If you need to handle numerous forecasts, you can use our dedicated tool.
This allows you to edit multiple forecasts at the lowest granularity level or in a multi-level mode.
This editor also enables you to have a representation of your forecasts at an aggregate level by selecting the fields for this aggregation.
You can start with a guided tour of this tab.
How to display your data in an aggregate format
Next, it is necessary to create an aggregation at the level that interests you, for example, at the product family level. If you have defined custom fields in your account, you should see them listed in the aggregation settings. You can use any combination of fields to create an aggregation.
Click on Aggregation fields...
Select the field for your aggregation. Here, we choose Family.
Click Confirm to calculate the aggregation.
Now, your data is organized to display your aggregations.
Save an aggregation on the platform
To avoid systematically recalculating an aggregation every time you display it, it is advisable to save it. This way, it will only be updated when necessary.
If your account contains thousands of forecasts, saving will prevent unnecessary delays associated with calculations when you modify the aggregation to be displayed.
In the previous example, we save the Family aggregation by clicking on the icon below.
Next, you need to name the aggregation. In our case, we call it Family.
Filter the displayed data
You can easily filter the data used as input for the calculation of your aggregation. This helps reduce the calculation time for substantial datasets and refine your display based on the elements of interest.
You should use the Filters (funnel-shaped icon) to narrow down this list.
Display configuration
You can switch the display of the table from "units" to "financial data" if you have provided this information for each item. Note that you can currently only modify values when "units" is selected.
You can sort each column by clicking on the corresponding field.
You can see in the blue section of the table your historical sales, and in the red section, the associated forecasts.
To hide or display actual sales (or real demand) in the table, you can use the buttons below.
You can also display the adjusted demand if you have modified historical values, for example, to smooth the effect of a promotion. The adjusted values are highlighted in blue in the table.
The cells highlighted in red represent the forecasts that you have adjusted (differ from the forecast baseline).
You can toggle the display of the table to see either the SKU Science baseline forecasts before your modifications, or the forecasts as you have adjusted them.
How to modify forecast values
By clicking on the red cells, you can edit forecast values and adjust them in two ways
By directly entering the new value using the keyboard and confirming the input.
By entering a percentage, for example, "120%" to increase the initial value by 20%.
Note that if, after one or more value changes, you delete the current value and confirm with the "Enter" key on your keyboard, you will revert to the initial value calculated by the platform (the baseline forecast value). The cell will return to a normal display, as if it had never been modified.
One way to return to the baseline data at all levels is to clear all adjustments by clicking on the rounded arrow at the beginning of the row you are working on. However, you will lose adjustments for all periods on that line.
Operation and Impacts of Distribution on Underlying Forecasts
When you modify a forecast cell, the platform distributes the new value as follows
The weight of each underlying forecast for that period is calculated. By underlying forecast, we mean the base value computed by SKU Science, potentially replaced by an adjustment if one exists at the forecast level.
Each underlying forecast retrieves the percentage of the input value in the aggregated cell, based on its weight in the total of the previously displayed forecast. This value is rounded.
Due to entered values and rounding, this distribution may not align perfectly with the underlying data it comprises. Additionally, the remaining distribution is added to the SKU with the highest weight in the distribution. If this method does not fully satisfy you, you can manually distribute this remainder among the SKUs of interest.
Here is an example to better understand
Family 1 includes 4 items. For the month of June, the values are 57, 469, 463, 28.
At this stage, the values are the base values, thus unmodified.
By aggregating by family, you indeed directly obtain a total of 1,017 in the June column.
By entering 130% in the location of the 1,017 and validating the input with the "Enter" key, you obtain 1 322 in the cell. Now, the cell is highlighted in red.
If we go back to the level of the 4 items, here's what we observe:
Each June value has indeed been increased by 30%, and the total does amount to 1 322.
If we now replace the 74 with 80 for item 022-1 and aggregate by family again, we obtain 1 328, which is 6 more than previously.
Now, it's your turn to play. 😀
You can use the demo base and make the modifications you want. You can always go back to the initial demo by clicking on Restore Demo.
Feel free to share your feedback or suggestions directly within the tool's interface.