Analyzing Customer Buying Behavior
In this competitive digital era with millions of products and services-based companies in the market, most of the businesses try hard to survive and gain a competitive advantage over others. Every company needs a marketing strategy which provides them an edge over other companies in the market. To have a strong edge over the other companies and to have a great marketing strategy it is essential for a product/services providing company to understand their customers and understand how they feel, think, reason, and select between different products and services available in the market. By better understanding its customers and their purchasing behavior, a company can increase its digital presence, improve user experiences, predict how customers will respond to its marketing strategies, retain loyal customers, develop/enhance marketing strategies to create new consuming markets, and increase sales revenue. The purpose of this research project is to investigate the buying behavior of mid-west tool manufacturing company’s customers to enable it to become one of the top brands in terms of providing woodworking plans and products in the market. The data analysis conducted on customer data collected through the company’s websites enabled understanding of the type of customers interested in buying its products, the kind of woodworking projects customers look for, the kind of tools and products (in terms of clamping, joining, routing, cutting, or measuring) they are interested in, the popular channels and sources directing customers to its websites, the type of customers who are likely to purchase its products and the factors contributing towards these consumer purchases. This research project also focusses on predicting whether a targeted customer will buy the company’s product or not if the company provides him/her with some special offers. Machine learning models were built using classification techniques: Logistic 8 Regression, Support Vector Machine, K-Nearest Neighbor, Decision Tree, Naïve Bayes, and Random Forest to predict customer purchases when given an offer and a comparison of these models was done to choose the best model to predict customer purchase behavior. Confusion matrices and prediction accuracy were used to evaluate the performance of the classifier on the test sample. The demographic variables such as age and annual salary played an important role in predicting whether the customer will buy a product or not if given an offer. The developed dashboards and models as part of this research project will enable the executives at mid-west tool manufacturing company to make informed decisions about the company’s future growth. Overall, this project analyzed the what, where, when, and how customers buy the company’s products.