Loading
Loading
Case study · Studio study
Shipments, exceptions, and a map that is not a toy.
Studio study with a working title. Not a named-client claim.
Overview
Harbor is a representative logistics control surface: shipments as a lifecycle, partners as APIs, exceptions as a queue. Automation proposes delay notices; dispatch confirms.
The challenge
Logistics software fails when the happy path is a truck that left on time. The product is the exception: late, split, damaged, unlabelled. Also: partners who send you XML at 2 a.m.
Objectives
A single shipment state machine
Inbound adapters that isolate partner chaos
Exception queue with SLAs
Deployable workers that can be replayed
Strategy
Event-first. Partners never write your core tables directly. Mapping lives in adapters. The UI is a radar for exceptions, not a form to re-enter the bill of lading.
Discovery
We listed every status a shipment can honestly have, including the embarrassing ones. If a status existed only in Slack, it went on the machine.
UX
A board of exceptions, a shipment timeline, and a partner health view. Search is an ID, a reference, or a fuzzy consignee — not a philosophy.
Design
Warehouse lighting: dark, high contrast, large status chips. Maroon for blocked. Map is secondary to the list; operators think in tables.
Engineering
Go or Node workers for ingest, Postgres for the machine, Redis for live presence, object storage for documents. CI replays a fixture pack of partner payloads.
Architecture
Ingest
SFTP, webhook, EDI-ish payloads.
Technology
Backend
Frontend
AI
Infra
Implementation
Result
Exception MTTA
Placeholder
Partner payload fixtures
Versioned
Replay
Supported
Replace with measured results after a client engagement.
Voice
“Replace with a dispatch lead: the product should make the ugly shipment visible, not the pretty one.”
Start
Tell us what you are trying to ship. We will come back with questions, a shape, and whether we are the right studio for it.