Case Studies

Discover Your Best Customers for Each Store Location

We can determine your target area by incorporating latitude and longitude right into the modeling process.

When consumers are making a major purchase, they consider various factors before buying. Some shop for the best price, some look for the highest quality, and some are concerned with convenience of shopping for the item. One multi-location furniture retailer was looking to identify these and other characteristics of potential customers to create a targeted mail campaign. The retailer turned to predictive modeling to help it identify who was most likely to make a major furniture purchase at each of its four store locations.

Use Trade Area Segmentation to Find Your Best Customers

Predictive modeling is used to create targeted prospect universes that are customized for every campaign. By applying over 750 variables to data on more than 120 million households, a model is generated to score and produce highly-targeted mailing lists that can improve response rates by 7.5-15% or more.

The furniture retailer was embarking on its first direct marketing campaign as a new way to reach out to its customers. With four locations spread out across two states, the retailer not only needed to find the people most likely to make a big purchase, but also who among these prospects was likely to shop at which store.

A separate model for each of the four locations was built to locate the most likely responders and a trade area analysis for each was done to determine how far the geographic target area around each store extended. This was accomplished through segmentation at the block group level using latitude and longitude to determine if there were geographic barriers that would prevent people from a certain neighborhoods from patronizing a specific store, like a river or one-way streets to circumnavigate that posed an inconvenience. The retailer sent 600,000 mail pieces to the best prospects for the four stores, with excellent results.

The Results

The response rate to the mailing was about 1%, resulting in $7.2 million in sales, with an average sale of $1,200. Having spent just $300,000 to model and deliver the campaign, the retailer was so satisfied with the results that it has repeatedly used predictive modeling for subsequent campaigns, with continued success.

Go Back