Product Affinity Analysis to Increase Sales using Machine Learning

Decision-making and understanding customer behavior has become critical and crucial for companies wanting to maintain their position in today's competitive markets. Every business aims to improve its revenue and profits by increasing sales. The objective of this study is to leverage customer firmographic data and product sales transaction data, which is drawn from the organization's internal Salesforce system, to build a solution to project the likelihood of a purchase from our existing customer base. The capacity of individuals to identify cross-sell and up-sell opportunities would be improved with a greater understanding of what was sold, why it was sold, and to whom. It also discovers the associations among products and predicts the products that could be projected for potential sales opportunities. Machine learning algorithms like Market Basket Analysis (MBA) using Apriori, Total Unduplicated Reach and Frequency Analysis (TURF) for frequency study, Chi-square Automatic Interaction Detector (CHAID) algorithm for the Decision tree, K-Nearest Neighbour (KNN) and Multilayer Perceptron (MLP) as part of Deep Learning are the techniques used to derive the desired outcome to the problem. The specific outcome includes product affinity analysis and recommendation of products that can be cross-sold or upsold to existing customers or new customers, where the decision tree algorithm achieves the best results among the other machine learning algorithms. Organizations can profile customers that belong to different categories based on these key drivers and propose the same for new customers who belong to any of these categories. Such products could be sold, thereby increasing the sales opportunities in the organization and enabling the organization to reach its goal of achieving sales targets, increasing the customer base and maintaining niche enterprise products.