KMA Predictive Modeling

Predictive models help marketers identify their best prospects by using statistical techniques to calculate the probability that a prospect will exhibit a desired behavior. Models vary by type, technique, and/or by the variables or algorithms used to develop them.

KMA provides several types of models, depending on the goals of the marketer, including response models, paid response models, payment or credit risk models, churn/attrition models, high-value-customer (LTV) models, and customer clustering. Several techniques are available, including Good Customer Match Models, Previously Mailed Models, and Mailed Regression Models. We can also develop models that can be used for pre-mailed optimization or suppressions, as well as attrition prevention and reactivation models for client’s house/customer files.

These models are built using our proprietary consumer data resources: CMC (Consumer Marketing Connection); and CHANNELS, our virtual data-shed platform of multi-category, multichannel streams of data live-linked data feeds.

KMA’s analytic staff will work with you on selecting the model that will uncover the most responsive universe for your particular offer and marketing objective. Our process to move forward with any type of model is simple and the only commitment we ask upon completion of the model is a test volume of 50,000 names.

When the model/analysis is complete we provide a report detailing the modeling process, key explanatory variables, gains charts, available name counts, and our recommendations for testing and additional analyses. The modeling process will take about 3 weeks from start to completion, depending on which model scenario is selected.

Basic information needed for modeling purposes:
(varies by type of model)

• Client’s best Customers
• Buyers/Responders
• Mailed Prospect File
• Buyers/Responders and Non-Responders from a Campaign