PREDICTING PURCHASE BEHAVIOR USING CUSTOMER JOURNEY DATA

• Main Author: Vivianni
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/58016
• Institution: Institut Teknologi Bandung

In the era of digitalization, many conventional fields are being digitized, including marketing strategy. By digitizing marketing strategy, more and more digital data available, and one of the most important data in digital strategy marketing is customer journey. Customer journey is the interaction between customer and marketing channel from the beginning until the customer finishes the transaction at the marketing channel. Predictive analytics became one of the most important analysis to be performed in the data as it resulted in knowing the information of product and behavior of the customer. One of predictive analytics that is useful for customer journey data is predicting purchase behavior based on the user journey. Because the form of customer journey can be different from one marketing channel to another, machine learning model that is used to predict customer behavior can have different performance. Hence, in this research, comparison of machine learning model to predict purchase behavior based on customer journey data will be performed. Machine learning models that are being compared are gradient boosting tree model, logistic regression model, and gradient boosting tree model as feature transformation and logistic regression as predictive model. From experiment, gradient boosting tree model as feature transformation and logistic regression has the best performance among other models.