There are many different types of regression models on postal addresses - which can each be used in many different ways. Here are just three of the many things you need to consider when using models for acquisition mailings.
1. Discuss the best type of data model and course of action with the modeler. This discussion should include the mailer’s goal for the model, the mailer’s offer and package(s), and whether or not seasonality is a factor. You may want to include more than one campaign (in the case of a mail match model), to include various packages and seasons. Or it may be necessary to build separate models for different packages or seasons.
2. Make sure that there is no upfront charge for the model build and determine what, if any, minimum quantity is required. Most list owners do not charge for the model build but often require that a minimum number of names be taken from the model.
3. Work with a broker who has extensive modeling experience. They can save you a lot of headaches and money and greatly improve your chances of success!
Want even more modeling tips and a guide on the many different types of regression models on postal addresses? Check out our whitepaper: What to Know Before You Model: The 4 Most Commonly Used Acquisition Models.
Previous PostRMI Attending 2014 Washington Nonprofit Conference