Backstory
Every day, Walmart department managers spent over 2 hours scanning every row, shelf, and aisle, asking the same questions:
- Is everything in stock?
- Are the shelf tags aligned?
- Are products displayed correctly?
It was a time-consuming process, and no one enjoyed it—not even the executives.
That’s when Alice, a VP of Store Operations who worked her way up from the sales floor, stepped in. She was determined to streamline operations, and section work was at the top of her list. I joined forces with Alice, a group of dedicated product managers, and a couple of talented data scientists to transform this process. Our mission was clear: make inventory management more efficient and eliminate the headaches of “inventory exceptions.”
Trying it for myself
I needed to understand the employee experience, so I spent time on the sales floor, shadowing department managers, observing their behaviors, and learning their tips and tricks.
When I tried the existing process, I found it tedious, time consuming, and error-prone. Department managers often dealt with countless inventory discrepancies, wasting hours locating missing or misaligned products. There was no sense of priority, and cascading inventory issues made the process feel like an endless chore.
The challenge was clear: how do we leverage data science to create a process that even the least experienced employee could follow and, ideally, eliminate section work altogether?
Making it better with paper prototyping
To start, I needed to validate that our data models worked. We tested the predictive algorithms across multiple stores, and they proved effective, with a failure rate of less than 8%.
Knowing where to look for problems cut down the time spent in each section and increased feelings of accomplishment. We enabled ourselves to mark cascading issues for follow-up, turning a sense of defeat into productive action.
Concept creation
Working closely with the data science team, I led the creation of prototypes directly in stores, focusing on the key pain points. Instead of scanning every product aimlessly, our design targeted the most pressing exceptions first. This reduced the time spent in each section and transformed the experience from frustrating to purposeful.
We developed multiple concepts for a mobile app that centered on retail workers’ needs, making the data models trustworthy, the instructions clear, and the process streamlined. By incorporating images, product descriptions, and organizing long numerical strings, we significantly improved how quickly employees could locate products.
After multiple rounds of testing, a process emerged
I created low-fidelity sketches and repeatedly tested them with employees, refining the experience based on their feedback. One key insight emerged: workers focused more on shelf tags than product images. By mirroring real-world shelf tags within the app, we dramatically improved the speed of locating exceptions.
A functional MVP
After multiple rounds of testing, we developed a high-fidelity prototype that offered a streamlined, step-by-step process for locating exceptions. The MVP included:
- Allowing users to select and favorite their assigned departments
- An augmented product scanner with expected values and coordinates
- Automatic data model updates when incorrect products were found
- Motivational messages to guide workers to their next task
The result? A solution that turned section work into a manageable, efficient process. By designing around the real-world behaviors and challenges of Walmart employees, we cut the average inventory count time by 30%, and department managers no longer dreaded their daily inventory checks.
As the lead product designer, I didn’t just create a tool; I built a solution that respected the user’s experience, transformed a painful process, and brought a sense of accomplishment to the daily tasks of Walmart’s store employees.