• Each retention decision is for and by a single customer

• Retention decisions must be relayed across channels

• There are lots of rules about retention

• Segmentation and other analytics can make retention more effective

• Retention rates are measured today and improvement is clearly related to better retention decision making

• The Value of Customer
Ideally the future value of the customer, not just the past value. Also the likely future value given plans for new products, pricing should be included.

• Risk of churn
Both how likely this customer to be retained if nothing is done and the overall target retention rate - after all, many companies are judged by retention rate and even retaining unprofitable customers may be necessary.

• Customer's likelihood of accepting various offers
Many different offers can be considered and each customer or customer segment will prefer some over others

• Regulations and Policies

• Automate Decision
Initially we have different channels and our approach to retention is probably different in each.
The first step, then, is to take control of the decision so we can make it consistently across channels.

• Apply rules
We should also use rules to describe it so that the decision can be automated correctly and managed by business staff, not IT.

• Close the Loop
Before we go any further we should ensure we can track the results of our retention decisions and see how well things are going, how they change, what the impact of changes is and so on.

• Segment customers
Not all customers are the same so we should analyze them and segment them so we can retain them differently depending on what is going to work. This is not going to stay static so we need to make sure we can keep analyzing and improving our segmentation.

• Adaptive Control
We are likely to have multiple opinions as to what might work and different potential segmentation schemes so we should institute adaptive control so that we can test our "champion" approach against some "challengers" to see if any of the challengers is more effective. This needs to be part of our ongoing improvement cycle.

• Predict risk, value
Segmenting based only on the data we have is interesting but it would be more useful if we could also use predictions as to their risk of leaving, lifetime value of them etc as part of our decision. Back to the data, then, to build predictive insights.

• Optimize decision
Working to continually improve the outcomes and ultimately to build models of the trade-offs that can be optimized mathematically and we end up with an optimized decision.