Is the road between forecasting and planning a two-way street?
“Should we work on our forecasts before investing in a planning or scheduling solution?” Time and time again I hear this question from companies starting their advanced analytics journey, regardless of their industry or functional area.
After a company has put its data in order and has already completed a descriptive analytics initiative, what is the next step? Increasingly, managers are seeing the value of using advanced analytics (1) to predict key variables, such as market demand (i.e., predictive analytics), and (2) to identify a preferred course of action, such as by deciding what product should be produced when and where (i.e., prescriptive analytics).
But this question is a natural one, for everyone understands the one-way causality between forecasts and plans: Good, sharp forecasts can indeed lead to plans that anticipate the future and thus use resources efficiently. Thus, many managers reason, investing in increasingly accurate forecasts can produce quicker and larger return on investment. But a key assumption behind this reasoning may not always be true. Can a “better” forecast be transformed into “better” action by an existing planning or scheduling process and system?
So whenever someone asks me this question, I conduct a little exercise. First I ask the stakeholders in the room to agree on a single product to serve as the focal point of the exercise. Then I ask them to assume that their brand-new forecasting solution predicts a 30 percent increase in demand in North America for the next quarter. Next, I ask each individual to write down, in only a sentence or two, a preferred course of action for dealing with this demand spike. But when stakeholders reveal their answers, I can usually count as many answers as there are people in the room, clearly demonstrating that enhanced forecasts cannot be immediately translated to action or bottom-line savings.
In many departments, planning and scheduling are achieved using manual processes (such as by using Excel spreadsheets) that require the planner to pick an action (in contrast with analytics, which provides a recommendation), that are error-prone, that are time-consuming, that offer no indication of how good a course of action is and that allow very little—or even no—comparison between alternatives. For such departments, investing in a “better” forecasting solution might not bring about the benefits that managers desire.
Conversely, investing in advanced planning and scheduling processes and solutions can bring significant benefits even when the forecast is lacking. At worst, the planning process can be used to create multiple contingency plans in advance or to quickly respond in dynamic environment. Enhanced resource use, timing and allocation often results in tangible savings in a matter of weeks—or even, in many industries, in only days.
To start your advanced analytics journey, explore what IBM Analytics can bring to your organization.