|
The Smithsonian Custom Modeling file is available for the purpose of improving results on marginal segments and/or to identify new, productive universes for expansion.
Smithsonian offers the following models:
Good Customer Match: This model finds prospects that most closely resemble a mailer's best customers/donors. The model is built by taking a cross-section of a mailer's "best donors/ customers" and matching them up against the Smithsonian file. A minimum of 100,000 best donors/customers should be provided. Regression analysis determines which variables are the most predictive for identifying those Smithsonian subscribers donrs who "look" the most like the mailer's best donors/customers.
Mailed Response: This model uses actual responders from the mailer's cross-section mailing of the Smithsonian file. A large enough cross section from Smithsonian must be tested in order to yield 1,000 (net) responders. The responders are returned along with a copy of the net names mailed file. The characteristics that are most likely to generate positive results are identified and a scoring equation is developed.
Mail-Match: Similar to the Response Model, but instead of mailing an Nth of the Smithsonian file, the mailer forwards a copy of their most recently completed campaign (flagging responders). Using all the data received from all of the lists mailed from that campaign (or multiple campaigns), a model is built.
Upon completion of the model build, a modeling report will be provided outlining the predictive variables, model gains, counts by segment and more. Model builds typically take 4-6 weeks to complete. |
