
Case study
Startselect
What can we do for you?
Using transactional data to generate operational insights and improve the retail experience.
More Case Studies

"When we were looking for a company that could help us build a Data Warehouse in the AWS cloud, Cambrian was recommended to us. When we started working with them, we understood why. Their highly-skilled data engineer(s) took the lead and communicated very well during the project. We’re happy with the end result and have experienced Cambrian as a true partner on our Data & Analytics journey. That’s why we still enjoy working with them to this day."
Geert van Zon - Teamlead Data, Startselect Tweet
The customer
Startselect is Europe’s largest reseller of mobile and game credit, gift cards, subscriptions and digital games. All their products are digital and delivered instantly. Customers can purchase and redeem their product within just a few minutes.The challenge
Startselect generates a trove of transactional data from its web store. Management needed to generate quantitative insights based on such data for performance dashboards, customer behaviour modeling, customer segmentation, and other numerical needs within the organization.The solution
Cambrian used AWS Glue, Amazon Redshift, AWS Lambda and Amazon S3 to build the Startselect data & analytics solution. Many business rules were built into AWS Glue in order to validate data and reject invalid information.
As the Startselect store keeps evolving, changes in the data model of the store translate into an evolving structure of stored data in Amazon Redshift. Redshift’s data model was designed in a way that allowed users to use dashboards that present all the available historical data across various historical periods of the store.
The results
Redshift views were built to allow business users to visualise data in easily accessible and understandable PowerBI dashboards. Additionally, Redshift is used to answer custom data requests throughout the company.Business users have gained a better understanding of Startselect’s performance and are therefore better able to make the right (data-driven) decisions.