Recommendation system managing inventory across 1,000+ stores

I designed an internal retail recommendation system for Stockwell’s retail team to optimize time and resources. Powered by machine learning and crowd-sourced data, the tool intelligently suggested which products to stock in each machine. It streamlined the team’s workflow, cutting planning time from three weeks down to just one.As the lead designer, I worked with 1 product manager, 2 engineers, and 2 retail buyers for a course of 4 weeks.

Problem

It was unsustainable for our 5 person retail team to manage 1,000+ Stockwell stores.

Hypothesis

If we give the retail team a tool that automates their workflow, we can save on headcount, increase revenue growth, cut operational costs, and manage large groups of stores with a small team.

Solution

Designed an MVP dashboard that simplifies the retail team’s workflow by automating recommendations.

Outcome

The MVP was launched after I left Stockwell in 2019. The Company closed in 2020, and I assumed the tool had a brief run before the company was immediately impacted and had to shut down due to covid-19.

The Designs

Dashboard overview

The dashboard opens to a list of stores, each displayed with its dataset in columns for easy comparison and optimization.

Store & Item Tabs

Retail Buyers can see the store schema on the left and the product table on the right. Selecting an item in the table highlights its location on the shelf.

Recommended optimizations

The ML algorithm suggests optimizations by listing which items to add and which to remove. Retail Buyers can review the metrics shown next to each product and decide whether to accept or reject the recommendation.

Contributing to Stockwell’s design system

While working on this project, I had the opportunity to work on other small projects at the side. I contributed to Stockwell’s design system and a snapshot of it is seen below.

Check out other works

Farther Finance