Applying “Collaborative Information” to Your Forecast

by | Dec 15, 2023 | Best Practices, Forecasting, Technology

Applying “Collaborative Information” to Your Forecast

By Jon & Matt Schreibfeder

Most demand planning systems base their forecasts of future sales on past usage history. This is known as “time series forecasting”. The idea is that what you sold in the past is a good indicator of what you will sell in the future. Unfortunately, this doesn’t always work. Situations change. For example:

    • You have an unexpected increase or decrease in sales or usage of an existing stock item.
    • A new customer projects they will buy large quantities of specific products.
    • You open new stocking locations or close some existing ones.
    • You introduce new items and discontinue other products.
    • You run promotions in an attempt to increase sales.
    • Events or other factors outside of your control that will affect future demand of specific items.

A good forecasting/replenishment system will alert a buyer or planner if actual sales or usage in the month or week just ended differs significantly from the forecasted quantity. You must determine what caused this spike (or trough) in usage. To accurately forecast future demand of the effected products, you must apply “collaborative” estimates to the forecast. This is information obtained from salespeople, customers, and other sources to determine the reason for the spike or trough in usage. Together with your salespeople, examine the transactions that comprise this unusual amount. Situations you may encounter include:

  1. One customer is buying a lot more or less of the product. Contact the customer if this unusual usage was due to a one time only situation, such as an unusual project. Or will they continue to purchase a lot more of less of the product:
    One time situation – Adjust usage to ignore this unusual activity. The adjusted usage quantity should reflect what you would normally sell.
    Start of a new trend – Adjust usage in each of the past 12 months to reflect this new activity. Make it reflect the customer’s new demand for the item. This will cause your system to properly forecast future demand.
  2. The unusual quantity is primarily sold to your current customers. This is normally the start of a new trend (see above).
  3. There are a lot of “typical quantity” sold to unfamiliar customers. Unless you are running a promotion or just decreased your price of the product, your competition is probably out of the item. Adjust usage as though this is a one-time situation (see above). We will discuss adjusting usage for promotions in a future newsletter.
  4. There are unusually large quantities sold to unfamiliar customers. The product is probably in short supply due to a supply chain disruption. Adjust usage for this one-time situation but be sure to reserve your remaining stock for your best customers.
  5. There is an unusually small usage quantity because you were out of stock. You must adjust the usage quantity to reflect what you would have sold had adequate stock.

If the unusual usage is due to a temporary situation, best practice is to adjust usage to equal the forecasted quantity. This is the amount you would have sold under “normal” circumstances. How do you know it is a temporary situation? Ask the appropriate salesperson or customer. On the other hand, if you suspect that it really isn’t unusual, talking with salespeople or customers will help you determine the new demand for the product. Salespeople and customers should have been aware of what is going on. Use their knowledge!