XShip + XMove

Fulfillment on easy mode.

An empty warehouse, running at full speed on day one.

Company
Flexe
Years
2018
Team
No PM, first version built with two engineers
Role
Product and design lead

At a glance.

Empty building → working warehouse.

As Flexe’s product and design lead, I spent a month on warehouse floors, then designed XShip and XMove: the desktop system leads use to run the warehouse, and the handheld app floor workers use to do the work. An empty building could be live in about two weeks, with new hires at peak proficiency in a day and a half. It made Flexe the world’s first on-demand fulfillment platform, across more than 2,000 warehouses.

Zero training required. Peak proficiency in a day and a half.

New hires were at 80% on day one and 100% by the middle of day two. A traditional system takes about three months.

The work finds the worker.

XShip creates every task the moment a truck calls in or an order comes in, and sends it to the right person.

Feels like the phone in your pocket.

The home screen was a grid of colored apps, modeled on the apps workers already used every day.

One step at a time.

XMove walks each worker through a task card by card, and tells them what to do when something goes wrong.

Live in as little as five days.

An empty building could be shipping orders in five days to two weeks, versus about four months.

99% shipped on time.

Across more than 2,000 warehouses worldwide.

Problem and opportunity

Untapped supply: converting dumb space into smart space.

There was no cost-effective way to scale an on-demand fulfillment network with existing software.

Research and discovery

Packing boxes and paper prototypes.

Before I designed anything, I spent a whole week working in the warehouse.

I interviewed everyone from operators and managers to leads and floor workers. Then, using a jobs-to-be-done approach, I worked every task in the warehouse myself, looking for where the software got in the way and how it could help instead.

“We don’t even use 95% of this.”

— Warehouse lead

“And the worst thing is when they update the software, then I have to redo all my notes.”

— Warehouse lead

“The software makes my job harder, not easier.”

— Warehouse lead

The software was so hard to use that people left themselves notes about which buttons to click. None of the systems were fast enough to support the on-demand model Flexe was seeking to create. What workers needed was far fewer independent functions, and more consistent, easily repeated workflows and automations that removed a lot of the busy work.

Validating our assumption

Real-world testing.

Testing on the floor showed us that a lean, tactical approach wasn’t a compromise. It was our advantage, and the only way to make on-demand work.

In about two weeks, using a jobs-to-be-done framework and lean methodology, we built a demo that proved we could build a working warehouse management system from scratch, one that worked better than systems that had been around for decades.

5warehouses

1 monthon site

Dozensof one-on-one interviews

Thousandsof boxes packed

Scope and influence

Little risk, big reward.

I led the effort from research to recommendation. With our engineers and tech leadership, I brought it to the rest of the business as an investment case: build a working version, and use Casper’s new sites as the base case. Casper was growing fast, and if we couldn’t scale with them, we’d lose them as a customer. One mattress in multiple sizes meant we still covered every workflow, from checking trucks in to loading them out, but only the simplest version of each.

One customer’s integration

6 months
8 engineers

Our first version

3 months
2 engineers

Design tenets

A recipe for success.

Based on my research, I created design tenets to make sure the product would succeed in our on-demand model, and to scale my expertise to the rest of the team.

  1. 1Ask first: can this be automated?
  2. 2No exception left behind.
  3. 3Training and documentation are a failure state.
  4. 4Day-one productivity.
  5. 5Use human terminology, not industry jargon.
  6. 6Don’t make me think.

On-demand space meant on-demand workers, often starting the same day they were hired. You can’t go on demand with software that requires three months of training. To be on demand, we had to be faster, more flexible, and easier to use.

Product and solution

Meet XShip + XMove.

Two apps, one system. XShip helps warehouse managers and leads run their operation, while XMove helps workers complete each task on time.

Everything you need to run a warehouse, from inbound to outbound.

  1. 01Truck check-in
  2. 02Trailer unload
  3. 03Receiving into inventory
  4. 04Put away
  5. 05Cycle counts
  6. 06Waving
  7. 07Picking
  8. 08Sortation
  9. 09Packing
  10. 10Trailer load

And here’s how XMove helps you get it done.

  1. 1Scan the bill of lading.
  2. 2Get the driver’s signature.
  3. 3Move each pallet to the staging location.
  4. 4Scan the staging location, then scan each pallet.
  5. 5Enter the number of pallets.
  6. 6Unwrap the pallets, scan all the SKUs, and enter the quantity of each.
  7. 7Wrap the pallets back up.
  8. 8Print a label.
  9. 9Apply the label to the pallet.

Task automation

If this, then that.

By investing in task automation, we let leads stop handing out work and start improving it.

When every job automatically created the work that followed, leads and managers no longer had to assign tasks, chase people down, or make sure everyone had something to do. They could spend their time improving what was already happening.

Placeholder: animated or stepped flow
  1. Wave criteria met
  2. Wave
  3. Batches
  4. Pick trips
  5. Assigned to workers
  6. Pack stations
  7. Boxes moved to the outbound door
  1. Truck ETA entered
  2. Dock assigned (or waiting)
  3. Unload task created for the right number of workers
  1. Cycle counts created automatically, so no one has to remember them

Labor planning (see the work coming tomorrow, next week, next month)

Guided tasks

We’ve got your back, every step of the way.

XMove was designed for simplicity, with no training required. We broke each task into small pieces and guided workers through them one at a time.

Beyond the pixels

Fulfillment in a box.

Different equipment meant different workflows, even for the same customer. So I designed a pop-up fulfillment kit to give every site the same baseline, and to launch faster with less time on site.

The details

The little things that matter.

XMove

Feels like your phone. The home screen is a grid of colored apps: picking, cycle counting, moving inventory.

Every app has a color. Find an abandoned device, or help someone who’s stuck, and you instantly know which task they’re in.

Unload truck, not Receive SKUs. Human words, not machine words.

Spanish in one tap. Top right of every screen.

Badge-scan login. Shared devices, ready in a second.

Apples and zebras. Every location kept its code and got an icon, faster to spot in any language.

Gloves on. Every screen works with a keyboard, not just touch.

Actionable errors. Instead of a code like “1274971 error,” XMove says what happened and what to do next: “Go find Miguel, your outbound manager,” or “Take your cart back to your station, leave the goods in it, and notify your lead. Then start a different task.”

XShip

Faces and big progress bars. The tasking page shows workers’ faces and names, with big, bold progress bars instead of tiny ones.

Color carries over. Each app’s color is the same in XShip and XMove.

Today, not October 29. Drop-offs and pickups say Today or Tomorrow instead of a date stamp.

Prep for tomorrow in one click. Wave tomorrow’s shipments by tapping Tomorrow instead of using a date picker.

Waving by picture. Tap the UPS logo and a picture of one box instead of typing codes.

Trucks you can see coming. Each truck travels along a bar in its carrier’s color (brown for UPS) toward its arrival time.

User setup in seconds. Take a photo of each worker’s ID, and their profile is created automatically. Pick the jobs they’ll do, and the right role and permissions are assigned. Choose how long they’ll be there, so access doesn’t last longer than it should. Print their scan card; on first scan they set a four-digit PIN, then log on to any device with their card and PIN. No passwords or email addresses.

Know your best people. Each worker’s page shows time on task, accuracy, and total work done, so warehouses could retain their best temps or move them into full-time roles.

The outcome

Live in two weeks, not four months.

2,000+fulfillment centers worldwide

2-weeklaunch time

99%shipped on time

80%win rate

FlexeTraditional WMS
Launch time
2 weeks4 months
Training to peak proficiency
1.5 days3 months
New managers to hire
NoneA full team
License cost
FreeAbout $200,000 a year
Floor square-foot cost
PlaceholderPlaceholder
Cost per touch
PlaceholderPlaceholder

Let's build incredible0→1 products together.

0→1 products =UrUser Research+UxUser Experience Design+PtPrototyping+TsTesting