Predicting Direct Marketing Results

Everyone wants to know what the future has in store for us. Most people just wait and see. But in direct marketing, we can often accurately predict the response rate, based on past customer behavior. The result can be greater efficiency of marketing dollars.

Our company has been maintaining a customer marketing database and providing marketing support to a major truck manufacturer for several years. As a result we have built up a wealth of detail from past marketing efforts. We use this information to predict what will happen when we do the next promotion to prospects and customers. Last year, we did a promotion to 5,644 truck prospects and customers. Before we began, we used a model to predict our results.

The methodology for the model was based on these variables:

• Employee size of business
• Vocation: eg landscaping, electrical contractor, towing, etc.
• Time since last purchase
• Total number of trucks owned

Using the model scores, we were able to divide the prospects into three groups based on our estimates of their purchasing a new truck:

• High probability
• Medium probability
• Low probability

While we were convinced that our method was accurate, we still had to convince the manufacturer to use the model scores to guide his marketing efforts. The test promotion to 5,644 prospects was used to prove the validity of our efforts. There was a lot riding on the outcome of this test both for us and for the manufacturer.

The trucks cost an average of $33,000 each. The truck manufacturer was hoping to convert 1.5% of the 5,644 prospects into truck buyers or 85 trucks. We thought that we could do better, using the results of our model. To sell the trucks, we developed a creative package that explained new truck features that had been developed in 1999, but had not been advertised widely. Using the scoring system, we sent an offer, with an incentive plus a survey to three equal groups plus a control group. Of the four groups, the control group was probably the most important. It represented truck sales which occur in the absence of any direct marketing offer. The 520 companies in the control group were selected equally from those prospects scoring high, medium and low on the sales probability model. The use of a control group answers the perennial naysayer, who says, “These companies would have bought trucks anyway without the direct mail. You wasted our money.” As you can see, the chart shows that without encouragement, one half of one percent of our selected prospects would have purchased a truck. All of our three groups did better than that.

Post Footer automatically generated by Add Post Footer Plugin for wordpress.