Change is always happening in business - sometimes those changes are more extreme and widespread - like now. As shopper habits change and as markets evolve your past demand planning models quickly become less relevant.
Could you imagine going to the archives and dusting off the analysis that was done 3 months ago and using that to plan today? It would be like using the 2018 form guide to pick a horse in the 2020 Kentucky derby.
Unfortunately - this is what often happens in business. A business may develop a model internally or outsource the process to a 3rd party - like a consulting firm. The focus of the work will be on developing the perfect model that represents the present, but the development process fails to recognize that things change. As days pass the model quickly becomes less than perfect.
Good modeling is not about having a model that is perfect at only one particular point in time - it's about having a good model and more importantly, a methodology that is agile, efficient, adaptable, and repeatable. You need something that allows you to address the next challenge, or recognizes the next opportunity. In an environment where change is constant, optimal planning is cognizant of this change.
Winston Churchill is quoted as saying "Perfection is the enemy of progress". This is true of statistical modeling in a number of ways ;
Firstly - developing the perfect model takes up precious time. You can create a good model quickly (a very good model very quickly with the right tools). Developing the perfect model takes time and energy. Time and energy you can't afford.
Secondly - the perfect model typically means complexity and complexity breeds confusion, resulting in reduced acceptance and curtailing progress.
Thirdly - the perfect model won't adapt to progress. It has been designed to meet the needs of a time now past.
So what's the key - find a modeling approach that recognizes that time is a resource not to be wasted. Speak to Cauself.
Comentários