R. L. Polk & Co.

	8.5 Higher Yield from Polk Predictive Model Situation: In February 2001, an automotive manufacturer was planning a direct mail campaign intended to increase traffic into dealer showrooms. They wanted their campaign to be highly targeted to reach the right audience for their vehicles. Solution: Polk recommended a custom profile Polk Predictive Segmentation Model to score the manufacturer’s internal prospect list. This process identified those individuals who look most like this client’s owner and hence, would most likely visit the dealerships and purchase their vehicles. The flat file provided to Polk for modeling required extensive, detailed processing prior to use, including identification and removal of owners to a separate file, identification of owners of MY 1997-2002 vehicles that were purchased new and still retained, removal of business names and addresses, and merge/purge processing to eliminate duplicates. Each of these households was then matched to Polk’s research database consisting of approximately 110 million households. The match allowed us to append demographic, lifestyle and vehicle information to each record. Custom automotive predictive variables (Polk Garage Predictors) were added as well. Once the file was categorized into two groups, buyers and non-buyers, the groups were compared to determine the key differentiating characteristics. Polk’s predictive model was built and households were scored and ranked based on a combination of their demographics, lifestyles, and vehicle ownership characteristics. The top score group in the model produced a buy rate index of 850, meaning that households in this group were 8.5 times more likely to buy the targeted vehicle than the average household. Results: Increased dealer showroom traffic Incremental sales of 164 vehicles sold that quarter Dealer appreciation for a highly-customized targeted direct mail campaign