Both constrained and unconstrained forecasting can be useful when predicting demand for a given period. However, each type is more suited to different applications.
For example, because unconstrained forecasting focuses more on actual demand, you can use it for applications like:
- Determining product launch feasibility.
- Identifying gaps between your potential and actual demand.
- Showing the effectiveness of your existing marketing strategies.
On the other hand, constrained forecasting is better for applications intended to optimize operations, such as:
- Understanding customer preferences and behaviors.
- Predicting total revenue for a given sales period.
- Revealing potential areas for budgeting improvements.
That said, you don’t have to settle for one or the other. You can also use both forecasting methods and compare your results to get a complete view of your potential demand.
How to Calculate Forecast Error
There are several forecast error formulas that inventory planners use to make their decisions. Some are simple, while others are pretty complex. We will discuss two more common forecast accuracy calculations: the MAPE and the MAD.
No. 1: MAPE Formula
MAPE stands for Mean Absolute Percent Error. This formula comprises two parts: one to retrieve the mean (M) and one to retrieve the absolute percent error (APE).
The MAPE formula is:
(Actual Demand – Forecasted Demand) / Actual Demand) x 100
With this formula, you can calculate the mean of all percentage errors that occur during a given period. MAPE is a measure of error, so a lower percentage should be the desired outcome.
No. 2: MAD Formula
MAD stands for Mean Absolute Deviation. The MAD Formula shows how forecasted demand differs from actual demand, measured in units.
For example, if the actual demand is 90, but your demand forecast is 95, your absolute error (or deviation) is 5. Calculate the deviations for multiple periods and then take the mean of those values to produce your error forecast.
Beyond the MAPE and MAD formulas, you can use many forecast error calculations. Regardless of which you choose to use, forecast error is critical to your business’s health, so be sure to use at least one!
The Benefits of Improving Forecast Accuracy
Forecast error data helps inventory planners perfect inventory purchasing. Here’s how:
- Reduce risk of future forecast errors: Your forecast error calculations provide a numerical estimate of past forecast quality, which can help you factor risk into future projections. This data reveals the necessary adjustments to your inventory management process, such as increasing the stock buffer or adjusting reorder points to cover gaps.
- Prioritize questionable forecasts: If you find a high forecast error for one of your products, you will want to pay special attention to that product. Closely monitor these products, look at their future demand, and adjust stock levels as needed. Forecast error values pinpoint which products need such attention.
- Refine and improve forecast accuracy: If you consistently realize high forecast error rates, there is an underlying problem. You may need to update your forecasting technique to alleviate the issue and ensure the accuracy of future forecasts.
Although manual calculations can lead to errors, you can further improve your accuracy with artificial intelligence-powered demand forecasting software. Even in the face of economic fluctuations, having an intelligent platform to support your supply chain planning can help your business stay afloat.
