Predictive Customer Value

The Predictive Customer Value module is an enhancement to our Core Offering. It surfaces fields not available in our base platform that are driven by proprietary data science models that allow WhatCounts to predict, with high confidence, key insights that will help you improve your retention marketing success rate. This upgrade to our Core Offering is available to qualified senders. The ideal candidates have 100,000+ unique customers. For more information about this module, click here or talk to your Client Success Manager.

With the purchase of the module, the fields below will be available in the WhatCounts Segment Builder, where it can be applied to automators, pushed to your ESP, or exported for use in other channels, such as display retargeting, Google Customer Match, Facebook Audience Match, and direct mail and sales call campaigns.

Glossary of Fields 
  1. Predicted Churn Probability: The likelihood a customer has already churned
  2. Predicted Next Order Date: The date we predict a customer to make their next purchase
  3. Predicted Future Orders: The number of orders we predict a customer to make over the next 12 months
  4. Predicted Future Spend: The total amount we predict a customer to spend over the next 12 months
  5. Predicted Future AOV: The average order value we predict for a customer over the next 12 months
  6. Predicted Future CLV: The Customer Lifetime Value we predict for a customer taking into account the next 12 months of predicted spending

All "predicted" fields are also accompanied by a quartile field (except Predicted Future Orders, which gets an accompany Rank field). These quartile fields your customer base down into quartiles based on their assigned value for each of those fields. Please note: While a quaritile can often behave similarly to a quarter (25%) they may not always be the same in cases where there is a lot of repetition in your data, as is the case with Predicted Future Orders (a lot of customers have only 1 order), which is why we give it a rank field.

 Use Cases for Predictive Data 


In Automators 

When you purchase the PCV module, your Best Customer and Win Back automators will be improved to include "Predicted Win Back" and "Predicted Best Customer" functionality, so that you can better target your recipients based on this new information. Some ideas include...
  • Target customers with an 80% or higher predicted churn probability for an aggressive Win Back campaign
  • Identify customers with a 50-75% predicted churn probability and append a moderately discounted coupon to get them to purchase again before they churn even further.
  • Send marketing promotions (with WhatCount's targeted product recommendations) to your customers who are in the first quartile of predicted future spend to capitalize on their aggressiveness and early loyalty to your brand.

In Segment Builder 

  • Build a segment of customers with a predicted next order date within the next 6 weeks; export it and upload it to the digital retargeting platform like AdRoll.
  • Build a segment of customers whose predicted next order date has passed and assign them to an offline channel, such as a sales call list or a direct mail campaign.

Frequently Asked Questions about Predictive Customer Value 

Q: Why don't all of my customers have predicted values?
A: WhatCounts can only assign predicted values for customers who have made a purchase in the past two years. Since this may not be true for all of your customers, you will see a discrepancy between your total customer count and the number of customers with a prediction. If someone that did not make a purchase in the past two years does make a purchase in the future, they will be reevaluated and given predicted values.

Was this article helpful?
0 out of 0 found this helpful
Have more questions? Submit a request


Powered by Zendesk