a customer's next order. To predict both the date range and constituent items of a customer's next order, this solution uses a combination of k-means clustering, collaborative filtering, and content-based filtering techniques. K-means clustering is used to bucket users into ranges of dates based on their past ordering behavior, while collaborative filtering examines the orders and items of users to predict their preference for previously unordered items. Co-clustering is used to group similar users and products together, further improving the accuracy of the predictions. Additionally, content-based filtering involves analyzing the content of items to determine their similarity and recommend items to users based on their past behavior or preferences. By using both collaborative and content-based filtering techniques, the solution provides personalized recommendations that are more accurate than predicting the brand's most popular items for all users. Generating personalized baskets of predicted items by user provides substantial gains in accuracy over predicting the brand's most popular items for all users, correctly selecting at least one item in a customer's next order out of a basket of three recommended items in 75% of the test set. An A/B test will illustrate the impact of this personalized prediction system by tracking increases in conversion rate and average order price of those shown recommendations informed by this prediction model vs those shown generic national advertisements.