Do high star ratings mean honest ads? We checked 10,000 ad lines
We split 10,406 ad lines by the star rating of the product they sell. Products rated 4.7 and up made weak claims about as often as products under 4.5.
A cream rated 4.8 stars feels like a safe bet, and it's tempting to assume its product page is honest too. It isn't a safe assumption. We split ad lines by the star rating of the product they sell, and the share of weak claims barely moved.
A star rating collects how satisfied users felt. Whether an ad line is true is a separate question. That answer lives in the ingredient list and the test conditions.
How did we split them?
As of September 27, 2026, we'd finished checking 2,049 products. We took the 796 that carry a star rating on a Korean review app with at least 50 reviews. The ratings were collected between August 27 and September 10, 2026. Of those, 764 products had ad lines, 10,406 lines in all, and we split them into three rating bands.
Each ad line already carries a check result, and we counted those results within each band. We didn't rank individual products. Only the share within each band.
"Weak claims" here groups four results together: overstated ad line, no source confirms it, says nothing, and sources disagree.
What does each rating band look like?
| Rating band | Products | Ad lines | Backed by evidence | Weak claims |
|---|---|---|---|---|
| 4.7 and up | 60 | 701 | 14.8% | 39.1% |
| 4.5 to under 4.7 | 344 | 4,547 | 10.1% | 40.4% |
| Under 4.5 | 360 | 5,158 | 12.1% | 37.8% |
Weak claims run at roughly four in ten in every band. The gap between bands is under three percentage points.
Backed by evidence doesn't move in one direction either. The top band is a little higher at 14.8%. The middle band is lowest at 10.1%, though, and the bottom band sits at 12.1%. A higher rating doesn't mean a more honest ad, and neither does a lower one.
The top band holds only 60 products. Swap a handful and its percentages could shift. So the thing to read in this table isn't the small gaps. It's that the three rows look alike.
Why can't star ratings filter ads?
They measure different things. A rating gathers experience: how a product feels going on, how it smells, whether it broke someone out. Checking an ad line asks whether a percentage on the page comes with a lab, a date window and a head count.
A cream that feels lovely can still attach nothing to a line about how many hours it hydrates. And a product that documents every claim carefully can lose stars over its texture. The two travel separately.
Ratings are also bunched tightly. Of the 796 products, 755 sit at 4.0 or higher, and the middle value is 4.5. When nearly everything is in the fours, stars alone don't separate products as much as they seem to.
What doesn't a star rating tell you?
What's inside. Ratings don't read ingredient lists. Whether there's fragrance, whether there's an essential oil, whether the advertised ingredient sits near the top or the bottom: only the list shows that.
How much is inside. Whether the advertised ingredient comes with a printed amount is also a list question. If it's printed, you can compare. If it isn't, all you know is that the name is there.
The conditions behind an ad number. How many people, over how many days, compared against what: a rating says none of it. We walked through how to read those footnotes in when the panel was twenty people.
So how should you use star ratings?
Keep them, but give them a job. Ratings and reviews are good at feel and at patterns of irritation. Whether an ad line is true is a separate check, done on the ingredient list and the test conditions.
Use ratings for: texture, scent, stickiness, and whether reports of irritation keep coming up.
Use the ingredient list for: whether the advertised ingredient is actually there, whether an amount is printed, and whether anything you avoid is on it.
Use the ad line itself for: whether a number has a lab, a date window and a head count in brackets next to it. If it doesn't, leave that number out of the decision. For lines that lean on dermatology or derma, see what "dermatologist recommended" means on a Korean label.
Common questions
Should I avoid products with low ratings?
Check why they're low first. If it's taste, like scent or texture, it may not matter to you. If irritation or breakouts keep coming up, that deserves weight.
Are ratings with more reviews more reliable?
More reviews make a rating less jumpy, which is why we only counted products with at least 50. No number of reviews stands in for the evidence behind an ad line, though.
Does this hold for ratings on every shopping site?
We looked at one review app's ratings, collected between late August and early September 2026. Other sites may spread differently. That a rating and an ad check measure different things holds anywhere.
If a product you're eyeing has great stars but you can't tell how far its page can be trusted, send us a label photo or the product name. We'll read what's behind its ad lines with you.
