Summary
Forecast Snapshots are copies of your forecast taken at certain times. These snapshots are useful for a retrospective view of what your forecasts were, and when they changed. This allows you to see how accurate your forecasts have been over time, and how their accuracy depends on how far from lead time they are set.
StockIQ will automatically take forecast snapshots:
- At the beginning of each forecast period (e.g. first of the month)
- Whenever a user saves a forecast in Forecast Manager
NOTE: You can change the date your forecast snapshot is taken in the Global Demand Forecast Settings.
These snapshots are displayed under the "Snapshots" tab in Forecast Manager, and in the Forecast Error Screen.
Snapshot Grid Units-Of-Measure
The snapshot grid can be converted to different display units, depending on what you want to see. Use the drop-down at the far right-hand side of the grid to switch between these views:
- Forecast Quantity - Shows the quantity of each forecasted period as it existed in that snapshot.
- Error Units - What are the errors of each forecast period relative to now-known actuals? Negative indicates the forecast was below actuals (under forecast), positive indicates it was above actuals (over forecast)
- Error Percent - What are the percent errors of each period relative to actuals? Again, under indicates that the forecast was low, positive indicates the forecast was high. The percentages are color-coded with the forecast error ranges for low/medium/high/very-high in your Global Demand Forecast Settings.
- Error Value - Converts the error units to currency (e.g. $) to indicate how impactful this error is to the business in economic terms.
Snapshot Grid Columns
In addition to the period-by-period units you can switch between, the first several columns are as follows:
- Fcst Series - Indicates from which forecast series this snapshot was taken. If you have multiple forecast series, you will see multiple entries here. You can use the filters at the top to filter only on the series you care about, such as your operational forecast.
- Snapshot Date - The date the snapshot was taken
- Name - The name of the snapshot. Typically, this is a system-generated snapshot name for the automated snapshots, describing for what node and when the snapshot was taken.
- Avg Err % - Average forecast error over your forecast measurement period (typically 1 year) as specified in your Global Demand Forecast Settings, backwards-looking as of the date of the snapshot. So, if the snapshot was taken in March of 2024, it would be calculated using March 2023 - Feb 2023 periods (assuming 1 year error measurement period)
- Err % vs Current - Shows the error of the snapshot using current date and current actuals as the starting point, so this is a combination of lag and backwards looking error calculation. So, if you have your March 2024 snapshot and today is June 1st. It is calculating error using June 2023 - May 2024.
- Lag Err % - The error from only the lagging periods included in your snapshot, e.g. the periods known now that were not known at the time the snapshot was taken. Using the same example snapshot from above, if you have a March 2024 snapshot, and today is June 2024, your lag periods are March, April, May, 3 periods over which the average error is calculated. Snapshots further back will have more lag periods and therefore more values to include in their Lag Err % calculation.
Troubleshooting
If you are investigating how a forecast may have changed there are some additional useful columns in the column chooser.
- User - Username that caused the snapshot. System is the default snapshot each month and user may have imported a forecast or updated using Forecast Manager > Edit.
- Note - Notes the user provided about the snapshot. Often made when importing forecast.