Direct marketers today use modeling as a tool to not only maximize customer file performance but also to optimize the selection process on large prospect files where traditional RFM selects may not be yielding adequate results. There are a number of different techniques that fall under the general category of predictive models.

As the name suggests, predictive models are used by direct marketers to try to predict future behavior. While there are several different types of models (good customer match, mail/responder, regression models) all generally provide direct marketers with the following:

1. A model equation that looks at groups of characteristics based on the data available to the modeler. A score is developed for each record which reflects both positive and negative characteristics and is used for the purpose of predicting behavior such as the likelihood to respond to a direct mail offer.

2. A ranked customer or prospect file divided into segments, with each segment having a different predicted behavior. The most common distribution divides the modeled universe into 10% subsets or deciles. The model helps direct marketers identify high and low performing decile segments. This is usually reflected in what the modeler presents as the gains chart.

A typical gains chart might look something like this:

Decile	Number of Individuals	Number of Responses	Decile Response Rate (%)	Cumulative Response Rate (%)	Cumulative Lift Index
Top	204974	14974	7.3	7.3	296
2	204982	7306	3.6	5.4	220
3	204972	5985	2.9	4.6	186
4	204989	4824	2.4	4	163
5	204967	3814	1.9	3.6	146
6	204985	3243	1.6	3.3	132
7	204975	2739	1.3	3	121
8	204973	3098	1.5	2.8	114
9	204996	2285	1.1	2.6	106
10	204979	2344	1.1	2.5	100
	2049792	50612

When making a decision, one of the key factors is the “Index” column. In this example, the index is calculated by dividing the decile’s response rate by the overall response rate, thus providing the direct marketer with a method to quantify performance.

Today, sophisticated direct marketers are looking to modeling solutions to help in this challenging business environment. Whether developing a customer-based model to reactivate lapsed house names or utilizing prospect modeling to open up new universes, many direct marketers don’t have the resources, technology, and statistical knowledge to explore this valuable marketing tool. LDSGroup, through our considerable experience in this field and our network of data relationship partners, can objectively find the right modeling solution.