Daily Price & Stock Intelligence Across 7 Marketplaces

Client Snapshot

A mid-sized consumer electronics retailer selling across seven marketplaces in North America and Europe. Roughly 12,000 active SKUs.

The Challenge

The client's pricing team was checking competitor listings by hand every morning. By the time a spreadsheet was compiled, prices had already moved. Flash discounts from competitors went unnoticed for 2–3 days, and stockouts on competitor listings — the single best moment to raise price — were invisible entirely.

Three specific pain points:

• Manual checks covered only the top 200 SKUs, leaving 98% of the catalog unmonitored.
• No historical record, so the team couldn't tell whether a competitor's price was a promotion or a permanent cut.
• Each marketplace showed different prices by region and device, making manual comparison unreliable.

Objectives

• Monitor all 12,000 SKUs across seven marketplaces daily
• Deliver structured JSON to the client's pricing engine
• Store history so price movements can be distinguished from one-off promos
• Alert within 60 minutes of any competitor price change beyond a set threshold

The Actowiz Approach

We built a scheduled extraction pipeline with per-marketplace collectors, each tuned to that site's structure and protection level. Requests were distributed through a managed proxy pool with geo-consistency so that regional pricing was captured accurately rather than averaged away.

Product matching — the hardest part — was handled with a multi-signal identity layer (GTIN/UPC where available, falling back to model number, brand and title similarity). Every match carried a confidence score, and anything below threshold was routed for review instead of silently entering the feed.

Extracted records were normalized into a single schema, deduplicated, and written to a time-series store. A change-detection layer compared each run against the last and fired webhooks on threshold breaches.

Compliance Note

Only publicly listed product and pricing information was collected. No personal data, no authenticated sessions, no circumvention of paywalls. Collection rates were kept well within reasonable load, and site terms were reviewed before each source was onboarded.