Our client has been building a desktop-only Java analytics ecosystem for the past 15 years, and it now consists of 20+ solutions tracking 400+ sales indicators. The entire system has become slow and inefficient, with time to apply a data filter of up to 300 seconds and with 30 days to add a new indicator
About the project
The client is a global retailer with several brands under its umbrella that carefully target adjacent market segments; it operates more than 7,000 stores in upwards of 90 markets worldwide.
After due diligence and feature roadmap definition, we put together a team of six software engineers to augment the client’s in-house team, to help them rewrite the software from plain Java to React and Spring. This setup quickly evolved into a full team on the Symphony side, guiding the client through the transformation and owning the project with only one back-end developer and a QA team on the client’s side. Besides rewriting the existing logic, our engineers are implementing new features based on the client’s business and tech pain points, as well as long-term business goals.
Improved time efficiency
In addition to facilitating new business achievements by creating a highly scalable microservice architecture, we’re building a web app that can be accessed from any device, regardless of the operating system or screen size, while considerably reducing the time to introduce new indicators for sales, purchases, warehouse, success, RFID stock, and distribution.
Higher productivity on every level
At the same time, we are working on introducing a composition service for accessing the database that will considerably improve performance and speed up data access. This will allow report generation in under 30 seconds for all key stakeholders from the store to the regional level, facilitate micro-segmentation for better decision-making, and improve comparability across items, product families, and campaigns, as well as offering fully customizable reporting areas and periods.
- Increased analytics speed 10x
- Analytics SLA dropped to 5 seconds with over 120 data sources