Towards Practical, Serverless, Cost-effective, Real-time Pricing for Retail E-Commerce

Real-time pricing, an implementation of dynamic pricing in real-time, is a potent tool with the potential to improve revenue and competitiveness for e-commerce sellers. The broader adoption of dynamic pricing has faced obstacles due to the scarcity of practical, flexible, cost-effective, and readily deployable solutions. The advent of cloud computing has promised to address these challenges by enabling scalable, resilient, and cost-effective software systems. In this paper, we present an innovative solution designed to address this need, aiming to create a practical real-time pricing system for e-commerce, powered by modern cloud computing techniques. Our approach is structured from the ground up, delineating the essential components of a pricing strategy, including prediction of demand and sales conversion ratio. To model and predict these parameters, we employ Linear and Polynomial Regression techniques. These building blocks are then utilized to formulate an optimal pricing strategy. The implemented strategy is deployed within the AWS environment, leveraging serverless services such as AWS Lambda and DynamoDB. These service choices are meticulously selected to optimize deployment costs, ensuring cost-effectiveness compared to alternative options. Our results demonstrate the successful forecasting of pricing parameters and the dynamic adjustment of prices in real-time. The cost-effectiveness of this strategy is further supported by our cost analysis, which finds that implementing dynamic pricing for a medium-volume product with approximately 1 million monthly product views results in a daily cloud bill of about $1 USD.