Data Depot

It’s a rainy day at the Home Depot in Secaucus, N.J. Contractors are milling about. Pigeons feed from an open bag of Wagner’s Four Season Sunflower Seed in an indoor garden aisle. Vanities with missing doors are part of the landscape. Near the bathroom fixture, a dusty placard tells potential customers “It’s all in the details.”


PDF Download

Little wonder Home Depot is looking over its shoulder at Lowe’s Home Improvement.

About a quarter of Home Depot’s 1,568 stores are more than seven years old. This one befriended by birds is nine years old. Store appearance and quality of merchandise are critical—when shoppers can go around the corner to your archcompetitor.

And Home Depot’s stores age quickly. For instance, Home Depot’s Paramus, N.J., store is nearly 13 years old, but considering the wear and tear it’s really about 25 years old, according to Home Depot. “If you’ve shopped the Paramus store, you’ll see what I mean,” Home Depot CFO Carol Tome said at a January meeting with analysts.

To compensate, Home Depot is increasing capital spending 48% to $4 billion in 2003, including 70% for new stores, 9% for technology upgrades and 6% for store remodeling.

And there are plans to add 200 new stores in 2003, including about 50 in the Northeast corridor between New York and Boston, where Lowe’s is just arriving. These stores also, though, will compete with themselves. The Northeast corridor already has 160 Home Depots and 85% of Home Depot stores end up within 10 miles of another.

The company says the concentration boosts sales. “Even the most mature areas can have the ability to capture more customers and business,” said Home Depot CEO Robert Nardelli, in a conference call earlier this year.

But it will strain its ability to pick the right locations to get that boost. Nardelli has championed the building of a big data warehouse that combines such factors as income data by ZIP code, population density, traffic patterns, home ownership, Home Depot market share, nearby store sales volume and other factors to find good locations even in the Northeast.

“With new data we are more confident of success,” Nardelli told financial analysts in January. “In the past we’ve done drive-bys and said ‘Let’s put a store here.'”

The data warehouse, built on IBM software and hardware, also will help the company make merchandising and staffing decisions. The company will spot hot-selling items, like a new line of cabinetry, a better-designed tool chest or emergency kits, faster; and decide whether it’s a local phenomenon or not.

Early returns are promising, say analysts. On a first quarter earnings conference call, Nardelli impressed Wall Street analysts with sales data by markets and product category. He also showed the payoff as Home Depot becomes more serious about merchandising high-priced items. Sales of John Deere tractors, rugs and ceiling fans selling for more than $200 pushed Home Depot’s first quarter average ticket to $51.29, an all-time high.

Nevertheless, Home Depot is still playing catch-up with other retailers, say analysts. Lowe’s, roughly half the size of Home Depot, has been using data warehousing since the early 1990s. Up until 2000, Home Depot had centralized data and could run reports selectively, but company-wide access wasn’t possible.

The reasons why Home Depot was slow to create a warehouse of data the entire company could use were two-fold. One, the system would have cost roughly 50% more than it does today, according to AMR Research. Two, Ron Griffin, CIO in the last half of the ’90s, didn’t get a big slice of the company’s capital expenditure budget, which was devoted mostly to opening new stores. The company also didn’t want to distract store managers by overloading them with information when they could be catering to customers, former managers say.

Home Depot began looking at NCR’s Teradata line of big storage devices and IBM as a potential partner for getting a data warehouse up and running. But the company hired Nardelli in December 2000 and a deal with IBM wasn’t completed until September 2002.

Home Depot CIO Robert DeRodes, who replaced Griffin in February 2002, though, is committed to an ambitious overhaul of the company’s information systems. What was good enough to get the company to its first $50 billion in sales is not going to get the company to $100 billion.

In the past, Home Depot developed 70% of its programs in-house. Now, it works with “top-tier I.T. companies” and buys rather than builds. For instance, Home Depot is shelving internal applications for managing its hiring and retention of employees with PeopleSoft “human capital” software. It is installing SAP’s enterprise planning software. IBM Global Services will manage much of the rollout.

When finished, Home Depot hopes, according to DeRodes, “to get the best technology to market that wasn’t available to the competition when they invested in things like data warehousing.”