Focusing on a narrow problem and throwing efforts in it for producing deeper and better insights may seem justified and prove very useful. However, it could be all in vain or a waste of time if a the entity under study is taken out of the context of connected entities and processes. A perfect solution could be found but the big picture could render it useless and put much of the efforts to waste. I never liked too much theorizing so let me lay an example: the manager of newly opened call center requests improvement of their inbound calls forecast. It is based on a sales forecast coming from a third party. The manager request the forecast improvement with the ultimate goal for optimizing the call center costs. Should the analysts jump onto revising and reworking the forecast? I do not think it is a good approach as even a perfect forecast could be waste of time and not lead to desired cost optimization.
A better approach would be to focus on the decision to be made and see all the elements in this case working together. The decision to be made is on the staffing. A good start should be with setting the service levels the company would like to provide for its customers. Putting these service levels together with the expected number of inbound calls would produce the optimal number of staff. Set of scenarios for call load should be explored to analyze the performance under different circumstances and outline the expectations for the accuracy of the inbound calls forecast. It would not come as a huge surprise if these expectations would not call for a very accurate forecast resulting time and money saved for the manager. The initial sales forecast should also be explored to find its confidence levels and accuracy to adjust the calls and staff that depend on it.
My call is for a systems approach in decision making and problem solving. Improvement just of expected number of calls would not necessarily produce the optimized staff the call center needed. Putting together of all factors and criteria together would certainly will certainly do. The systems thinking approach of provides better solutions on problems involving more than one interconnected parts. One of its features I find most useful is it provides ranges of the system's parameters describing that guarantee the decision still holds validity. A good starting point for exploring systems approach could be here. However, as in the case with my client made a fine example that application of this approach could be hampered by organizational, domain-related, political and agency issue.