Written by Ruvisha Pillay and Mark Whiteacre

Forecast risk and offset are a measure of the variance between sales history and forecast history. On an item’s Inquiry screen, information regarding forecast risk can be found in the Safety stock panel. This is because forecast risk is one of the inputs that affect safety stock, and this article aims to help you understand how.

If the days offset is positive (as in the example below), there is a tendency for the item to be under-forecasted. Days are therefore added to the safety stock. This is to reduce the risk on that item as a result of selling more units than forecasted. If the days offset is negative, this means that the item has a tendency to be over-forecasted, therefore days are removed from the safety stock.

Under-forecasting → sales exceeds forecast → stock out risk

Positive offset days → adds to safety stock

Over-forecasting → forecast exceeds sales → excess stock risk

Negative offset days → reduces safety stock

The risk is derived from the Forecast history tab. The average of the historical forecast attempts over the cover forward period is used to compare with the actual sales for the month in order to assess forecast risk. Cover forward is the number of days you have when you sum the lead time, safety stock and replenishment cycle days, also known as “order up to” days. It represents a period of time over which forecasts will have an impact on the immediate order recommendation. Any forecasts beyond the cover forward are therefore less important to measure.

The sales and forecast comparison incorporates "upstream" demand for dependent locations and related finished goods. Specifically:

Including upstream demand ensures an appropriate risk and offset is computed. This is especially important for supplying locations and raw materials which have no direct sales of their own.

Olive line → forecast

Purple line → sales

Gray shaded area → forecast attempts

The table below shows a variance of the average of the forecast attempts within a certain time frame compared with sales history. In the example, the data points depicted in green are used (added together, divided by 6) to get the average forecast of 10 for May 2021.

A negative figure in the variance row shows that the forecast is less than sales. That represents a risk of stocking out. A standard deviation is applied which calculates the under-forecast in units, which is then converted to a risk percentage.

The higher the forecast risk % → the higher the safety stock.

The offset is the difference between the average sales and average forecast expressed in days. The offset is also impacted by the future weight setting. Future weight is designed to allow safety stock to react to how the current forecast compares with sales history. For example, if our current forecast is double what has been sold previously (actual sales), we are potentially over-forecasting and effectively building safety stock into the forecast itself. In this scenario, the future weight will reduce the offset (and consequently safety stock) to compensate. The expected range of values for the future weight setting is from 0 (no impact) to 2 (exclusive impact).

The higher the forecast risk offset days → the higher the safety stock.

The minimum and maximum settings will constrain the risk and offset to a reasonable range. The Risk limits panel below shows an allowed risk range. For example, if the calculation results in a value above the maximum, the risk percentage would get capped at the maximum allowed value.

Settings → Configuration → Risk limits

Forecast risk will be calculated only for items with an age greater than 6 months. This age check will incorporate the maximum age found in upstream demand items. Items with less than 6 months of history will attract the default risk and offset specified in settings. This can be set up in the Risk defaults panel.

Settings → Configuration → Risk default

Forecast risk is calculated by using factors such as sales history and historic forecasts. Parameters may be set (in the “Risk limits” panel described above) to cap the permissible range of the risk value, which protects you from holding too much safety stock. The higher the risk, the greater the level of safety stock needed.