[CASE STUDY] In-depth Plan Profitability Calculations for Violeta

 

Production and distribution companies such as Violeta have thousands of SKUs and hundreds of sales/distributor locations. This results in tens of millions of time series, which is somewhat impractical for making detailed plans. In Farseer, it is possible to enter a number of units sold, for example, on any level – SKU, client, business unit/branch, or any combination of these variables. The optimization algorithm then distributes the amounts across all relevant dimensions, solving systems of several thousand equations. Unlike standard machine learning systems, Farseer’s optimization algorithm does not work as a black box, generating random results that cannot be influenced in any way, but it places users in the center of the process, giving them a chance to affect any number and adapt it to the real-world conditions. The rest of the plan optimizes automatically and adapts to the number that is manually entered.

 

For example, planners in Violeta can set a general number of 500 000 SKUs sold in the next year, and, based on all available data, the system automatically calculates sales goals for every customer, together with amounts, detailed prices, and specific rebates for each one of them. If the collaboration with a client goes better than initially planned, users can manually enter the SKU number just for them, and all other values across all other relevant dimensions automatically and momentarily adapt to the newly entered value.