The following values are used in various tables and graphs and are calculated using Retailigence algorithms.
- Missing Lost Sales – Where a product is not sold in some stores within a cluster, this number is a prediction of the additional sales that adding the product to the missing stores could bring.
- Undersold Lost Sales – By comparing products, product groups and clusters, the machine learning algorithm can find trends and patterns. Within these patterns there are anomalous data points where a product is selling less than the trends would suggest. This could be due to out of stock periods or a merchandising / product location issue within the store. The algorithm then projects what the sales value would be upon rectifying the issue.
- Extrapolated Sales – Where a product has not been sold for the full time period selected eg. a line might have been introduced into the range after the start date of the plan. The algorithm will extrapolate the sales of the product to account for the period when the product was not selling.
- Projected Sales – The projected sales is a value consisting of the actual sales of a product, plus the previous 3 values. This value is used as the key value for ranking and sorting the assortment.