Real Estate Site Selection Program

Client: Off-Price Fashion Retailer


  • Identified three predictors of store success
  • Described replicable “sweet spot” markets
  • Simplified site selection process


Our client, a regional department store chain offering moderately priced brand name and private label apparel, accessories, cosmetics and footwear, typically locates its stores in small and midsize towns. The company was interested in expansion and, knowing that store location was crucial to success, sought data analysis assistance to determine the potential profitability of three markets and locations it had identified.


The goal of the real estate site selection program was to answer the question, “What are the best markets for us, and what is the ideal store location in a given market?” The study was designed to:

  • Understand performance of existing stores
  • Predict sales at potential locations
  • Optimize real estate planning to support new store locations


A study was first undertaken to determine the characteristics that correlate with store profitability. Three classifications clearly correlated with success: geographic classification (rural or metro), customer classification (how many trade area customers matched the chain’s best customer profiles) and market classification (what percentage of the trade area population matched the chain’s best customer clusters).


An optimization strategy is essential before you open your next location.

Find A Solution Today


In analyzing the chain’s three potential market locations, the results were clear. Of the three possibilities, only one epitomized the chain’s successful niche in rural-moderate markets. However, further analysis identified a potential problem with that location. If the chain were to locate a new store there, it would share 30% of its best customers with an existing location of the same chain, risking significant “cannibalization” of loyal customers from the neighboring store.

As a further test of the accuracy of the model, Lift361 looked back at the stores the chain had opened the year before to see what the model would have predicted about the chosen locations. Of the 17 locations opened the prior year, the model would have green-lighted only six. Unfortunately for the retailer, the model was correct: while the winning six locations were indeed doing well, averaging a 6% profit, the other 11 locations had lost an average of 10% in year one.