Putting Stores Where the Customers Are

What are the three criteria for picking a site for a store? The old adage is: location, location, location. Now, with real estate management software, you can add “customers” to the list.

Sophisticated software from vendors such as SAS Institute, PeopleSoft and Siebel Systems helps retailers figure out who their most profitable customers are and where they live. But one dollar store has found there can be unintended consequences from getting to know its customers well.

Family Dollar Stores, the second largest player in the category, surprised analysts in April when it reported it was 30 stores behind in its ambitious expansion plans for the year. The company had hoped to open 565 stores in fiscal 2004, but now it estimates it will be more like 525. Furthermore, instead of a 60-40 split between urban and rural locations, the Matthews, N.C., company said it would end the year with at least two-thirds of its new stores in urban locations (cities with populations greater than 75,000).

The reason for fewer stores overall and far fewer in rural areas? Better site selection technology, says David Alexander, the firm’s president and chief operating officer.

“We’ve continued to raise our standards for a store to pass muster and be approved,” he said in the company’s second-quarter conference call. “When a submittal is sent in, we go through an extensive review of the trade area, we review the demographics, we do sales projections, evaluation of rent, the terms of the lease . . . and from all that we either approve, reject or qualify the deals that come in.”

Family Dollar uses a combination of SAS software and the AnySite real estate selection product from MapInfo in its real estate analysis. The SAS software analyzes existing store performance and customer metrics, looking for patterns such as whether strip mall locations outperform enclosed mall outlets, whether outlets next to grocery stores draw shoppers from wider markets, and why the average store pulls in $8.85 per sale but certain rural stores pull in closer to $10 per sale. AnySite matches these metrics with potential store locations, predicting performance through a range of factors such as customer average age, household income and ethnicity; street traffic counts; and summary data for nearby businesses.

Family Dollar encountered some unexpected results from the analysis. After setting its standards for new store locations, such as an urban site that has a minimum of 8,000 people with family incomes of $30,000 or lower within a one-mile radius, or a rural site with the same features within a 10-mile radius, real estate managers found that urban sites were much more likely to gain approval than rural or small-town sites. Family Dollar will purchase sites and spend up to $400,000 on construction, but prefers to look for existing buildings, ranging in size from 7,500 to 9,500 square feet, that it can lease and renovate for about $40,000.

This led to the unexpected emphasis on urban locales. But why the drop in total store openings? Another unintended consequence of the analysis, Alexander says. It takes longer to open an urban location than a rural or small-town site. Permits take longer to get approved, shopping centers take longer to build—or fail to get developed—and traffic concerns often have to wend their way through city hall.

“In urban markets, when it comes to getting certificates of occupancy and things like that, there are a lot more surprises than I’d like to see,” he says.

Larry Maves, who handles the Family Dollar account for MapInfo, said that while the company is working out the kinks in its site selection process, it’s a nice problem to have. “If you set your goals high, you may have fewer store openings than you planned, but your results are also much more reliable,” Maves says. Family Dollar has been able to boost its new-store sales average (for stores in their first year of operation) from 81% of the chain average two years ago to 96% in 2003. Stores overall average $1 million in sales per year.