PrefSolve™:
The Core Decision Support Tool that Undergirds Our Strategic Approach
to a Wide Range of Business
Issues.
Our
Approach to Customer Retention
Retaining customers requires three critical success
factors. First, the datamart needs to include better predictors of
attrition. For instance, it is common knowledge that in many industries
customers switch to competitors because of a series of negative service
incidents. Not coincidentially, these same service incidents are precisely
those not captured in most enterprise data-stores. Second, there is
a marked association between attrition and value-based segmentation.
These value-based segmentation markers must also be added to the datamart.
Third, high performance predictive outcomes require that truly exhaustive
multivariate customer retention modeling be performed. We will address
each of these issues in turn.
First, adding service history performance to the
datamart involves capturing "service touches." That is,
the goal of the truly customer-centric firm is to capture any relevant
contact with the customer, and place it onto the datamart. The steps
to add service touches to the datamart are relatively well understood
and will not be discussed here.
The remaining two success factors (i.e. adding
value-based segments to the datamart and performing high performance
data mining) are made possible by three proprietary protocols developed
at Decision Support Sciences. Each of these protocols can be further understood
by clicking on its respective link.
Adding value-based segments to the datamart requires
two sub-steps: capturing the "hot buttons" of customers
using Value-Based Segmentation and generalizing those scores to the
entire customer datamart using Attitudinal
Imputation Modeling (AIM)SM.
Last, once these critical "missing pieces" have been added
to the datamart, customer attrition models can be quickly developed
using Ultra High
Performance Data MiningSM.
