Revenue and acquisition costs are important aspects of a direct marketing campaign. Sending promotional materials to fewer people who are more likely to respond brings in the revenue you need and minimizes costs. We know this, but how can you make the most of every mailing?

Different marketing strategies are used to identify prospects likely to respond.  One strategy is to build a regression model and assign a response score to each prospect. Prospects are sorted by their scores and divided into groups. The next step is to select prospects or the group of prospects with the highest response score. What if that next step isn’t this simple? What if prospects most likely to respond to a direct marketing campaign are associated with a higher risk rate, higher acquisition cost or lower revenue? If the answer to any of these questions is “yes”, building a second model to guide prospect selections might be the next best step.

For example, let’s say a store has been in business for over 10 years. They have conducted a number of surveys and collected a lot of information regarding their consumers.  They want to offer a new promotion to boost sales. Instead of offering the promotion to any person walking by, the store uses its existing consumer behavior data to build a model to predict how likely a person is to respond. When customers with the highest scores are identified, the store finds that all of the high score customers didn’t spend the same amount of money during a store visit. Some of them spent up to three times the amount of money spent by others. In this example, selecting customers with the highest response score wasn’t predicative of overall spend, because some of them were associated with a lower revenue. So the store can utilize a second model to predict revenue.

So here’s how the secondary model works – Two scores are assigned to each customer - a response score and a revenue score. Next, customers are arranged in a matrix like the one to the right. Customers with the highest response and revenue scores are in cell 1. Customers with the highest response score and lower revenue score are in cell 2. Customers with the highest revenue score and lower response score are in cell 3 and so on. Customers with the lowest response and revenue scores are in the cell with the bright red X. The store selects customers with the highest response and revenue scores in cells 1, 2, 3 and 4.

Each cell can be treated as a segment. Just like a segment is profiled, a cell can be profiled across attitudes, behaviors and demographics.  The profile can provide a new layer of information to help:

  • Craft an effective message to compel the consumer to pay attention
  • Identify the right channels to communicate the offer
  • Place advertisements on the right channels at the right time to increase awareness and reinforce the message and
  • Assess feasibility of word-of-mouth marketing strategies

Long story short, the matrix helps you:

  • Combine two different models to pick the best prospects
  • Build an attractive message and
  • Simplify a multi-channel marketing strategy by focusing on the best consumer segments

Using these matrix methodologies can do so more than just identifying the right person.