EXPLORING CONSUMER BEHAVIOUR PATTERNS AND MODELLING CHURN PREDICTION IN THE FOOD DELIVERY SERVICE INDUSTRY: A CASE STUDY

Abstract

In the current market, businesses are always looking for ways to sustain and maintain their stability. One of the key strategies is to maintain consumer loyalty and create a peak consumer experience. The study focuses on a third-party delivery service that collaborates with restaurants and fast-food stalls, which is finding it increasingly challenging to maintain individual relationships with its growing consumer base. To address this issue, the study uses churn analysis to extract semantic consumer behavioural information from transaction details, restaurant information, and consumer profiles. The churn analysis process involves six stages: forming objectives, data acquisition, data pre-processing, machine learning, data visualization, and data reporting. Datasets were obtained from a food delivery company and pre-processed using various techniques, including data cleaning, data integration, data transformation, and data reduction. During data pre-processing, the churn period was defined, and various machine learning algorithms such as Artificial Neural Network, Support Vector Machine, AdaBoost Classifier, Gradient Boosting Classifier, XGboost Classifier, and Deep Neural Network were used to predict outcomes. The models were evaluated using metrics such as Area Under the Curve, Confusion Matrix, and Receiver Operating Characteristic. The testing of classifiers was performed with varying feature reduction methods. The findings provide valuable insights into the churn analysis process and how it can be used to extract semantic consumer behavioural information, which can help improve consumer retention and satisfaction, leading to increased revenue and business growth.