A leading US based Discount Apparel Store Chain wanted to introduce a new department in some of its more recent stores. Due to the significant investments involved, the management needed to accurately forecast sales for the department. The challenge was that only Census level demographic data could be used to develop the forecast as these were brand new locations.

• A 2 member AbsolutData team worked closely with the client for a month to deliver the model
• The model was built using a sophisticated regression based technique that simultaneously segmented the market into three groups as well as developed separate models for each market type
• Extensive exploratory analysis helped to identify the potential different market segments. Latent class segmentation technique was applied improving model fit by 25%
• The 3 segments were characterized based on response parameters and demographics
• We also came up with a user friendly Excel based interface that the client team could use to forecast department sales

• The model results have been validated by the actual sales in the new stores
• The model is still being used by the client after more than 2 years