Dealblob is a fully automated deal aggregator. Instead of editorial curation, every deal you see passes a hard-coded algorithm. Here's step by step how it works – and where we make money.
A Playwright-based scraper running a real Chromium browser scans Amazon every 6 hours. We simulate human browsing with randomized delays (2–4s per request) so we don't hammer the retailers' servers.
Every product is run through our quality score. Rejected: items with fewer than 10 reviews, a star rating below 3.5, prices under $10, or no real strikethrough price. Typical low-quality clones are filtered out automatically.
We only accept deals with a verifiable strikethrough price. The current price must be at least 5% below the retailer's listed MSRP. Fake markdowns with invented “was” prices get rejected. For long-running products we store price history in our database.
Deals the scraper doesn't see for two consecutive runs (about 8 hours offline) are automatically hidden. This prevents “deal expired” clicks – nothing is more frustrating than chasing a price that no longer exists.
What survives lands on the site – including the retailer's original star rating, review count, discount percentage, and where available a price-history chart. Each deal links directly to the retailer.
Amazon deals: We participate in the Amazon Associates Program. If you click an Amazon link and buy the product, we earn a small commission from the retailer. Your price stays the same. This commission funds hosting, scraper servers and ongoing development.
Amazon only: Until August 2026 we also listed deals from Target, Walmart, Best Buy and others. We never earned anything on those, because there is no affiliate programme for them. Since 25 August 2026 we stick to Amazon.
No paid placements: No retailer can buy a spot in our listings. Sort order is purely driven by the quality score. Ranking and commission are decoupled.
Honesty is our whole promise – so we are equally open about how the content on this site is made.
The price and deal data comes solely from our own automated collection: we read the price, list price, rating and availability straight from the retailer and re-check every active deal every two hours. Those numbers are measured, not estimated.
The prose – product verdicts, guides and comparisons – is generated with AI assistance from exactly that data. It is not individually reviewed by a person, and we label it visibly on each piece. We do not physically test products and never claim otherwise.
The price verdict on every deal page ("lowest in 90 days", "has been cheaper") is pure arithmetic from the price history we recorded – no AI involved.
