Home Depot said Monday that it's building a massive new database that will give it a clearer picture of its customers and operations.
The new system will store information about customers, sales, inventory and company finances from a multitude of computer systems. The world's biggest home-improvement retailer, Home Depot expects to get the database up and running by the end of the year.
The system will replace 50 data mining and analysis systems now in use across the company. The steadily expanding chain, which is adding 200 new stores this year, is outgrowing it current systems, said Bob DeRodes, chief information officer at Home Depot.
Data mining, the use of statistical analysis to uncover hidden patterns in otherwise random information, remains a top priority for many North American businesses, according to a recent Forrester Research study on information technology spending. Through data mining, marketers can target customers with personalized special promotions and other information they are likely to use, dramatically reducing advertising budgets and boosting revenue.
Home Depot anticipates its new system will enable business managers for the first time to quickly gather complex data and analyze trends across the company, such as the products that sold best in a particular week, DeRodes said.
In addition to providing sales and inventory data, the database will collect information about customer purchasing and shopping patterns and will enable the company to analyze the patterns of individual customers. DeRodes said Home Depot expects this to improve customer service by helping stores better match merchandise and inventory to customer buying patterns.
Each of Home Depot's more than 1,400 stores is visited by an average of 30,000 customers every week, he said.
Protection of customer privacy is a major concern for the company, DeRodes added. Some retailers have recently faced criticismfor expanded customer-data gathering tactics.
"We are doing extensive testing and customer feedback," said DeRodes, referring to new ways the company is using customer data. "If it's not acceptable to the Home Depot customer, we won't do it."
The home-improvement company, which runs many of its information system on IBM mainframe computers, will use an IBM DB2 database as the foundation of the new data-mining system. Big Blue last year released a new version of DB2 that expands the product's data-mining features; it competes with products from Oracle and Microsoft.
Home Depot is still in the midst of selecting other data-mining software it will need for the project, DeRodes said. It's currently evaluating data visualization tools, which present information in easy-to-understand charts and graphs, as well as data extraction and cleansing tools, which help gather data from many computer systems and format it in a consistent way. DeRodes declined to name these products or their developers.
So-called dirty data plagues many data-mining projects. Seventy-five percent of the information technology directors polled by PricewaterhouseCoopers last year said they experienced problems related to faulty data. Only one-third of information technology managers at large corporations said they felt "very confident" about their company's data quality. Dirty data can cause retailers to make mistakes such as send multiple catalogs or other promotional mailings to the same customer.