Sable & Row

Business problem

Most storefront platforms treat every order as either "fine" or "support ticket," which buries fulfillment problems (damage, misdelivery, short-shipment) inside a generic support inbox instead of tracking them as operational exceptions with their own resolution path.

Users & stakeholders

• Shoppers browsing and checking out on the storefront
• Warehouse/fulfillment staff packing and shipping orders
• Operations staff resolving fulfillment exceptions and carrier claims
• Support staff processing returns and exchanges

Proposed system

CASSINI separated three concerns that a generic order table collapses into one: normal order fulfillment, fulfillment exceptions (delivery disputes, damage, short-shipment), and the returns/exchange pipeline — each with the view type that matches how staff actually work it.

Public-facing experience

The storefront organizes around four trip-specific collections with the technical detail (pack weight, temperature rating) a buyer actually needs to decide, rather than a generic product grid, and states the return window plainly rather than burying it in a policy page.

Admin / operations experience

Orders use a table because staff scan status across many orders. Fulfillment exceptions use a kanban because each exception moves through a real investigation workflow (flagged → investigating → carrier claim → resolved). Returns use a checklist because the operative state per item is a short, ordered sequence (item received → inspected → refunded), not an open-ended board.

Workflow & integration model

Fulfillment exceptions and returns are modeled as two separate pipelines rather than one, because a fulfillment exception (carrier's problem) and a return (customer's choice) trigger different internal owners and different resolutions — collapsing them would misroute work.

CASSINI capabilities demonstrated

Demo disclosure: Sable & Row is a fictional organization created by CASSINI to demonstrate an end-to-end system build for the ecommerce industry. It is not a real client, and no real people, records, or performance metrics are represented here.