House of Fraser brings together a wide number of popular major retail stores, including Army and Navy and Kendals, and the information needed to run such a diverse business resided in a number of places. "There were disparate pieces of information in small systems--the Finance management information system (MIS), the merchandising MIS. Each spewed out information in the format they chose and typically we had nasty little systems to pull that together."

With nearly a thousand departments within 50 stores, getting to a lower-level view to plan for things like branding and sizing was proving impossible. However, selling an IT-driven data warehouse project to the business was not going to be easy, so Bond says his team introduced the concept in stages, piecing together an Oracle back-end. "We went with a raft of individual medium-term projects, and used it as the starting point." But Bond emphasises the importance of planning the layout of the warehouse from the outset, ensuring it will be able to handle the volume of data further down the line.

House of Fraser began by focusing on supply chain planning in 1996, looking at gross profits, seasonality and forecasting. Immediately, the impact was profound. "We had never before had non-aggregated data." The same year, the transaction-level data began to feed Oracle Express online analytical processing (OLAP) reporting to 800 to 900 managers, while another system helped to monitor the businesses' key performance indicators (KPIs). In 1999, a further application came online to monitor suppliers' performance. Each of these sits on Oracle's Sales Analyser tool.

The focus now is on supporting the company's multifaceted CRM strategy. At post time, the company is launching a "Recognition" loyalty card campaign and is using the history it already has on its Frasercard customers to decide how to introduce it. A call centre coming online to support the loyalty card program will feed the warehouse so it builds a customer history.

It is also introducing a personal shopping service for busy professionals, and based on past buying behaviour can preselect choices of size, brand and colour. In addition, the warehouse is being used to inform preliminary research for a Web service using clickstream analysis of customers on its Website to identify areas of interest. The warehouse has given the company confidence, "so when we make a decision based on the data we know it's informed and we have the confidence to make the changes".