BI on the line for Carphone Warehouse

The project aims to improve the performance and functionality of the data warehouse while reducing maintenance overheads.

Explained Klaus Zumdick, Head of Business Intelligence at The Phone House, “Our current data warehouse is failing to deliver queries in an acceptable time frame, and the latency of data over our network means it’s no longer fit for purpose. The [new] Netezza Performance Server (NPS) system will allow us to massively reduce the time it takes us to load and recover business critical data. We currently have 500 GB of indexes that take hours to load. Using the [new] index-free data warehouse appliance will shrink these processes to literally minutes.”

The firm is undergoing a business transformation project in which the data warehouse plays a key part. Being able to perform detailed analytics in a telecommunications business is absolutely business critical. The previous platform was not able to offer satisfactory analytics or keep pace with our capacity requirements.

By early next year, Acuma, one of Europe ’s largest dedicated providers of information management solutions and services, will have migrated The Phone House CDR and billing analysis to the NPS system. This will be followed by financial applications and commission flows, until the warehouse is live across the entire German enterprise. There will also be an upgrade to Informatica Version 8 across the group, who will continue to use Business Objects as its main BI software tool, along with SAS for insight and data mining reports.

The upshot will be that, for the first time, the Phone House will be able to analyse data such as calling patterns and customer churn which may be used to stabilise and manage its customer upgrade processes.

The NPS system is a data warehouse appliance built specifically to analyse terabytes of detailed data. It stores, filters and processes terabytes of data within a single unit, analysing only the relevant information for each query. This allows the NPS system to speed through processes that would occupy most data warehouse systems for hours, or even days, thereby enabling dramatic increases in productivity.