Should you factor returns in your forecast error calculation? In this article, we’ll use a sample data set, to demonstrate if you should consider returns when calculating your forecast errors.
To set the foundation for this discussion, let us first look at the definition of order lead time. Order lead time is the time gap between the date when a customer places an order and when they expect to receive the product. Typically, in a B2B environment, the expectation is that there will be some gap between the two dates, and in many cases, this gap can be negotiated.
Imagine a demand planner working with 10,000 unique combinations. One of the not so envious tasks for this person would be to generate statistical forecast for all these combinations. These days, the statistical forecasting tools available on the market can forecast these combinations using a list of forecasting methods and figure out which method works best for a particular combination.
An open order is defined as an order placed by the customer which is under process and is yet to be fulfilled by the supplier. For effective analysis, open order data needs to be recorded daily in an ERP system. A minimum of twelve to eighteen months of open order history is required for your sales forecasting analysis and fine-tuning process. A shorter period could render unreliable and skew the results of your data analysis.
In an ideal world, demand and supply would be steady and predictable, resulting in optimal capacity utilization and no back orders or missed customer orders. However, in the actual supply chain world variations in actual sales vs. projected sales result in lower forecast accuracy, and either overstocking or stock out situations.
Demand sensing helps us identify the actual customer order trends and helps us improve the near-term forecast. Strategic actions like demand shaping for knowingly increasing or decreasing the demand for the product can be undertaken by sensing demand signals.
Over the past few months, we’ve been running simulation tests on different demand forecasting methods: Winter’s additive & multiplicative, seasonal and robust seasonal. Then, we used MAPE to determine the forecast accuracy for each method. Here’s what we found.
Seasonal method is a regression method that fits a linear trend along with sine and cosine curves. These sine and cosine portions of the regression can fit any seasonal deviations from the linear trend. Robust seasonal method also fits a trend along with sine and cosine curves, however this method uses linear programming to fit a seasonal series in a way that compared to the regular seasonal method is less likely to be thrown off by noisy values that depart from the trend or seasonality.
While major supply chain consequences aren’t usually life threatening, you can bet they have a huge impact on you, your organization, and your customers. Your goal is to prevent problems early but it doesn’t always happen.
For most businesses that rely on demand forecasts for supply and capacity planning, improving demand forecast accuracy is critical. There are many methods to measure forecast bias and the accuracy of supply chain forecasts including using statistical methods like the Mean Absolute Percent Error or MAPE that we’ve discussed in our previous blogs.