Table of Content
- The Client
- Client Challenges
- Xoriant Solution
- Architecture/Data Flow Diagram
- Implementation Approach
- Client Benefits
Our client is a Silicon Valley based eCommerce solutions provider, with comprehensive portfolio of solutions which improve the effectiveness of ecommerce and e-marketing activities of small to medium online retailers. One of our client's solutions allows the e-marketers to precisely measure the effectiveness of their promotions as well as their creative campaigns, thus minimizing the guesswork. This solution enables an online retailer to present the most precise targeted content to the customers and automatically keep track of which combination of choices leads to the greatest "conversions." Our client's solution enables e-marketers to run different combinations of creative using a simple, easy-to-use web interface. Marketers can simply login to the hosted web site, define their tests, and then check the results, examining the best combinations of offers that yield the highest business metrics.
Our client's current solution had good data gathering capabilities and high class usability, but lacked a comprehensive and flexible decision making module. Our client's customers required information that would enable their decisions about the image and the text to be posted on a website for better hits. Arriving at a comprehensive decision model required robust analysis of multiple and changing permutations of different decision parameters. The project required modeling and programming the input process of these parameters and designing and implementing the decision making algorithms, which would then be adaptable to incorporate the changing parameters.
Xoriant team architected, designed, and built a fully functional system with high scalability requirements using Microsoft Internet product suite. Xoriant team spent considerable time in studying the current systems, and different use cases to understand how typical users leveraged the different product features to make decisions. Our teams developed the domain and statistical expertise required to build the product. Based on the observations, we worked with the client to build an exhaustive requirements document. Following this, a multi-talented team performed classic product engineering tasks encompassing architecture, design, development and testing. Our approach included:
Detailed real time participation with the client, to understand the actual needs and concept behind the product.
A mix of both client server and web based architecture that achieved mutually orthogonal demands of ubiquity and network performance on one side and security and connectivity on the other side. The seamless interoperability between MS-Windows client-server and .Net internet environments made this mixed architecture functional.
Partner approach with experts in design and architecture followed by product engineering by the Xoriant development team.
ARCHITECTURE/DATA FLOW DIAGRAM
Some of the highlights of the architecture were:
Layered approach: The solution was broken down into different layers of class libraries to facilitate scalability and speed. There was a unique set of class libraries at the database level that provided a standardized access point to other modules and guaranteed efficient usage of processing power. Another set of class libraries was built to encapsulate the business logic. These libraries acted as links between the database server and the client and made the system flexible enough to accommodate any changes in business rules.
Windows services for resource hungry processes: We took out the routines requiring higher processor power and implemented them as Windows services. The services were executed during the off peak hours when the server had more resources available. This also allowed us to make any changes to the processing routine without endangering user data or without affecting the performance.
The initial development was done on Xoriant servers in India. This helped our client to visualize the product and suggest changes at early changes. This resulted in major cost savings to our client in terms of logistics.
Our client's solution was enriched by a decision-making module with a proven and tested algorithm that helped our client's e-marketing customer plan the emarketing strategy more definitively and profitably. While keeping the usability of the solution intact, we were able to implement very complex statistical algorithms to increase the accuracy of the decisions, while completely hiding the complexity from the user.