Grocery retailer uses machine learning to adjust daily prices to maximize revenue and profits
Posted by | Shashank Kulshrestha
Background
A US-based Grocery and Fresh retailer was grappling with retail giants like Walmart and national grocery chains, and wanted to leverage Machine Learning to optimize pricing.
Objective
- The company wanted to leverage Machine Learning (ML) pricing optimization to model price-demand to maximize profit or revenue, based on the objectives of store owners.
- They also wanted to execute price optimizations on a weekly basis, while ensuring adherence to the rules and guardrails defined at the enterprise, category, and location levels.
Solution
- Fuld’s team of analysts designed jobs to integrate the weekly sales, product and sales hierarchy, competitor pricing, and other SKU metadata into the SAP database for both the SKUs and individual stores
- Using Python and Alteryx we created ML jobs to model a demand-price response and calculate predicted sales, volume, and gross profit for each SKU and price-point
- Once the jobs were in place, we designed the system to select optimal prices based on:
- Pre-defined price rules and guardrails
- Targets of balancing the revenue and profit goals
- As part of the verification stage of the project, we triggered and executed optimization jobs daily to recalculate optimal pricing which we then shared with our client’s pricing team to evaluate and approve
- We completed the project by creating dashboards in Power BI to help the client visually monitor revenue, profitability, opportunity curve, and the cost of rules
Outcome
- The prices were successfully optimized for each store location, product category and SKU family on a daily basis
- The retailer was able to meet and/or maintain the revenue and gross profit targets for 80% of its stores within the first 3-months of deployment of the new system.
Tools Used
Tags: Artificial Intelligence, Consumer Products & Retail, Machine Learning, ML, Pricing Strategy
