Company
Lune
Role
Product Designer
Scope
Research · Workflow design · UI
Logistics Data Upload
Coming out of the product consolidation, two import tools needed to become one. Both ran emissions calculations at scale, but neither was suitable as-is for both user types, and the gap was already being filled by the customer success team manually uploading data.
Two tools, one problem
Lune's import tool was proprietary, built in-house, simple, and reliable. It did what it said, but it was relatively rigid: columns had to match, formats had to be exact, and there was limited guidance when something went wrong. Experienced users liked it, but there was no flexibility to include optional, additional fields.
Gryn's import tool was a third-party integration, flexible and technically capable, but expensive and genuinely difficult to use. The configuration overhead was high, error messages were opaque, and the learning curve was steep enough that customers were regularly calling in for help.
The situation
Simple but rigid
Proprietary, fast, and reliable, but inflexible on format. Strong for power users. Poor onboarding for new ones.
Flexible but broken
Third-party, highly configurable, but expensive, hard to use, and generating more support load than it was worth.
Where it actually broke down
The friction point was structural, not just usability. Shippers, the companies generating the shipment data, often needed their LSPs to actually perform the upload on their behalf. The LSP had the technical access and was responsible for the calculation. The shipper had the data.
But LSPs had no strong incentive to do this reliably. When the tool was hard to use, they'd delay, deprioritise, or do it badly. Shippers chased. Eventually, customer success stepped in and uploaded the data themselves, which was fast, but wasn't sustainable, and meant the tool had effectively failed at its core job.
"When the tool is hard to use, the customer success team becomes the tool. That's not a support problem. It's a product problem."
Design goals
One tool, two users
A single import experience that worked for both LSPs and shippers, with the interface adapting to surface only what was relevant to each.
Simpler without losing flexibility
Reduce the configuration overhead of Gryn's tool and the rigidity of Lune's, keeping what was useful from both and dropping what wasn't.
Reduce CS burden
Make the tool good enough that customers could use it independently, so the support team wasn't filling the gap.
Process
The workflow analysis came first. I mapped the full import journey for each user type, not just the steps inside the tool, but the upstream context: where the data came from, who prepared it, who uploaded it, and what happened when something went wrong. That's where the incentive misalignment between shippers and LSPs became clear.
From there, the design work focused on progressive disclosure, showing each user type the options relevant to them, at the moment they needed them, without burying the flexibility that power users depended on.
Design
The unified flow was validated by stress-testing it against multiple scenarios: different user types, different data shapes, different error states. The question was always: does this hold up when the data isn't clean, the user isn't experienced, and the LSP is doing this on behalf of a shipper who isn't in the room?
Real users were brought in to validate the flows. The feedback was positive, and the customer success team, who'd been living with the old tools longest, were notably enthusiastic about what the new approach would mean for their workload.
Final designs
Outcome
The project was in progress when I left Lune, so there are no shipped metrics. What exists is a validated design, tested with real users, stress-tested across scenarios, and signed off by the team. The customer success team's reaction was the clearest signal: they could see exactly how it would change their day-to-day.
Sometimes the most meaningful outcome of design work is the moment a support team stops seeing themselves as the backup plan.
What I took from it
The workflow mapping was where this project was won or lost. Understanding that the upload friction wasn't just a UX problem, it was an incentive problem between two parties, and that changed what the design needed to solve. A better UI alone wouldn't have fixed it.
The other lesson: progressive disclosure is genuinely hard to get right. Hiding complexity from users who don't need it, without removing it for users who do, requires a precise understanding of who's in the room at each step. That understanding came from the workflow research, not from intuition.