The sixth part of the article series by PwC and Dastani Consulting deals with the potential of Artificial Intelligence (AI) in churn management. An AI system can find the customers that indicate a willingness to change. But not every B2B customer churn is painful.
The artificial word “churn” is a fusion of the words “change” and “turn”: the customer changes his behavior and turns away from the previous provider.
An AI system can autonomously find, among all possible criteria, those that actually indicate a willingness to change. Churn management at the data level begins by selecting those customers whose real sales are significantly below the expected value. The expected value is derived from the Customer Value Prediction. In a second step, we evaluate the available sales: if a customer migrates with a probability of 10 percent and we expect sales of 100,000 euros, the risk potential is 10,000 euros. A sales expectation of 20,000 Euro with a fifty percent probability of churn carries the same risk potential, also 10,000 Euro.
The need for action therefore results neither from the probability of migration nor from the current turnover alone, but from the risk potential. It is not worth classifying someone with only 200 euros in sales as at risk and paying special attention to them in the form of expensive customer loyalty measures. It is mainly decisive to address the valuable and endangered customers in such a way that they remain loyal to the company. The success of the applied loyalty strategies can be measured with data analyses. Based on the effectiveness of individual packages of measures, these can then be further optimized.
This article is part of a series on LinkedIn about #PredictiveSales:
1. the potential lies in forecasting
2. technical requirements
3. forecast the purchase probability of potential customers
4. identify cross- and up-selling potentials
5. discover the turnover potential of the customers
6. not every goodbye hurts
7. concrete use in sales