- New Korean groups look 69% down since 2015. We are not publishing that, and here is what killed it
Sat Aug 22 — Wikidata records 691 Korean musical groups with a formation year. The 2023-25 average is 31% of the 2015-17 average. Then we ran the same count for Japan, the United States and the United Kingdom — and all three fell too.
- In Southeast Asia, a Korean record label is looked up more than Seoul
Fri Aug 14 — We measured 2,238 places in Korea against four Southeast Asian Wikipedias. YG Entertainment scores 111.47 and Seoul 106.54. Twelve companies out-read 520 cities and districts on the median. What this cannot see is a single restaurant.
- Kim Jong-kook is the most evenly read Korean name in Southeast Asia. K-pop groups are the most even category
Fri Aug 14 — We measured concentration — the share of a name's four-country reading that sits in its biggest country. K-pop groups come closest to even at 41.7%. The most concentrated row turned out to be an artefact of Wikipedia, not of attention.
- Malaysia over-reads SISTAR and Ahn So-hee while the region leads with BTS
Fri Aug 14 — Groups, solo musicians, actors, athletes — the Malay Wikipedia takes 7.6% to 8.6% of the four-country reading for each. For luxury and car brands it takes 23%. Four categories in a narrow band, one well outside it.
- Faker and Peanut get a year of Southeast Asian readers in one month. Son Heung-min never does
Thu Aug 13 — Split Indonesian, Vietnamese, Thai and Malay Wikipedia reads by month. All eleven esports players in the panel peak in the same month, and they agree in three of the four years we measured. The footballers never once agree.
- What we called an esports pattern turns out to be one game
Thu Aug 13 — We said Korean esports players in the Southeast Asian Wikipedias all peak in one month. Split by game, 85.7% of them play League of Legends. Korea has more StarCraft II players, and four of them have an article in those editions.
- We tried to check whether the Korean esports pattern was really about Korea. There was nobody to check it against
Thu Aug 13 — Chinese, Swedish and American esports players have articles in the Southeast Asian Wikipedias at roughly the rate Korean players do. 9.1% of theirs draw enough reads to measure, against 71.4% of the Korean ones.
- Half the Korean actors on a Netflix chart are there for exactly one title
Thu Aug 13 — We tried to test whether debuting younger buys a wider career. The answer did not come out in one direction, so we went looking for why — and found that for 658 of 1,329 actors, the chart only ever lit them once.
- Son Heung-min is read more in Southeast Asia than all 1,023 Korean actors we measured
Thu Aug 13 — We measured 1,972 Koreans with one instrument. Son Heung-min sits above all 1,023 actors. But 22 actors clear 100 against two athletes, and the two groups are read in different countries. Musicians who never acted are not in this panel.
- Vietnam reads Faker and Kim Sang-sik where its neighbours read Son Heung-min
Thu Aug 13 — Four Southeast Asian Wikipedias, twelve months, 926 Korean athletes. Three of the four read almost only football. In the fourth, esports takes 60.6% of the reads and the most-read Korean athlete is not a footballer.
- Undercover took 48 chart places in our Korean panel. It is a Belgian series
Mon Aug 10 — Two works can carry the same English title, and a chart row does not say which one it is. Thirteen titles in our Korean panel turned out to be foreign works. The test that found them is not the name — it is where they charted.
- Korean films leave a Netflix chart exactly as fast as everyone else's. Korean series do not
Mon Aug 10 — Across 210,355 finished runs, a Korean film holds a country chart for 1.96 weeks on average and every other film for 2.02. Korean series run 3.7 weeks against 2.71 — and removing each group's biggest title barely moves it.
- We wrote that a market has one Korean slot. Measured, two-thirds of that was arithmetic
Mon Aug 10 — When a Korean title reaches a market's top three, the other Korean titles there fall by 0.403. Non-Korean arrivals cost 0.013. That looks decisive until you notice the two groups start from different heights.
- Two in three Koreans have had their fortune read. Here is what they are actually looking at
Mon Aug 10 — The animal year and saju are two different things, and Koreans use both. We counted the animal years of 1,047 Korean actors who reached a Netflix chart: chance would put 87.3 in each sign, and the count runs 76 to 103.
- A Korean series that opens at number one lasts six weeks. One that opens at ten lasts one
Mon Aug 10 — Opening position is the only fact anyone has on the first Monday. Across 8,324 Korean series runs it moves expected length by 0.56 weeks a step — against 0.44 for other series, 0.33 for Korean films and 0.22 for other films.
- Korean series are widest in week one only 20% of the time
Mon Aug 10 — Hold run length fixed and most of that gap closes — among titles that lasted five to eight weeks it is 26.6% against 28.6%. Among those that lasted nine weeks or more it does not close: 9.7% against 24%.
- On Korea's own chart a Korean title needs no other market. In Estonia the median one had 77
Mon Aug 10 — For each of 93 markets we took the median breadth of the Korean titles that reached it. Korea's own chart: 2. Estonia, Slovenia, Switzerland: 77, 76, 72. Most of that gap is how often a market plays Korean titles at all — but not all of it.
- At number one a Korean title has a 1.2% chance of being gone next week. At number ten it is 70.2%
Sun Aug 09 — Everything written about a chart is about getting on it. We measured 11,722 departures instead — the position a Korean title held the week before it vanished, against the 10.0% each rank would take if position did not matter.
- Our headline 7.9% is not one place a week — it is nothing, and then several at once
Sun Aug 09 — Korean titles hold 7.9% of Netflix country chart places. Spread evenly that would put at least one on 56.3% of country-weeks; the real figure is 42.0%. Empty weeks run 58.0% against 43.7%, and crowded weeks 4.0% against 0.6%.
- Korean drama costs rose 27-fold. The chart concentration the argument predicts did not appear
Sun Aug 09 — The case that Netflix gained leverage over Korean drama runs through cost: $360,000 an episode in 2015 against $9.8m in 2024. We hold no budgets. We do hold the charts, and from 2022 to 2025 concentration fell rather than rose.
- What a chart position does not tell you — and why it tells you least on the chart Korean titles sit on
Sun Aug 09 — Netflix attaches hours viewed to 10,720 global chart rows and to 0 of its 499,180 country rows. Knowing a rank narrows the possible hours by 47.3% on the non-English series chart — least of the four, and the one Korean titles sit on.