AI product strategy
A Databricks optimization platform for brand mix and facing allocation, with what-if simulations and global performance insight.
Problem
Brand-mix and facing-allocation decisions ran on fragmented, Excel-based processes: heavy manual effort, little ability to simulate scenarios or compare alternatives, and no structured history. Hard to optimize space consistently or align decisions across regions and banners.
Approach
I led the evolution from a pure facing-allocation logic into an AI product-strategy platform on Databricks. Optimization algorithms recommend the best space allocation across brands and products, maximizing performance under business constraints, weighing productivity, marginality and store-specific traits. On top: what-if scenario simulation, new-brand and rebalancing suggestions, intra-year adjustments, and versioned tracking of strategy runs. A Power BI reporting layer gives a unified, comparable, historical view across scenarios.
Impact
Turned product strategy from a manual, one-off task into a scalable, continuously evolving system. Merchandisers can simulate strategies, compare outcomes and decide with clear trade-offs; global visibility and history tracking improve governance and cross-country alignment, raising decision speed, consistency and transparency.
Built on Databricks.