DoubleClick Prospects for New Business
PredictiveMail targets opt-in e-mail consumers using statistically analyzed purchase data for e-mail prospecting or retention offers. Since DoubleClick introduced PredictiveMail earlier this year, the service -- which functions as a members-only alliance -- has grown to include upward of 200 publishers and 5.5 million opt-in e-mail names in the shared database.
DoubleClick describes PredictiveMail as "opt-in e-mail offer prospecting lists," according to Jonathan Shapiro, senior vice president and general manager at Abacus Online.
"What we are determining is, 'Who based on DoubleClick information is most likely to respond to this offer?' " Shapiro said. "We're confident that the ability to build a good statistical model that actually does determine who is the appropriate audience for the offer helps the marketers and helps the consumer, because they're only getting offers that they like."
The statistical models are at the root of PredictiveMail's ability to offer alliance members a supposedly more precise target. The information on which DoubleClick is basing these models is a combination of the e-mail names those members provide, and any relevant online or offline purchasing history that can be determined from customer files and Abacus profiles.
"Once you have a name and an address, there is lots of information out there that can be used for statistical modeling," Shapiro said. Unlike the last time DoubleClick started talking about merging online and offline information -- which resulted in a public relations and regulatory debacle -- this time the New York-based online ad firm is focusing its efforts strictly on consumers who opt in to receive third-party promotions.
The members do not pay to join the PredictiveMail database or to request a statistically driven list. The charges come when the mailing is executed. The current rate for database members is $150 per thousand names, Shapiro said.
DoubleClick is not revealing the names of alliance members, though it claims to send millions of e-mails each month. According to the company, one member -- a high-end apparel cataloger -- began with a small, sample e-mail test. Based on those results, the cataloger followed with a prospect mailing of 700,000 names and then 500,000 e-mail names on top of that.
Shapiro also noted that each mailing provides the user with the opportunity to opt out of future mailings. If the user chooses to opt out, he is removed from the PredictiveMail database but remains on the list of whichever alliance members submitted the name.
"Clearly we want to be very sensitive to privacy issues and ultimately how consumers respond," Shapiro said.