Logistic Regression for Campaign Management

Campaign management example (using logistic regression)

Recall the campaign management scenario described in Data Mining Services: Overview. Your company wants to improve the effectiveness of its marketing campaigns, with the goals of reducing costs and increasing the percent of positive responses. The results of a previous campaign will be analyzed to determine what factors, if any, can be used to predict the performance of a similar future campaign. Use logistic regression analysis to generate a predictive model. Logistic regression selects the most likely outcome from a set of distinct possibilities.

A recent back-to-school sale campaign produced hundreds of respondents from a pool of thousands of customers. The campaign was based on the following:

•

Age

•

Gender

•

Household count

•

Income range

To predict future campaigns based on the back-to-school sale campaign, you want to use all of these attributes as predictors in the predictive model. Therefore, you must create metrics for each attribute form. Some example metrics for this report are as follows:

 Max([Customer Age Range]@ID) {Customer}

 Max([Customer Gender]@DESC) {Customer}

 Max([Customer Household Count]@DESC)

 {Customer}

The example Tutorial project includes reports, metrics, and other objects created for this campaign management example (search the project for “Campaign Management”). You can use the objects in the Tutorial project to step through the example and determine how it can be applied to your reporting environment.

Use the Training Metric Wizard to design a training metric, following the procedure below.

To create a training metric for logistic regression analysis

This procedure assumes you have already created a Back-to-School Sale Responder metric to use as the dependent metric.

1

In MicroStrategy Developer, select Training Metric Wizard from the Tools menu. The Training Metric Wizard opens on the Introduction page.

To skip the Introduction page when creating training metrics in the future, select the Don’t show this message next timecheck box.

2

Click Next to open the Select Type of Analysis page.

3

Select Logistic regression as the type of analysis.

4

Click Next to open the Select Metrics page.

5

Select Back-to-School Sale Responder as the Dependent Metric.

6

Add the Age Range, Gender, and Household Count metrics to the list of Independent Metrics.

7

Click Next to open the Select Output page.

8

Select the Automatically create on report execution check box.

9

Select Predicted Value.

10

Click Finish to save and create the metric. You can now include the metric in a training metric to create a predictive metric, as described in Creating a predictive model using MicroStrategy.

11

Create a new report with the training metric, Back-to-School Sale Responder metric, and the Customer and Order attributes.

12

Filter the report to include only orders dated during the back-to-school promotional period. For example, you can create a filter that only includes the months of August and September.

13

Execute the report to generate a logistic regression model.

A predictive metric is created in the folder you specified in the Training Metric Wizard. The default location is the My Objects folder.

By adding the predictive metric to a new report along with the Customer attribute and the Back-to-School Sale Responder metric, the accuracy of the prediction is shown to be correct almost 100% of the time; there were only a few incorrect predictions out of thousands of customers.

The predictive metric is ready to be used to target customers who are likely to respond to a future campaign.