Home › Methodology

How we pick Korean beauty brands before the US catches on

The short version: we look for brands doing better than their peers in Korea, then keep the ones that are still small in US Google searches. Here's exactly how, what it got right in a backtest, and where it can go wrong.

1. Follow brands in Korea

We follow 62 Korean beauty brands. For each one we look at two signals from Korean platforms: how much people search for it, and what share of beauty shopping clicks it gets in online stores. We compare the last four full months with the year before, so seasonal swings mostly cancel out. Using the share of clicks, not raw clicks, keeps platform-wide changes from looking like brand growth.

2. Keep the top fifth

A brand makes the shortlist if it lands in the top 20% of brands on either signal. Brands strong on both come first. Because this is a ranking, a brand can make the list in a slow year by falling less than everyone else. When that's the case, its page says so.

3. Check the US

We drop brands that are already big in the US, meaning brands whose US Google search interest is 30% or more of COSRX's, a widely known K-beauty brand we use as a yardstick. For the rest we chart US Google Trends on its own, smoothed over four weeks.

4. Check the facts

Every brand page is checked by hand-built research: is it really a Korean beauty brand, can US shoppers buy it from a store that ships here, and is every fact on the page backed by a source. Claims that come from the brand itself are labeled that way. We don't publish efficacy claims.

Backtest: did it work in the past?

We ran the same rule on brands as they stood in 2021, 2022 and 2023 and checked which ones went on to become the fastest-growing fifth of brands in US Google search.

Limits you should know about

• It's a small sample: 24 picks per signal. Treat it as a good hint, not a promise.
• The 2022 starting point was the one we used to design the rule, so it isn't a clean out-of-sample test. The 2023 picks only had about eight months of follow-up.
• Google Trends is relative and rounded. Small brands can bounce around, and US data had a break in spring 2026 that we work around by comparing year over year.
• Some brand names are also everyday words, which can inflate their search numbers.
• Doing well in Korea says nothing about whether a product suits your skin. Nothing here is skincare or medical advice.

What we publish and what we don't

Charts and percentages on this site come from Google Trends and are labeled that way. The Korean platform data is used only to make the picks; we don't republish it.