Syndicated from ebizQ
Robert Grossman had a nice post – Five Common Mistakes in Analytic Projects that made me think about the role of decision management in putting predictive analytics to work. Of Robert’s 5 mistakes, 3 are directly addressed by decision management:
- There is not a good plan for deploying the model
Because decision management is focused on applying models in operational systems and because it combines a focus on business rules to automate decisions with analytics to make them smarter, there is always a plan for deploying the model with decision management.
- Working backwards, instead of starting with an analytic strategy
Decision management always begins with the decision in mind – focusing on the desired outcome, using that to drive the definition of the analytic project and thus the data to be integrated etc.
- The predictions of the model are not actionable
Decision management provides a framework for doing exactly what Robert suggests – understanding the measures for a decision (part of decision discovery), defining the actions to be taken – the rules (part of decision service development) and analyzing the results to understand what works and what does not (decision analysis).
Analytic projects can still go wrong, even when conducted in a decision management framework but decision management can help maximize the value and minimize the risk of your analytic projects.