Dun and Bradstreet Completes Acquisition of Lattice Engines. Learn More

3 Examples to Help You Understand the Power of Predictive Marketing

3 Examples to Help You Understand the Power of Predictive Marketing

There has been an explosion of interest in predictive marketing and using machine learning to build models to predict buyer intent. When you are considering making an investment in predictive marketing, it is important to remember to set yourself up for success. This means ensuring you think through how the analytic insights will ultimately shape the decisions you make within your marketing and sales strategies and understand what drives those decisions.

All analytical models are all built with some set of assumptions about what they are trying to predict, i.e. which prospects are ready to buy, which customers are likely to attrite, and so on.

Here are three examples to help illustrate the impact of machine learning and predictive marketing to help solve your top challenges.

Example 1: Imagine that you have a large number of prospects that you’d like to engage, but not enough resources to get it accomplished efficiently. Your inside sales or lead development team is short-staffed and doesn’t have enough bandwidth to follow up with each prospect. Most likely, you will want to use a model that can predict high conversion rates for a small percentage of your prospect universe. You’ll want a model to identify the prospects with the highest likelihood of converting to paying customers. In the lift chart below, the prospects noted by the green bar have the highest likelihood of converting.

Example 2: On the other hand, imagine that your business has a strong recurring sales model, with half of your customers purchase in the next sales cycle. For the sake of a simple example, let’s assume that the potential value of these new purchases is distributed from $1.00, $5.00 and $10.00 with an average selling price of $2.00. The best predictive models may give you a segment of customers that renews at 80 percent, which would help you focus your attention on the customers most likely to renew. On the other hand, it may be possible to build a model to uncover which customers are likely to renew with an ASP of $4.00, double the average. In comparing these two cases, the improvement in productivity ranges as follows:

From focusing on the higher conversion rate is 80 percent/50 percent = 1.6x.However, the improvement for focusing on the higher value segment is $4.00/$2.00 = 2x.

Example 3: Imagine your company sells a high value product and has a modest set of inbound leads. Your inbound leads are passed directly to your inside sales team because the amount is reasonable and manageable, and it is the job of the inside sales team to qualify the leads. The qualified leads are then passed to a much higher-cost field sales team. In this example, there is little value in creating a score for lead conversion to filter leads for the inside sales team because it makes sense to follow up on every lead. Instead, it probably makes more sense to have a model to calculate expected lead value. This can be used throughout the sales cycle to make sure that the right effort is applied to both connect with the lead initially, and to stay focused on these leads as they are handed off to the field team. You may also consider having a separate model for leads once they reach the opportunity stage.

Ultimately applying predictive analytics to your marketing efforts can have a tremendous impact on your conversion and revenue goals. For a deeper dive on common modeling techniques, read more here: Not All Predictive Models are Created Equal.

Written by

Paolo Massimi
May 6, 2014

Comments