Next Generation Store Receiving
Overview:
My Role: Sole Designer
Tools: Pen and paper, Sketch, Protopie
Activities: In-person research, User Flows, Wireframes, Prototyping, Usability Testing
Team: Myself, Product Manager, Business Operations Partner, Engineering Manager
The Problem: Reduce the time it takes suppliers to deliver product to the store
Press play on these two videos to see prototypes of the product.
This is a screen recording of the happy path of the prototype.
This is a recording of when an audit would be triggered. This would flag an associate to count the merchandise before the delivery could be completed.
This is a concept for a new way of receiving product into Walmart stores. This project was in a proof-of-concept phase being run in 5 stores when I left the company, but it was being met with great success.
The Problem
It took too long for deliveries that come directly from suppliers to our stores to be dropped off and accounted for. The process would last a minimum of 30 minutes in the best circumstances, and would regularly last for 1.5-2 hours. We were looking for a way to shorten that time.
Understanding the Problem
Every day, suppliers send trucks to our stores where they wait for our associates to check them in and validate what they brought. The idea was brought to me to allow the drivers to bring in their merchandise and use a kiosk we had set up to check themselves in. Our associates would only get involved if they needed to validate that the merchandise was correct.
This was the first project I worked on at Walmart where the user was not a Walmart associate. Rather, the users were the drivers of the trucks that come directly from suppliers instead of Walmart distribution centers.
After mapping out the journey suggested for the driver by my product team, I realized that it was essentially the process of our Self Check-Out area where customers buy product. I reached out to the Point of Sale design team to learn about their designs and to see what research had been done around hardware. I took their designs and quickly turned around some wireframes of a possible "Self Check-In" to do some quick validation.
Luckily, the designs of our Self Check-Out point of sale systems only required moderate tweaks to become a reasonable set of wireframes to communicate a concept to my partners.
I prototyped this, put it on an old Android tablet I had, and brought it to the stores between 4am-12pm (this is when supplier receiving happens) to see if the drivers could use it. I had no troubles in this stage, likely because I was able to benefit from the Point of Sale team's extensive testing and really tight designs.
Pivoting
The next item to be solved was figuring out what device this should be on. We looked at a lot of options, and traveled to meet a number of vendors for them to show us what hardware they had for us. Though the product team was close to sold on a number of the options we saw, I became more dissatisfied with each one we looked at. Having spent time in the store doing the job of the receiving associate, I realized that this solution needed to be able to move. Between the massive pallets rolling around and the number of people in the area, a kiosk would be far too restricting.
I sketched out what I thought should be the solution to this problem and presented it to my partners.
There was some resistance to this idea from my product partners that were really sold on the idea of a kiosk. They had paid for a prototype of the kiosk to be made and, though they had encountered a number of issues, they thought the cons of a kiosk would have less of an impact than switching to a mobile solution at this point in the process.
I needed to gather some data to back up my concept. I built two prototypes: one for a mobile solution using the store devices, and one using the driver's devices.
The wireframes for the store devices. The devices have a built in scanner which makes them great for dealing with a lot of merchandise at once. The drawbacks are that they are expensive, and it is often difficult to get your hands on as associates sometimes hide them to use for themselves.
The wireframes for an external app that the drivers could download to their own phone. This was my ideal solution, though I needed to figure out a way to make scanning with a camera as efficient as possible.
Both of the prototypes I created tested very well with the drivers and associates, and helped convince my partners on the project to pursue this concept. At the same time as I was conducting this validation, the prototype of the kiosk was placed in a store and used by one of our vendors. The feedback was poor, and reflected a number of my concerns with a stationary solution in such a chaotic environment.
In order to mimic the efficiency of scanning with the store devices with their dedicated scan button and their built-in scanner, I designed the camera scanner solution to run the camera and only react to the barcodes we knew to look for. Luckily, we knew what was being sent so we could ignore any barcodes we registered that did not match the anticipated deliveries.
I presented this concept to a few of Walmart's biggest suppliers as part of a forum introducing some of the ways we wanted to improve our methods of working together moving forward. There was a great deal of excitement at the prospect of massively reducing the time their drivers would spend in the store.
5 of our largest vendors, including Coke and Pepsi volunteered to be part of a pilot program in 5 stores with an MVP of the product. The results speak for themselves:
Driver wait time was lowered from more than 1 hour on average to under 20 minutes if there was an audit, and less than 5 minutes if there was no audit.
Conclusion
Unfortunately, there were some delays with this project as COVID-19 hit while development was happening on the tools. I left the company before it could be launched store-wide, though we were able to see huge success in a proof of concept that was run. This MVP version of the tool brought driver dwell times down from that hour+ wait time to less than 20 minutes in every case. When there was no audit, the delivery time was often less than 5 minutes.
