What’s The Correct Business Math For Decisioning Purposes?
Determining the best business math is another important element that is too often
overlooked. As most business executives don’t typically model regressions or
probabilities themselves, the key is to connect the outputs from technical models
to business decision-making processes. This connection can and should be made
in the design phase and not during implementation. The output of this step of the
methodology is a compelling business case which answers the question: “How does
this insight make money”.
The logic of the model must be consistent with business conditions and current
operations. Line and field personnel should be consulted to reach an agreement
on changes to decision making that are both feasible and practical. They should
also help identify where new data and insights can add value in, for instance, highly
iterative underwriting processes. Insights must be actionable – not just academic. If
a modeling result doesn’t make sense or can’t easily explain how current decisioning
processes should be modified, then there may be cause to re-evaluate the models.
Note that this design step may not result in a single outcome; there will most likely
be options for such business changes. The modeling activity will not provide these
options; rather, it’s up to the managers of the business to develop and evaluate the
options for implementing the analytics insights.
No comments:
Post a Comment