We ran the same test on 284 Korean brands. The gap is global.

⏰ The short version

• Same method, bigger sample: 284 Korean DTC brands, 50 questions each.
• 65.5% (186 brands) scored zero. Average: 0.65 out of 50.
• Beauty was worst at 86% zero — despite being the category that publishes the most content.
• Food scored best at 51% zero, and the reason is instructive.
• Two markets, two languages, one pattern: brand pages describe the brand; AI is looking for answers.

Why look at a second market

We started with 20 US indie brands and found 90% at zero. Small sample, so the obvious move was to test something bigger. Korea gave us 284 brands with completed diagnostics.

Different language, different platforms — Korea has Naver AI Briefing alongside ChatGPT — and a much larger sample. If the gap were a quirk of one market, this is where it would show.

The result

Questions answered	Brands	Share
0	186	65.5%
1	52	18.3%
2	27	9.5%
3	11	3.9%
4+	8	2.8%

Only 4 brands out of 284 answered 5 or more.

Known, established brands were in the zero group. Recognition and AI citation are unrelated — AI doesn't know who's famous. It checks whether this page answers this question.

The category breakdown is the interesting part

Category	Brands	Zero rate	Average
Beauty / skincare	58	86%	0.36
Supplements	58	74%	0.57
Home & living	57	60%	0.68
Pet	54	56%	0.85
Food & drink	57	51%	0.79

Beauty is worst by a wide margin. Which is strange, because beauty publishes more content than any other category — blogs, long product pages, ingredient explainers.

The problem isn't volume. It's that beauty copy is written to feel something.

"Deeply replenishing moisture." "All-day luminous glow."

Good brand writing. Impossible to cite — it isn't the answer to any question anyone asks.

Food scores best for the mirror-image reason. Storage, prep, sourcing, shelf life — food brands write answer-shaped sentences without trying, because their customers ask practical questions. So AI can quote them.

The same pattern in both markets

Across 304 brands in two markets:

Brand pages answer "who we are." Founder story, philosophy, sourcing standards.

AI is searching for "what's the answer to this." Which sunscreen works for sensitive skin? Grain-free dog treats — which ones?

That mismatch isn't a translation problem. It's a purpose problem, so it reproduces identically in Korean and English.

What English-language brands should take from this

Two things.

One: you're not behind, and neither is anyone else. Both markets are mostly empty. The shelf is short — AI names two or three brands — and most of those slots are unclaimed.

Two: check your category before you assume. If you're in beauty, the odds you're invisible are higher, and the fix is specific: keep the atmospheric copy, but add one factual sentence underneath it. Both can live on the same page.

Honest footnotes

• Korea: 370 brands attempted, 284 completed — the rest failed on crawl blocks or page structure. That failure rate is itself a finding.
• Korea used 50 questions, the US test used 30. Zero-rates compare; averages do not.
• Category labels were assigned by us and may not match how a brand sees itself.
• Judgments were made by a language model. Borderline cases exist where a human would score differently.
• Data as of July 2026.