We saved every price for 44 days to find out
How often sale prices actually change
A discount percentage compares two numbers the seller chose. It says nothing about whether the price moved. So we stopped guessing and kept a record: 190,465 daily price readings, taken from August 7, 2026 to September 19, 2026.
Nearly three quarters of the listings we watched for a month or longer held exactly the same price the entire time.
Get the sale price dataCSV, 4,157 rows, free to reuse with credit.
44 days
of daily observations, with no gap in the record
190,465
price readings across 6,609 discounted listings
72.3%
of the 4,157 listings we tracked for a month or more never changed price
The part we did not expect
The bigger the discount, the less the price moved
We grouped every tracked listing by the discount it advertised, then asked how many in each group ever showed a different price. The pattern runs one way the whole distance. Listings shouting the loudest markdowns were the ones whose prices sat completely still.
20 to 39 percent off
1,340 of 2,168 never changed
61.8%
40 to 59 percent off
710 of 999 never changed
71.1%
60 to 79 percent off
872 of 908 never changed
96%
80 percent off or more
82 of 82 never changed
100%
All 82 listings advertised at 80 percent off or more showed one price, every day, for the whole time we watched them. Across everything that held its price, the average advertised discount was 44.9 percent, and 1,462 of those listings claimed half off or better while never moving a cent.
A shopper reading those labels would reasonably think the price had just dropped. In this record, the number on the label and the behaviour of the price were pulling in opposite directions.
What this record can and cannot show
Gazoobie saves one row per tracked offer per day. For this study we took every row up to September 19, 2026, kept the 4,157 listings that appeared on at least 30 separate days, and counted how many distinct prices each one showed. Every one of the 190,465 rows passed a validity check before it was counted, so nothing was dropped to make the finding tidier.
These are feed observations, not checkout tests
Each row is what the retailer supplied through its affiliate product feed on that date, saved once a day. We did not add the item to a cart. A feed can repeat an older price, and shipping, tax, member pricing and seller changes are outside the record.
The sample is listings that were already discounted
Both feeds carry sale inventory, so every listing here arrived with a discount attached. That is why we report nothing about how much of retail is on sale. The question this record can answer is narrower and more useful: once something is advertised at a discount, does its price move?
Two feeds, not the whole market
The CJ and Sovrn affiliate product feeds are the two sources that state a per-product price. Coupon offers without a price are absent by design. These findings describe the listings we track, and we are not going to stretch them into a claim about every shop online.
A month of history cannot show a record low
A price that held for six weeks may have been lower in March. We publish what we observed between the two dates above and nothing either side of them, which is also why no figure on this page is described as a lowest-ever price.
This page reports a pattern across thousands of listings. It is not an accusation against any shop, and we have deliberately published no retailer names with it. A single listing that held its price proves nothing on its own, which is exactly why we counted 4,157 of them. More on the method behind the record: how we verify deals.
Cite this study
The per-listing results are free to reuse under a Creative Commons Attribution 4.0 licence. Quote the figures, chart them, check our arithmetic. Journalists and researchers who want a cut of the data we have not published here can ask us for it.
Suggested citation
Gazoobie. "How Often Sale Prices Actually Change." September 20, 2026. https://gazoobie.com/research/how-often-sale-prices-change
Questions about the findings
How often do sale prices actually change?
Less often than the label suggests. Of 4,157 discounted listings that Gazoobie observed on at least 30 separate days between August 7, 2026 and September 19, 2026, 3,004 of them (72.3 percent) showed the same price on every single observation.
Does a bigger discount mean a better deal?
In this record it meant the opposite. Listings advertised at 20 to 39 percent off changed price 38.2 percent of the time. Listings advertised at 80 percent off or more changed price in none of the 82 cases observed. The deeper the advertised discount, the more likely the price behind it sat still.
Does this mean retailers are running fake sales?
That is not a claim this data can support, and Gazoobie does not make it. A price that holds steady has many innocent explanations, including a genuine long promotion or a feed that repeats an unchanged value. What the record does show is that the size of a discount label tells a shopper very little about whether the price recently moved.
Can I use this data?
Yes. The per-listing results are published as a CSV under a Creative Commons Attribution 4.0 licence. Cite Gazoobie and link to this page.
This is why we throw most deals away
A discount label is not evidence. The Real Deals board is what survives after we stop treating it as evidence.
See the Real Deals board