Business Analytics for Sales and Marketing Managers: How to Compete in the Information Age

CREATING SIMPLE WARNING SYSTEMS

If you used the decision tree at the beginning of this chapter (Exhibit 6.1), you were directed to this section because you indicated that you want to start to use churn prediction and integrate it into your business intelligence systems. By doing this, you can make early warning systems that inform the relevant staff in case customers have been exposed to service failures or simply are showing behavior that typically precedes customer churn.

From a process perspective, it is very logical to have the system contact call center agents or sales reps as soon as a customer shows increased churn risk. This is opposed to using monthly call lists or the like; a customer might already have terminated his or her contract by the time retention calls finally comes around.

This pop-up functionality is standard for DWs, which can push these warnings out to specific email accounts in customer service or at the key client manager teams. An example also comes from the telecom industry, where we set an algorithm to scan the CRM systems every fifth minute looking for whether any business customers had made one of a series of inquiries, such as requesting prices for new mobile phones for its employees or asking for a total review of the bills they had paid within the last six months. These types of questions were focused on because churn prediction models had found that these sorts of requests were strong indicators of churn behavior. Therefore, the DW would automatically ...