Half or Nothing
50% off or it doesn't make the page.
Published 2026-08-30 · Updated 2026-09-15 · Written by Checarlos Cruz · 13 min read
Most online price drops aren't actual discounts. They're a number set high on purpose so a lower number looks generous next to it. You've felt the pattern before: a price on Amazon or eBay that moves without warning, an “only 2 left” banner that's been there for a week, a strikethrough price that never seems to have been real. Telling a genuine clearance from an inflated starting point isn't hard once you know what to check, it's just not something most shopping sites want you to be able to do easily.
This is the checklist I actually use, the same one behind every 50%+ badge on Half or Nothing. I built it for myself first, honestly, I got burned by a fake “original price” on my own purchase before I ever thought about turning it into a filter for other people.
A deal is legitimate when four things check out at once: the “original” price is verifiably real (it shows up in the item's own 90-day price history, not just on the listing itself), the seller has an identifiable track record, the return and warranty terms are stated plainly rather than buried, and the total landed cost (item plus shipping) is still below what the item has actually sold for elsewhere recently. If any one of those four is missing or vague, treat the discount as unproven until you check it, not as fake outright.
In This Article
A high “original price” changes how you judge everything under it, that's the whole point of putting one there. In e-commerce, that number is something a retailer set, not a historical fact. Manufacturer's Suggested Retail Price (MSRP) gets inflated all the time, specifically so a 50% or 70% badge looks bigger next to it. If the anchor is high, whatever price sits under it feels like a win, even when it isn't one.
Marketplaces actually track two different numbers here: a “List Price” (the highest figure a manufacturer suggests) and a “Typical Price” (the median over the last 90 days). A misleading listing leans on the List Price and ignores the Typical Price entirely, which is how you get a strike-through number with no real transaction history behind it.
Give it a 90-day window before you trust a discount. Plenty of items sit “on sale” 100% of the time, that's not a discount, that's just the price. Check the same product across a couple of independent retailers: if they all show the same “discounted” price, that's the actual market value, not a deal. And watch for a price spike right before a sale, retailers sometimes raise the number 2 to 4 weeks ahead of a big event specifically so they can advertise a bigger drop without actually lowering the real cost.
A ghost discount is a sale that never ends. It leans on “limited time” pressure with a countdown that quietly resets every 24 hours, code doing a job, not a reflection of real inventory. If a deal is always available, it isn't a deal, it's just the baseline price with a clock glued to it. A real drop means the current price is meaningfully below the 90-day median, not sitting right at it with a badge slapped on. This is exactly why there's no countdown timer anywhere on Half or Nothing, I don't want to sell you urgency, I want to show you the number.
You can't tell a real deal from a banner. It has to come from the numbers. For anything you're seriously considering:
Artificial scarcity is the main trick used to short-circuit this process, so the FTC's own guide to safe online shopping is worth reading once. A real deal survives the five minutes it takes to check it. If a listing is priced well below the real floor but comes from a third-party seller with zero history, that's a different risk calculation, not a reason to skip the check.
Each of these is testable in under a minute, not a matter of opinion:
Evidence of each failure mode is common enough to be worth naming directly: inflated list prices that never appear in real transaction history, listings with no stated return terms at all, seller accounts with a rating pulled from a different product category entirely, and “discounts” that quietly disappear once shipping is added at checkout. Any one of these on its own is a reason to slow down, not necessarily walk away.
Browser extensions that overlay price history directly on Amazon and eBay listings are the fastest way to do this. Keepa is the one most people mean when they say this for Amazon. Something like this at scale is exactly what the hourly refresh on this site is doing behind the scenes, pulling items that don't clear the bar back out of the feed.
A sawtooth chart, regular spikes followed by an immediate “sale”, means the MSRP is decorative, not real. The floor (the lowest price in the last year) is your actual target: anything within 5% of it is a genuine deal. As a rule of thumb, I look for at least 15% below the 1-year average before a discount is worth the click.
They hide different things, so the checklist isn't identical.
“Sold by Amazon” means the platform itself owns the inventory. Third-party sellers on Amazon carry more pricing and risk variance. Check the seller rating, anything under 90% over the last year is worth a second look. Watch for review merging (a good rating inherited from a completely different product), and compare the “Coupon” checkbox against the “List Price” separately, stacking both sometimes gets you further than either alone. Ignore the “Lightning Deal” progress bar entirely, it's built to rush you, and check the price against the 90-day median instead. The “Other Sellers” section on a listing sometimes has a better price from a seller who just isn't paying for Buy Box placement.
Listed prices on eBay are aspirational until proven otherwise. Filter by “Completed Items” to see what people actually paid. Check the “Best Offer” history too, if an item consistently sells 20% under “Buy It Now,” the listed price is inflated. “Open Box” electronics are one of the best legitimate sources of a real 90% discount, usually new with damaged retail packaging. “Certified Refurbished” carries a manufacturer warranty, which is worth more than a plain “Used” listing at a similar price.
If something's listed at 50% off, it's worth asking why: often the product is being phased out and the seller is clearing inventory ahead of a new SKU. A real 50% off deal sits at least 30% below the 90-day average. If the current price is only 5% under the historical median, that 50% badge is doing more work than the actual price is.
Here's what that actually looks like side by side: a real 55% off retinol serum right now, next to a similar serum at its normal price. Neither is fake, one's just a genuine markdown and the other isn't.
SkinMedica Retinol Complex 1.0 Anti-Aging Serum, 1 oz
$43.20$96.0055% OFF
Half or Nothing may earn commissions on qualifying purchases.
CeraVe Resurfacing Retinol Serum, 1 oz
$10.99 regular price
Half or Nothing may earn commissions on qualifying purchases.
An automated scraper working purely on volume will happily include expired or dead listings, it has no way to know better. That's how you end up clicking a “deal” that decayed hours ago. A person checking the feed catches that. High-frequency updates plus a real set of eyes is what actually prevents it, which is why the hourly refresh exists in the first place, every link active, every price checked against current Amazon and eBay data.
Half or Nothing is this checklist running as the actual filter for every listing on the site. No hype badges, no fake urgency, just the 50% line applied the same way every time. That line isn't negotiable for me, I'd rather post fewer deals than water down the one rule that makes this site worth your time.
Concretely, that means every listing on this site already passes the six-signal check above before it's published: the price is checked against real transaction history (not just the seller's own claimed original price), the retailer or seller is identifiable on the source listing, and nothing with a buried or missing return policy makes the cut. See how we verify every deal for the full methodology, and what actually counts as real 50% off for how the threshold itself is calculated.
A deal that's real at 9am can be gone by 10am. A static list can't keep up with that, so the feed rebuilds every hour instead of once a day. Listings that don't clear the bar against the 365-day floor get pulled, no exceptions for a slow news day.
Amazon's own affiliate API stopped working for accounts like mine (a flat 403 on every request), and the site actively blocks scrapers, so there's no honest way to automate this part. Every Amazon Pick on this site is something I personally checked and added, one at a time. It's slower, but it means I'm never showing you a number I can't stand behind.
Check the 90-day price history. If the “original” price is higher than anything actually charged in the last three months, it's an inflated anchor. Amazon's own “Typical Price” label reflects the real 90-day median, if the discount is calculated off a List Price that hasn't been active recently, the savings aren't real.
Use eBay's own “Completed Items” filter to see actual realized sale prices, that's the most reliable source since it's the platform's own transaction data. External SKU trackers can add historical trend lines on top of that. Avoid anything that only shows the current listing price with no history behind it.
No. It's only real if the starting price actually matches the 90-day median. Plenty of retailers raise the price right before a sale specifically to manufacture a bigger percentage drop. I look for at least 30% below the 1-year average before I trust a 50% badge.
Because pricing on Amazon and eBay moves in real time, and any aggregator (including this one) can lag behind that by a few minutes between refreshes. That's exactly why Half or Nothing refreshes hourly instead of daily, to keep that gap as small as it can be.
An aggregator filters real, live marketplace listings for actual price drops. A coupon site hands out codes that frequently fail at checkout. I don't run a generic coupon database, on eBay the discount is always already baked into the listing price, and on Amazon, when a real coupon exists, I find it, verify it, and add it myself instead of pulling it from a scraped feed.
Every listing is hand-picked from real Amazon and eBay inventory, the feed refreshes hourly to drop anything expired or repriced, and nothing with an inflated MSRP or a sketchy seller rating makes the cut.
Usually, yes, once you check the seller's return policy. “Open Box” is often a genuine 90% discount caused by damaged retail packaging, not a damaged product. “Certified Refurbished” goes a step further with a manufacturer warranty attached, which is worth more than a plain “Used” listing at a similar price.
Assume it's a timing gap, not a scam. If the marketplace price is higher than what a deal site shows, the discount has probably already expired since your last refresh. Don't buy off the stale number, check the timestamp and confirm the live price first.
This same six-signal check is what runs behind every category page on the site, including discounted electronics and gadgets and discounted fashion clothing, where it's applied to live inventory instead of a hypothetical listing.
Every pick we've pulled at 50% off or more, in one place.
Visit our Amazon store ›
How to verify a jewelry discount is real, what return and authenticity terms actually protect you, and when outlet, resale, and marketplace pricing is genuine versus repackaged markup.
2026-09-15
Every verified statistic on fake discounts and false reference pricing, sourced to the regulators, researchers and laws that produced them.
2026-09-13
The difference between signal and noise in deal aggregation, and why Half or Nothing runs on data instead of urgency banners.
2026-08-30