Understanding the inventory forecasting Excel report

Understanding the inventory forecasting Excel report

Home » Inventory Forecasting » Here

This page documents the Excel inventory forecast report as typically produced by Lokad. While Lokad can also exports its forecasts as flat files, the Excel report is typically easier to navigate, especially when the number of items (i.e. products, SKUs, barcodes, depending on the applicable terminology) is not too large. The Excel report contains both data directly copied from the original data source and the forecasts generated by the Forecasting Engine of the Salescast.


Download sample Excel report: salescast-northwind-sample.xlsx

Big Picture

Lokad starts a project run by retrieving the data from files stored in your account. Then, quantile and/or classic forecasts are produced. Finally Lokad creates a consolidated Excel forecasting report gathering key inventory optimization metrics.

This page documents a typical inventory forecasting Excel report produced by Lokad. More details are provided below for each section. Depending on the information present in your IT systems, your mileage may vary. Salescast is flexible and can produce reports with extra columns and eventually less. We do not cover those aspects here.

Context data

The data columns, in gray in the schema here above, represents the context, that is to say all the information directly imported from the original data without any statistical processing involved.

a) Item ID: identifier for each product / SKU / barcode, depending on the applicable terminology.

b) Item Name: human-readable entry for the item. This field is not required, however it illustrates the capability of Lokad to insert extra fields in the Excel report for pure reporting purposes.

c) Service Level: desired probability of not hitting a shortage. This value represents the performance goal in term of stock availability. Read more on setting the right service level. This value is provided as input to Lokad.

d) Lead Time: delay expressed in calendar days that represent the sum of the reordering delay plus the supply delay. This value typically depends on your supplier. Read more on calculating lead times. This value is provided as input to Lokad.

e) Stock On Hand, Stock On Orders: number of units readily available, resp. number of units already reordered.

f) History: aggregated sales (per day, week or month) over the last 12 periods. Those values are provided to facilitate quick validation of the sales forecasts. History not displayed is still used while computing the forecasts (both classic and quantiles).

Forecasts and optimization metrics

The forecasts are produced by Lokad. Based on those forecasts, we also infers a few inventory optimization metrics. Those are represented over an orange background in the schema here above.

1) Reorder Point: The number of units that should trigger a replenishment order when the stock on hand gets strictly lower than the reorder point value. This value is calculated by Lokad, if quantile forecasts are active.

2) Order Quantity: The number of units that should be reordered. If zero then Lokad suggests that the item should not be reordered at this point of time. This value is defined by the relationship ReorderQuantity = ReorderPoint - StockOnHand - StockOnOrder (minimal value is zero). If LotMultiplier is also provided, Lokad will round up OrderQuantity to be a multiple of it.

3) Lead Demand (short-hand for lead time demand): the number of units that will be sold during the lead time. This value is inferred from the classic forecasts combined with the lead time.

4) Stock Cover: the number of days left before stock-out if no reorder is made. This value is inferred from the classic forecasts and the stock on hand.

5) Accuracy (followed by Classic Forecasts): The expected accuracy of the forecasts delivered by Lokad, which can be enabled in the project settings. This value is a percentage, and included between zero and one. The higher the value the more accurate forecasts are. We have the following relationship Accuracy = 1 - MAPE, where MAPE is the Mean Absolute Percentage Error. The expected accuracy is computed by Lokad - like the forecasts themselves - through advanced statistical analysis. Check Measuring forecast accuracy for more details.

6) Classic Forecasts: The classic demand forecasts produced by Lokad at the daily, weekly or monthly level. These classic forecasts can be absent if the quantile-only option is selected in Lokad. Check our Forecasting Technology page for more details.