How Deal Score works
Transparency, on purposeRetailers tell you something is on sale. Daily Deals Detector tells you whether it's actually a deal — by checking the price against its own history, the product against its reviews, and the seller against their record, then turning that into one plain decision:
How we find deals
We pull real listings from retailer and marketplace feeds (Amazon via Keepa's price-tracking data, Woot, eBay and others), plus trending products matched against the same sources. Every listing is checked for a usable price before it's stored. Reseller "mystery box" lots, multi-quantity variants whose price we can't pin to a single unit, and titles that fail a brand-safety check are filtered out before they can appear anywhere.
What goes into the score
Eight factors, each with a maximum weight. A factor only counts when we actually have the data for it — a missing factor is reported as missing, never guessed.
| Factor | Max | What it measures |
|---|---|---|
| Historical price | 25 | Where the current price sits against the item's 90-day (or 30-day) average price. |
| True discount | 20 | How far below its typical price it is — not how far below a retailer's chosen "was" price. A retailer-claimed discount with no history behind it counts for at most half. |
| Lowest-price proximity | 15 | How close the price is to the lowest we've ever recorded for it, and when that low was. |
| Product quality | 15 | Star rating weighed by review depth, so five stars from two reviews doesn't beat 4.6 from three thousand. |
| Seller trust | 10 | Who's selling: Amazon or Woot directly, an Amazon marketplace seller, or an eBay seller's public feedback score. |
| Competitor price | 5 | Whether other retailers we track sell the same product for less. Rare today — we only compare when identifiers (UPC/EAN/ASIN) match exactly. |
| Community | 5 | Real votes from signed-in members. Counted only when votes exist. |
| Availability | 5 | Whether it's in stock, how often it goes out of stock, and whether the offer has an end date. |
Confidence — and why a thin score can't be a BUY NOW
Confidence is the share of that 100-point evidence weight we actually had for a deal. If we only know a retailer's own "was" price and who's selling, confidence is low, the score is pulled toward the middle, and the verdict is capped at Good deal — no matter how big the claimed markdown. Deals with real price history, product reviews and a known seller are the only ones that can reach Buy now. Every deal page shows the factor table and which evidence was missing.
What "price history" means here
For Amazon items we use Keepa's recorded price series: 30/90/180-day and 1-year averages, the lowest and highest price on record and when they happened. For every deal we also record our own price observation daily while it's live. Where we don't yet have history, the deal says Limited price history— we never invent one.
"True Discount" vs. what the retailer says
A retailer's "60% off" is measured against a reference price the retailer chose. Our True Discount is measured against what the item has typically sold for. When both are available you'll see them side by side. We use factual language only — a gap between the two is reported, not accused.
Editor-verified
A deal marked Editor-verified was opened and checked by a person. It isn't routine on every deal yet, and it is shown only when it actually happened.
How we make money — and what it can't buy
Some outbound links are affiliate links; we may earn a commission at no cost to you. Clicks go straight to the merchant: no browser extension, no overwriting anyone else's links.
Merchants can pay for visibility. They cannot pay for a better Deal Score.
Any paid placement will be labeled SPONSORED and, if scored, carry the note "Deal Score independently calculated." A paying merchant can still receive a 43/100 and a SKIP if the data says so. There are no sponsored placements on the site today; this is the rule they will be held to.
Our promises
No cookie hijacking
No extension, no overwriting anyone's affiliate links. Clicks go straight to the merchant.
No invented history
Averages, lows and highs come from recorded prices. Missing data is labeled missing.
Verdicts you can read
Buy / Wait / Skip in words, with the factor table behind every score.