We checked 45,776 Prime Day prices. Most never went on sale at all.
Amazon Prime Day 2026 ran June 23–26. We pulled the real price history of 45,776 products — across electronics, baby, pet, tools, fashion and more — and compared what they cost during the event to what they cost the week before and over the previous three months. The short version: most "deals" weren't.
This study draws on Thrifle's continuously-updated database of Amazon product price histories — the same first-party data behind our free price-check tool. It covers the 45,776 products in our tracking dataset, not all of Amazon: there were certainly deals we didn't observe. Amazon itself describes Prime Day as "millions of member-exclusive deals worldwide" (aboutamazon.com) — our study measures how widely that discounting ran across a large tracked slice of the store.
Where this study fits: this is our first, deliberately broad study — it measures every product we track, not only items claimed as deals, because Prime Day is marketed as a storewide sale. Readers rightly asked the follow-up question: how good were the actual deals? That's part two — The Prime Day Verdict, a more specific, deals-only audit, now live.
Are Prime Day deals real? In an audit of 45,776 Amazon products, only 19.6% were genuinely cheaper than their normal 90-day price during Prime Day 2026. Just 7.8% beat the price from the week before, and 12.1% actually cost more during the event. The real savings were concentrated in electronics and on big-ticket items.
"Up to 70% off" is the kind of phrase that sells a sale. The trouble is that a percentage off only means something if you know what it's off of — and Amazon, like most retailers, gets to pick that reference number (the "list price," or MSRP). So we ignored the marketing entirely and went to the one thing that can't be staged: the actual price each item has charged, day after day, for the past year.
Read that way, Prime Day looks a lot less generous. Out of 45,776 products, only 19.6% were meaningfully below their own typical price during the event. More were raised than were genuinely cut to a fresh low. And the deals that were real tended to be modest — the median genuine discount was about 11.1%, not the eye-popping numbers on the banner.
8 things we learned
- Most "deals" aren't. Only 19.6% of products were genuinely below their normal price. The other ~80% were the usual price wearing a sale sticker.
- Beating last week was even rarer. Just 7.8% were cheaper than they'd been the week before Prime Day.
- 1 in 8 went the wrong way. 12.1% of products cost more during Prime Day than the week before — and those higher prices were still there after the event ended.
- Electronics is the one category that delivers. 13.5% cheaper and nearly 30% real cuts — about double everything else.
- The cheaper the item, the faker the deal. Sub-$10 products were cheaper just 5.8% of the time; $250–$500 items, almost 17%.
- When a deal is real, it's small. Half of the genuine discounts were under 11.1%. Only the top 10% beat ~27.9%.
- Real deals do exist — they just hide. A handful of laptops, monitors and baby gear hit honest lows. You'd never spot them without the price history.
- Shoppers felt it. Verified-shopper spending fell 17% and satisfaction slid from 68% to 59% year-over-year (Numerator). The numbers explain the mood.
Out of every 100 products on Prime Day…
Compared with the price the week before the event, here's where all 45,776 products actually landed.
Where the real deals were (and weren't)
Prime Day isn't one event — it's a different event in every department. Electronics is the only category where a meaningful share of products were truly cheaper than the week before. In pet, baby, tools and fashion, the typical product's price barely flinched. The dotted line is the all-category average.
Share genuinely cheaper than the week before, by category: Electronics 13.5%, Automotive 8.6%, Baby Products 8.1%, Pet Supplies 7.4%, Tools 7%, Health 6.6%, Clothing 6.5%, Industrial 6.1%.
The cheaper the item, the faker the deal
This is the pattern that surprised us most. The under-$10 impulse buys that fill a Prime Day cart were almost never genuinely cheaper. The real savings — when they happened — were on the expensive stuff people actually research: laptops, appliances, big baby gear.
Share genuinely cheaper than the week before, by price band: $250–500 16.9%, $500–1k 14.2%, $100–250 13%, $1k+ 11.1%, $50–100 9.5%, $25–50 7.4%, $10–25 6.6%, $0–10 5.8%.
The deals that went the wrong way
We can't read Amazon's mind, so we won't claim intent. But the data is the data: 12.1% of products were more expensive during Prime Day than the week before, and we only counted the ones where the higher price stayed after the event — so these aren't momentary glitches or dynamic-pricing blips.
Case study: Apple Watch SE 3 price inflation
The starkest example in our sample was the Apple Watch SE 3, which went from $249 the week before to $369 during the "sale" — a 48% jump that was still $369 days after Prime Day ended. If you were watching that model and bought during the event, you paid the worst price of the month.
Tap any product to see its real price history on Thrifle.
To be fair: the real deals were real
None of this means you should skip Prime Day. It means you should aim. A real minority of products hit honest lows — and some were excellent. These are genuine: each price below was meaningfully under the item's normal selling price, verified against its own history.
Even on big-ticket items — the place most people assume Prime Day must be a trap — a few were the real thing:
What shoppers actually felt
Our price data lined up neatly with how people reacted. According to Numerator's verified-shopper tracker, average Prime Day household spending fell to $105, down 17% from a year earlier, and the average order dropped to $47. Satisfaction slid from 68% to 59%, and only 41% of shoppers said Prime Day was their main reason for buying on Amazon — down from 52% in 2025.
Verified shoppers "viewed this year's deals less favorably than in prior years as inflation grows," Numerator reported mid-event.
The trade and tech press came to the same place. Coverage shifted from "best deals" to "deals to skip," and one widely-shared write-up of a personal price-tracking experiment was blunt enough to be titled, simply, "We've Been Played." When the people running the sale's own numbers and the people shopping it both cool off in the same year, it's usually because the math stopped working.
How we actually measured this
Every product carries a complete price timeline — an event log of price changes, not a daily snapshot. So a product that didn't change price during Prime Day has no data point on those dates, even though it absolutely had a price. Naively matching on the exact dates would have silently dropped ~85% of the catalog. Instead we use carry-forward: the price "in effect" on any given day is the most recent change on or before it. That recovers a real Prime Day price for essentially every product.
Distribution of the genuine discounts. Among the products that were cheaper, discount depth was right-skewed: 25th percentile 6.2%, median 11.1%, 75th percentile 19.9%, 90th percentile 27.9%. In plain terms, the typical real deal is a single-digit-to-low-teens discount, and the headline-grade cuts are a thin tail.
Limitations, stated plainly. The sample is popular, well-tracked products — roughly what people actually shop — not a random draw from Amazon's full catalog, and it includes third-party marketplace listings whose prices move for reasons beyond Prime Day. Prices are read at change-event granularity, so a sub-hour lightning deal on an obscure item can fall between observations. The aggregate percentages are stable across tens of thousands of products; individual examples are illustrative.
Want the raw data behind this study?
All 62,000+ products as a clean CSV — each item's real price before, during, and after Prime Day, plus the honest-discount flags. Built for researchers, journalists, and analysts who want to run the numbers themselves. The report and its charts stay free to read and cite; this is the underlying data, and it's a paid download.
Stop guessing whether a "deal" is a deal
The whole problem with a sale is that the discount you see is measured against a number the seller chose. The fix is boring and it works: check the item's real price history before you buy. Paste any Amazon link into Thrifle and we'll tell you, in one line, whether today's price is genuinely low or just dressed up — based on its actual history, not a list price.
Check a price free →Frequently asked
Methodology and figures © Thrifle, 2026. Based on Thrifle's Amazon price-history database, covering 45,776 products through Prime Day 2026 (June 23–26). Shopper sentiment and spending figures via Numerator's Prime Day 2026 tracker. Charts and figures are free to republish with a link back to this page. Want every product's numbers? Get the full per-product dataset (CSV, 62,000+ products — $45), or email [email protected]. Data updated June 29, 2026.
Corrections & updates
July 6, 2026 — headline revised. This study originally ran under the headline "Here's how many 'deals' were real." Readers correctly pointed out that the word "deals" implied we had audited claimed deals only — when this study deliberately measures our whole tracked catalog, because Prime Day is marketed store-wide. Same data, same numbers, clearer words: the headline now states the actual finding (most products never went on sale). The deals-only question got its own study: part two, The Prime Day Verdict.
Also added: Amazon's own description of the event for context, an explicit note on dataset scope, and the disclosure below. We publish corrections openly — being checkable is the whole point of this research.
Disclosure & independence: Thrifle participates in Amazon's affiliate program — purchases through our links may earn us a commission. We are not here to bash Amazon; we exist to be an impartial layer of truth between shoppers and the stores they buy from. We publish what the data says either way — including when it favors Amazon (in part two: 87% of the deals Amazon promoted were genuine).