What should a good manager do if it becomes apparent that a decision that has already been made is clearly not working or solving the situation?
It seems odd that on one hand we would argue for using historic data to build decision models and on the other hand argue that we need to use decision models because the situation is changing from the past. Be that as it may, the risks associated with model building....unusual base data, untested algorithms and, often, the inability of the knowledgeable 'user' to understand which assumptions and factors used by the 'unknowing/separated' programmer (therefore what is missing from the model that must be made up by professional judgement) make me think that it is at best a tool of overall decision making rather than the means.