Skip to main content
Back to all work
Dealerships

E-Commerce Inventory Automation

Replaced 2 hours of daily manual parts filtering across 1.2M+ line items with a rule-based automation pipeline, with a turnaround of days, not weeks.

JavaJavaFXSFTPREST APICSV Processing

The Problem

The dealership group sold parts through an e-commerce platform, but getting inventory listed was a brutal manual process. The vendor ERP offered no direct database access, only an export tool that dumped a CSV of the full active parts catalog: 1.2 to 1.5 million line items.

Every morning, two staff members would pull that export and spend 1.5–2 hours manually scanning through it line by line, removing parts the company didn't want listed (open repair order parts, low-volume items, parts with no sourcing, and other criteria that shifted daily). They were scrolling through a million-row spreadsheet by hand, every single day.

A previous contractor had attempted a solution that automated the file transfer but still left the manual filtering untouched, cutting out maybe five minutes of a two-hour problem. The two staff members doing this work were experienced parts specialists whose actual value was on the phone with distributors, not buried in Excel.

What We Built

Rule Engine & Management Interface

Built a JavaFX desktop application where staff could maintain an exclusion list of parts the company didn't want sold online. The app gave them a simple interface to manage exceptions that the automated criteria engine didn't catch, without ever opening a spreadsheet.

Automated Pipeline

Scheduled the vendor ERP's export tool to drop the catalog CSV to an SFTP server. The pipeline pulls the file, parses 1.2M+ line items, applies the criteria engine to strip out parts that shouldn't be listed (open ROs, low-volume, unsourceable, and other dynamic rules), then runs a second validation pass against the manually maintained exclusion list. The cleaned catalog gets pushed to the e-commerce platform via API. The entire process runs unattended.

The e-commerce API was undocumented, so integration took some trial and error to get right. The previous contractor's solution had never gotten past the file transfer step.

The Outcome

Built and deployed the full solution in days, not weeks
Recovered 1.5–2 hours of daily labor from two experienced parts specialists, putting them back on distributor calls where they actually added value
Replaced a previous contractor's partial solution that had only automated the file transfer while leaving the real problem untouched
Gave the business a maintainable, rule-based system processing 1.2M+ parts daily instead of a manual task performed indefinitely