The purpose of this study was to investigate both customer segmentation and recommendation systems and how they can be used in conjunction. There are efficiency limitations with recommendation systems, especially with the common method of collaborative filtering. There is a tradeoff between the quality of the recommendation and the scalability. Using clustering with collaborative filtering can help to improve efficiency and maintain quality. Moreover, the output of the clustering model can be used in the recommendation messaging for increased personalization. Customer segmentation and recommendation techniques were studied. A k-means clustering model was built using an order dataset from an on-demand food delivery service. Cluster groups were interpreted, analyzed, and then used as a filter for a collaborative filtering recommendation system. The results indicated that predefined cluster groups can be used to improve collaborative filtering efficiency and quality. Furthermore, a discussion on how to implement the customer group definitions into the recommendation messaging was included in the study.