North Star Logic GroupStart a Project
← All projects

An internal dashboard for a logistics operations team

Designed and built a single-pane operations dashboard that replaced four disconnected spreadsheets and a tribal-knowledge runbook.

Date
September 2025
Client
Halyard Freight
Role
Solo engineer
Year
2025
Stack
  • Next.js
  • TypeScript
  • PostgreSQL
  • Drizzle
  • Tailwind

The starting point

Halyard's operations team coordinated dozens of freight shipments per day across half a dozen carriers, each with its own portal, file format, and timing quirks. The institutional knowledge for keeping it all moving lived in two senior operators' heads, four spreadsheets, and a Slack channel.

When one of those operators announced they were moving on in three months, the company needed to externalize that knowledge — fast.

What we did

We started by sitting next to the senior operators for a week. Not to design anything yet; just to understand what they actually did, where the spreadsheets fit, and which steps were judgment calls versus rote process.

From there we designed a dashboard around the day of an operator, not the shape of the data. The home screen surfaced the shipments that needed attention right now. Each shipment had a single canonical detail view with carrier-specific actions inlined where they belonged. The spreadsheets — every one of them — became reports that derived from the live system, not sources of truth in their own right.

We built it as a single Next.js app on Postgres, deployed to the team's existing Vercel project. No new infrastructure for them to learn.

The hard parts

The carriers were the hard part. Three of them had APIs, two had EDI feeds, and one was a person you emailed a CSV to. We built carrier-specific adapters with a shared interface, so the operators didn't have to think about which carrier was which — they just acted on shipments.

The other hard part was earning trust. The operators had been burned by software before, and they were skeptical that anything could capture what they actually did. We deployed in a read-only mode for the first three weeks: the dashboard showed everything, but the spreadsheets were still the source of truth. Once the team confirmed the dashboard matched reality, we flipped the direction.

Outcome

After two months the spreadsheets were retired. The departing operator handed off to a new hire who was running shipments solo by week two — something that had previously taken six months of shadowing. Halyard's operations team now uses the dashboard as their primary tool, and they have the source code, the schema, and the documentation to extend it themselves.