Case study 02 · Wayfair · Freight forwarding

Transforming Freight Forwarding at Wayfair: The Journey of NaviOS

Three systems, a spreadsheet and no single source of truth, replaced with one platform the operations team could work in.

Role
Lead Product Designer.
Time
Two years, 2022 to 2024.
Team
Director of product, four engineering teams, operations leadership.
Scope
Web based, responsive.

01 · Background

CastleGate Forwarding

CastleGate Forwarding is Wayfair’s own logistics operation. It moves product from factories around the world to customers across North America and Europe, consolidating shipments upstream so there are fewer touches and less time between the factory and the person who bought the sofa.

Over 23,000 suppliers and more than 80,000 containers. By volume that is the equivalent of a top 20 importer into the United States. NaviOS is the platform the operations associates use to run it.

02 · The problem

The same shipment, typed three times

Before NaviOS there were three places the work lived: NCT, CargoWise and spreadsheets. None of them checked each other, so there was no single source of truth and associates re-entered the same shipment up to three times, once per system. Over 40 fields, by hand, per shipment.

Nothing validated any of it. Errors stayed invisible until they surfaced at customs, which is the most expensive place to find them.

  • Manual data entry

    Typing into spreadsheets produced frequent errors and inconsistencies, and customs and routing depended on that data being right.

  • No integrations

    The tools sat in silos. Data had to be moved between them by hand, which duplicated effort and multiplied the chances of a mistake.

  • Time lost to routine

    Associates spent a significant part of the day on tasks that could have been automated, which delayed everything downstream.

The fourth problem was the one that made the others worse. The tools were cumbersome enough that associates were frustrated by them, and a frustrated user works around a system rather than with it.

03 · Research

Twenty associates, three methods

To understand what the work actually looked like, I used three methods rather than one. Each answered a different question.

  • Interviews

    One-on-one with more than 20 operations associates, to get detail on daily tasks and where they got stuck.

  • Shadowing

    Time spent watching people work in their own environment, which is where the workarounds show up.

  • Surveys

    Sent to a broader group, to find out how common the problems were and to catch anything the interviews missed.

A workshop room with interview notes grouped into columns of sticky notes on a black board, flip chart paper of handwritten sprint questions taped to the wall beside it, and a facilitator standing at the board.
Making sense of it. Notes grouped on the board, and the questions the sprint still had to answer written up on the wall next to them.

04 · Ideation

From workshops to flows

I ran brainstorming sessions using mind mapping and affinity diagrams so that every idea got captured and grouped, then a Crazy 8s workshop to push past the obvious ones.

Four things came out of it: a centralized dashboard with real-time shipment updates, automated notifications for changes and delays, simplified data entry forms, and integrations that let data move between systems instead of being carried between them.

Then I mapped the flows. The existing ones were long and fragile. A customs specialist entered the same data into several systems in sequence, and any inconsistency introduced early survived all the way through. The redesigned flows integrate the systems, automate validation and notifications, and cut the number of steps a person has to hold in their head.

05 · The solution

One platform, three changes

After the research, ideation and prototyping I implemented a set of changes in NaviOS aimed at usability, efficiency and accuracy, with systems integrated and routine work automated.

The NaviOS shipments list. Each row shows status, carrier, bill of lading, vessel ETA and port of discharge together.
The shipments list. Status, carrier, bill of lading, vessel ETA and port of discharge on one row, so an associate can scan 234 open shipments without opening any of them.

Integrated data entry forms

Reduces manual data entry and errors by automating data validation and integrating with existing systems.

Automated notifications

Keeps operations teams informed in real time about changes, delays and required actions, without manual monitoring.

Enhanced user interface

Optimizes the process of entering, updating and tracking shipment information.

Data entry screens including address verification with a suggested match, a basic shipment information form with inline validation, and a prepull edit dialog.
Data entry. Address mismatches are caught and offered as a choice rather than a free text field, and the forms validate against the systems they write to.
A shipment detail view showing rate types with their status, alongside links to booking requests, house bills of lading and ocean bookings.
Shipment detail. Rate status is explicit, including what failed and why, and the related records are one click away instead of in another system.
Design artifacts including a tags versus flags model, a shipment information card, and a dynamic date filter with its overview and flow.
Interface work underneath the screens. Tags and flags do different jobs, so I separated them, and the date filter got a defined model rather than an ad hoc one.

06 · Delivery

Documentation as part of the design

Four engineering teams were building from these designs, so handoff was not a final step. I prioritized documentation and collaborative practice throughout.

  • Embedded comments

    Designs carried comments and guidelines for developers and stakeholders, in the file, next to the thing being described.

  • Written specification

    User flows, functionality, reasoning and visual guidelines, so a decision could be understood months later by someone who was not in the room.

  • Shared language

    Consistent communication across teams, which is what stops four teams building four slightly different versions of the same pattern.

07 · Results

Measured a quarter after launch

Measured one quarter post-launch with a subset of current and new NaviOS users.

50%

Reduction in manual data entry.

42%

Decrease in data entry errors.

20%

Increase in compliance.

NaviOS has significantly reduced the time I spend on repetitive tasks, allowing me to focus on more important work.

Operations associate, drayage

A satisfaction survey ran alongside. Associates rated their overall experience against NCT and CargoWise at 4.1 out of 5, performance at 3.6, and ease of use at 3.4.

Ease of use at 3.4 was the lowest of the three and the most useful number in the set. The platform was better than what it replaced and still harder to learn than it should have been, which is what set the backlog priorities afterwards.

08 · What came next

After the launch

What I did

Started weekly check-ins with operations stakeholders for backlog refinement and prioritized requests. Addressed three of the four highest priority items and shipped them. The rest went onto the roadmap.

Looking forward

Additional user training, including Lunch and Learns and a NaviOS power user bootcamp. A quarterly measurement process so the product team could track metrics and find the gaps rather than wait to be told about them.

Let’s talk.

I’m open to UX and product design roles, Seattle or remote.