How EPAM built a recommendation engine for brand portfolio decision making
The challenge: optimize merchandise levels
A trial and error approach to optimizing merchandise levels for a single retail store is inefficient and ineffective. To reduce the amount of unsold stock, Merchandising Managers and Brand Portfolio Managers must have a good idea of what will sell and what will not. Capitalizing on regional data to predict brands or products with higher sales potential puts any retail company on a better path.
The solution: Recommendation engine and Guided Analytics
A data science team uses data analytics platform, KNIME, to create a solution similar to collaborative filtering. It makes automatic predictions of individual customer interests by collecting preferences from many customers. The workflow starts with ETL and other data preparation steps, before creating a recommendation engine, and lastly through to determining interaction points for the Analytical Application. The workflow is then deployed on KNIME Server as a Guided Analytics Application.
⇒ Download Workflow from KNIME Hub
Why KNIME?
Tasks like ETL and data prep require a certain degree of technical knowledge - as does creating a recommendation engine. In this case, data scientists can focus on creating and deploying an Analytical Application from which Merchandising and Brand Portfolio Managers can draw insights and make decisions.