Back to work

Product · Wholesale sourcing

Sourcely

EZCLOUD

Stack
  • Vue 3
  • TypeScript
  • Python
  • Django

Analysts upload a supplier price list. Sourcely matches it to Amazon listings, prices in every fee and moves each product through Review, Invest or Don't invest.

The Challenge

Supplier files are large and messy, while Amazon and Keepa limit API calls. Decisions had to be fast, shared across the team and explainable: why a product was bought or skipped.

The Approach

  • Price lists parsed in a Web Worker with an instant preview and resumable enrichment that stays inside Amazon and Keepa rate limits.
  • UPC-to-ASIN matching with 90-day Keepa history. FBA/FBM, referral and seasonal storage fees computed down to net ROI.
  • A rules engine ("buy if…") previews how many products a rule will hit before saving. Manual overrides keep a reversible link to the rule.
  • Kanban pipeline with drag and drop and keyboard control, synced live for the whole team.
  • Migrated about 150 components from Vue 2 to Vue 3 and TypeScript over 3 months, with no feature freeze.
  • Built on the shared EZCLOUD design system.

The Outcome

100k

products per catalog

10

analysts using it daily

Pipeline: every product in Review, Invest, Don't invest or Snoozed
Analyze: matched products ranked by profit and ROI, with a decision on each
Product: listing data, offers and 90-day sales rank
Imports: every supplier file with matched rows and errors
Marketing site: the landing page runs real product pieces