Intuendi manages your data applying a hierarchy where a category contains more products which sell over multiple regions or channels. Hence, a product can belong to a unique category, but a region can belong to multiple products.
If you’re not a Shopify merchant, you have to generate the import files described in https://intuendi.com/data-import/. Once done, you can import via API (see the data import page) or using the application.
The frequency of your sales history and of your forecasting data represents how your data will be aggregated during the import phase. If you choose a monthly frequency, you will find your history data organized by months and your forecasting will be monthly, consequently.
It is the number of periods to be forecasted according to the chosen frequency. If your sales history is organized by months and you’ve chosen a horizon equal to 6, then your forecast will be of 6 months length.
In the case you don’t provide the lead time info of a product, then the system will use the default lead time value in days for the inventory forecasting and purchase order functionalities. However, it is strongly recommended to set the lead time info for each product of your catalog.
In the case you don’t provide the coverage info of a product, then the system will use the default coverage value in days for the purchase order functionality. However, it is strongly recommended to set the coverage info for each product of your catalog.
The system allows you to forecast new products, i.e. products with few data points. If checked, the system will automatically detect new products and you will be able to forecast them, otherwise it will consider them as canonical products.
Intuendi uses a Machine Learning engine which chooses the best forecasting model given a time series. Our engine exploits several statistical methods and proprietary algorithms for catching trends and seasonal patterns.
Intuendi does a bottom-up forecasting strategy. We do the forecast at the lowest levels of the catalog, then we summed them up the highest considered level. For example, the aggregated forecast for a product will be a summation over all the regions considered by the product.
Once the stock reaches the level indicated by the reorder point, the user should place an order. Such stock level is the combination of the lead time demand (the expected forecast demand during the lead time) and the safety stock level (a quantity based on the variability of the forecast).
The service level is the probability of not going stockout during the lead time. The larger the service level is, the larger is the the safety stock level. However, large values of the service levels may lead to extra levels of the stock.
Once the system suggests an estimation date of your replenishment, then an estimation of the stock arrival will be derived consequently on the base of the lead time length. From that moment, the forecasted demand until the coverage end is considered as the quantity of stock that you will need to reorder to satisfy the coverage.
In this case, an estimation date of the replenishment can’t be found. The system assumes that you will make an order immediately. The purchase order quantity is computed considering the forecasted demand from the moment the stock reaches zero to the end of the coverage period. However, it is strongly recommended to set the lead time info for each product of your catalog.
If a purchase order is marked as unnecessary means that the stock level will not reach the reorder point during the forecast horizon. Thus, no replenishment is needed since there is no risk of stockout.
The reorder point value considers the safety stock level which reflects the uncertainty of the forecast. It may happen that the system has forecasted a very small demand during the coverage period (even equal to zero) such that the stock will never go to zero. However, the safety stock level tells you to replenish in order to deal with unexpected future peaks of the demand.
Because the coverage period ends after the forecasted horizon. Thus, you will see only a part of it. However, the purchase quantity considers the extension of the forecast beyond the horizon in order to suggest you a reasonable quantity.
When a purchase order is overdue, the system assumes that you are going to place an order immediately. If it is not the case, the day after, the purchase order quantity should be recomputed since the coverage windows will shift one day ahead.