Direct Mail : Get More for Less

How to leverage saturation mail for localized, targeted campaigns

Another method is to aggregate the data elements for each known individual in each carrier route, then calculate the selected metric's probability for each by applying the coefficients and intercept employed by the individual response model to those independent variable values. In this case, values for each independent variable contained within the mailer's response model would be averaged at a carrier route level and then input to the model's algorithm to determine response probability for each carrier route. The resulting probabilities then would be used to rank-order carrier routes.

Alternatively, data elements for all known individuals in the carrier route can be aggregated, then used as independent variables in a modeling effort whose dependant variable would be the aggregated metric from all individuals mailed in each carrier route. Whereas the previous two methods allow the mailer to employ models already in existence, this third alternative offers the opportunity to customize the application to saturation direct mail.

Support a Local Presence
Because those located within closer proximity of a location are more likely to visit it, differing targeting approaches may be required to optimize results based upon distance from mailing address to local site.

Consider how such a scheme might look in the table on page 29. Note that the table is a fictitious example of the ultimate selection methodology a mailer might use to generate new walk-in business. Determining the ideal targeting structure for each physical location would ideally require that:

• Random saturation and targeted mailings extending to 20 miles from each local site be made in the introductory support efforts for each site. The number of carrier routes to be included in the random group can be used to control the circulation dedicated to saturation exposures. A sample of the remaining addresses can be mailed via traditional Standard Mail means.

• Results from the random mailings will be used for multiple purposes: 1) To quantify the response (or sales rate)/distance function in terms that can be translated into a factor applied to modeled response probabilities; 2) to validate or build saturation mail models specific to the site; and 3) to validate or build targeted mail models specific to the site.

• Armed with the distance factors and saturation and targeted gains curves, the mailer will be positioned to determine the optimal mix of saturation/targeted mail to be employed at a distance/model decile level.

• If things proceed smoothly, a mailer may be able to develop single retail/saturation models that can be applied to all brick-and-mortar outlets, recognizing that the distance factors applied to the modeled probabilities will adjust for differences in behavior attributable to each unique geography.

Despite restrictions applied by the USPS, analytics can be employed to maximize the probability that marketing objectives can be attained over time. The methodologies are somewhat different—but strongly related—to those used to optimize targeted mail campaigns, and can be applied with increasing success as experience is gained with saturational mail techniques.

Joe DeLago is the founder and director of strategy for the full-service direct marketing agency TPG Direct, an Omnicom company. He can be reached at jdelago@tpgphl.com or (215) 592-8381.

• MORE ON
• DIRECT MAIL & POSTAL