The Lowest Allegro Prices from 150K Eans Collected

The client ran a price optimization operation on Allegro.pl. The business logic behind it was short: take the lowest price a product was currently offered at on the marketplace, put that number next to the price on their supplier's list, and check whether they could go below the current market rate and still earn on the sale. Everything downstream depended on the first number in that comparison being current, so the market side had to be collected fresh every day rather than sampled now and then.

Sources: allegro.pl
Coverage: EANs supplied by the client, up to 1.2 million on the list; for each, the cheapest offer for a new product sold with a VAT invoice
Frequency: daily, from early 2025 to May 2026
Volume: roughly 150,000 EANs checked per day in 2025
Delivery: a daily file in XLSX, JSON and CSV; from October 2025 also a price and stock history in the client's BI dashboard on their own server
Setup and launch: 12 working days of development and testing

The client's list defined the job, not the marketplace catalog

This was not a category crawl. The client sent over a list of EANs, and our task was to resolve every one of them on Allegro.pl and return the lowest price found for it. The input set stayed under their control, which meant building around identifier lookup at scale, with no assumption about which products would show up on tomorrow's list. A product we failed to resolve was not just a blank row in a file - it was an item they could not make a pricing decision on that day.

The list also changed size: about 237,000 EANs in September 2025, then 1.2 million after 11 October 2025, roughly a million of them never checked before. Those went to the front of the queue. Only some offers counted: the cheapest new product sold with a VAT invoice, so used and refurbished listings were left out.

Sustaining 150,000 Allegro lookups a day was where the engineering went

Allegro.pl runs strong anti-bot protection, including a slider captcha, and working with protection of that class required a dedicated solution on our side. The genuinely difficult part was different: keeping roughly 150,000 lookups per day running without the daily cycle degrading. Far more of the work went into run stability at that volume than into reading a product page.

A typical day was 8 September 2025: 157,567 EANs processed, 129,167 found and 28,400 with no matching offer. Both outcomes went into the file with a check date, so the client could tell an item with no offer from one we had not reached.

Built and proven in 12 working days

Development and testing took 12 working days. Testing carried real weight here, because the thing being validated was not whether a single price could be extracted correctly - that part is straightforward - but whether the full daily volume would come back complete and on time, repeatedly. Once that held, the setup moved into daily operation: list in, lowest prices out.

From a daily file to an Allegro price and stock history

By October 2025 the daily file held about 1.3 million rows, more than a spreadsheet can take, and one row per EAN no longer showed what the client wanted most: how price and stock changed over time. So the client provided their own server, and we set up a database and the client's BI dashboard on it. Every check of an EAN went into its history, price and stock included; later the dashboard also listed each EAN's sellers with their prices and delivery dates. In December 2025 the client confirmed the history worked and asked only that it be saved daily.

What the client had

Every day the client received the lowest market price for each EAN on their list, refreshed for roughly 150,000 EANs a day in 2025. It went straight into their pipeline, where each market price was set against supplier pricing to decide whether a deal was worth taking, and from October 2025 the same checks also built the price and stock history. Until the project closed in May 2026, pricing decisions rested on prices collected that day, not on a snapshot from last week, and the comparison that drove their model ran as a daily routine instead of a manual research effort.

If you price against Allegro.pl, see our Allegro scraper, read about e-commerce data collection, or send us your EAN list.