Quick Summary: Monthly forecast shots provide snapshots of generated forecasts to help identify over- or under-forecasting patterns. Review performance against actual sales to understand impacts on forecast risk, offset, and safety stock behavior.
What Are Forecast Shots?
A forecast shot is a saved snapshot of the forecast generated at a specific point in time (usually during the monthly forecast run).
Each new forecast replaces the active forecast, but the previous shots are retained for history and analysis.
Forecast shots allow you to compare what the system predicted with what actually happened.
Think of forecast shots like “forecast versions” that build a historical record of accuracy.
Forecast Shots vs. Forecast History
It's important to differentiate between these two terms:
Forecast shots are historical snapshots of the forecast at a specific point in time. For example, the forecast that was made in August 2023 for the month of April 2024 is a single "shot." These snapshots are what the system uses to calculate the average forecast that drove your past orders.
Forecast history is a calculated value. It is the average of the forecast shots over the Cover Forward period. This value is what you see in the Forecast history tab and represents the forecast that was driving your order recommendations in a given month.
When Are Forecast Shots & Forecast History Is Generated?
A new forecast "shot" is generated and saved whenever the forecast for an item is recalculated. Similar to shots, Forecast history is recalculated whenever forecasts are regenerated. This happens:
Automatically for all items at the start of each new month.
When Regenerate forecasts is triggered from Data Refresh or Settings > Configuration > Forecast.
It is also recomputed during regular daily batches for:
Newly downloaded items.
Items with changed supersessions.
Items whose classifications change to stocked.
Items with age <= 4, when the Configuration setting: Forecast Young Items Daily is enabled.
The number of forecast shots included in the average depends on the Cover Forward period, so values may change as that period shifts.
Forecast history also incorporates frozen and macro-adjusted forecasts to ensure that historical records accurately reflect all forecast decisions.
Forecast History vs. Sales History
Don't confuse your forecast history with your sales history.
Sales history is a record of what you actually sold in the past.
Forecast history is a record of what the app predicted you would sell.
Comparing these values is the foundation for variance analysis and for calculating forecast risk and offset.
The Role of the Cover Forward Period
Forecast shots are most important within the cover forward horizon:
Cover forward = Lead Time (LT) + Safety Stock (SS) + Replenishment Cycle (RC).
This is the period where forecasts directly affect replenishment recommendations.
Forecasts outside this window are useful for planning but less critical for immediate order decisions.
➜ For more on this topic, read: Cover Forward Period Explained
The Forecast History Tab
The Forecast History tab on the Enquiry screen shows how past forecasts compared with actual sales.
Olive line = Forecast.
Purple line = Actual sales.
Gray shading = Historical forecast attempts (shots).
Snowflake icon = Months where forecasts were manually changed and frozen.
By reviewing this history, you can see whether the system consistently under-forecasted (sales > forecast) or over-forecasted (forecast > sales).
Reading the Forecast History Table
The Forecast History table, located in the Forecast history tab on an item's Inquiry screen, provides a powerful view into past performance.
Columns: Each column represents a forecast month.
Rows: Each row represents a forecast attempt over time (a "shot").
Cover Fwd row: This row is especially important, as it shows the duration of the Cover Forward period (rounded to the nearest month) used for the forecast in each month.
Variance: This column shows the difference between the forecast and the actual sales for a given month.
Positive Variance: Indicates over-forecasting (forecast was higher than actual sales). This would have led to a larger order quantity and an increase in safety stock.
Negative Variance: Indicates under-forecasting (forecast was lower than actual sales). This could have resulted in a smaller order quantity, potentially leading to a stockout.
Red Snowflake: A red snowflake icon in the table is a crucial visual cue. It indicates that the forecast for that specific month was manually changed and frozen.
How Variance Is Calculated
For each period, the system compares the average of the forecast shots within the cover forward horizon, known as the forecast history, against actual sales.
Variance units = Forecast – Sales.
Variance % = Variance units ÷ the lower of forecast units or sales units.
Negative variance = under-forecast (stockout risk).
Positive variance = over-forecast (excess risk).
This information feeds into forecast risk and offset calculations, which then influence safety stock.
⚠️ Watchouts
Scope of Data: Forecast history reflects the final forecasts used at the time, including both system-generated and manually adjusted (frozen) forecasts. A snowflake icon indicates where manual adjustments were made.
💡 Tips
Spot Bias Early: Use the variance rows in the Forecast History table to identify consistent under- or over-forecasting before it affects replenishment.
Forget about these 👇 😞 😐 😃 Have your say here!