Prequisites before uploading your dataset
To successfully generate a forecast, make sure your dataset adheres to the following.
At least one product_id has a sales history of at least four times the forecast time horizon provided in the outbound_order_line dataset. For example, if the forecast time horizon is 26 weeks, the minimum order data requirement is 26*4 = 104 weeks.
Product_id under the product data entity should not contain any incomplete data (null or empty string) or duplicates.
All the additional columns selected for granularity in the forecast configuration (that are conditionally required ‘) does not contain incomplete data (null or empty string).
The column id across all data entities (for example, product_id, site_id, ship_from_site_id) does not contain special characters, such as asterisk (*) and double quotes (" ").
The order_date does not contain invalid date. For example, 2/29/2023, that is 29th February 2023 is only valid on a leap year.
To improve forecast accuracy, Demand Planning highly recommends the following.
Upload two to three years of outbound order line history as input to generate an accurate forecast. This duration allows the forecasting models to capture your business cycles and ensure a more robust and reliable prediction.
For improved forecast accuracy, it is also recommended to include product attributes such as brand, color, product_group_id, product_introduction_day and discontinue_day in the product data entity.
You can provide additional demand drivers information through the supplementary_time_series data entity. Note, only numerical values are supported.
You provide alternate product mapping when you have similar products or previous version for a new product.
Remove any non-recurring or one-time event such as COVID before uploading the historical sales data.
