In the past, selling needs brick and mortar store and it is limited to a local customer base. But today, ecommerce websites make it easy for the customers around the world to shop by virtual store. Performance of ecommerce websites depend not only the quality and price of the product but also the customer centric suggestions about the product and services and retrieval speed of websites. There is more possibility to buy a product brand suggested by friends or relatives than by viewing commends from unknown people. We can implement such a customer centric approach using Neo4j database, which provides easy and fast retrieval of information based on multiple customer attributes and relations than relational databases. This system is also compared with other NoSQL database such as MongoDB. MongoDB is well suited for application like E-Commerce because it support semi structured data and high availability. Integrating Ecommerce solution with social network MongoDB is slower and not provides an optimal solution. The MongoDB imposes certain limitations for fast and efficient retrieval. This paper presents the advantages of using Neo4j over MongoDB for E-Commerce applications. Experiment shows the enhancement of performance in terms of time and complexity.