Predictive models are the bread and butter of direct and database marketing. They help marketers focus their efforts and resources on those prospects most likely to take up the product or the offer. These models uncover the best prospects, letting marketers ignore those less likely to take action. With this type of analytics, we will typically use the model algorithm to score an entire prospect population and then select the best contacts that meet the client’s marketing criteria. The modeling is simply identifying prospects that look most like our client’s new customers. Often these models are produced without the benefit of data that identifies a prospect’s interest in the product.
Following development of the predictive models, we compare our predictions to see how much more likely our targeted groups actually purchased the product or service. Consistently we find that over the long term, the predictive models do a good job defining those most likely to act. However, we have seen that prospect scoring alone is not a strong predictor of campaign response. If we can add information about who is currently interested in the product – who is in the consideration cycle – the results jump.
By mixing predictive modeling with approaches that uncover those who are interested in a client’s products we can increase the yield on our marketing efforts. To identify those with a demonstrated interest, we can bring in data about companies who have searched a client’s website, and examine email campaign responses and look for some level of email responsiveness – those who have opened the email and demonstrated some interest.
A web visitor’s IP address can identify which domain they are coming from, and the domain can often identify the prospect company. We look at the number of people from the same IP addresses that are visiting our client’s website and how involved they are in their products. When we notice above-average interest we include them in an appropriate direct marketing campaign.
We also look to the client’s prospect email campaigns to uncover who is interested. Usually with big marketing efforts, multiple people from the same company are in the campaigns. The more contacts we can reach in targeted companies, the more likely we are to achieve a lead and have a sales entry point. We then look to identify the “company’s” interest based on the total responders’ activities within the email campaign. Shown at the left is a typical interest scale by activity.
Depending on the campaign requirements, we normally recommend that those in the “sweet spot” should be touched with more effort. That typically involves some level of tele-contact activity.