Korean pop culture, in numbers

How lopsided is attention on Korean acts? 24 of 774 hold half of it

A Korean forum thread this week asked why lesser-known idol groups are said to lose money on every event they play. We cannot answer that — we hold no fee, ticket or revenue data for any act here, and this page makes no claim about anybody's finances. What we can do is measure the shape of attention, and the shape is why the question exists.

In the single month 2026-08, 774 Korean music acts drew 2,670,482 reads on English Wikipedia between them. 24 of those acts hold half of it — 3.1% of the list. The lower half of the list, all 387 acts together, holds 3.1%.

Counting down from the top

The most-readShare of the month's reading
Top 1 of 7749.7%
Top 3 of 77418.0%
Top 5 of 77422.9%
Top 10 of 77433.5%
Top 20 of 77447.0%
Top 50 of 77464.6%
Top 100 of 77478.4%
Top 200 of 77489.8%
Reads for the median act519

One act — BTS — accounts for 9.7% on its own. The median act on the list was read 519 times in the month, which is 0.0194% of the total.

The most-read 25

#ActFoundedReads in 2026-08Share
1BTS2013260,0859.7%
2Blackpink2016140,8815.3%
3Stray Kids201780,1993.0%
4BigBang200670,1342.6%
5Ateez201861,2922.3%
6Le Sserafim202258,9162.2%
7Aespa202057,9542.2%
8Twice57,7712.2%
9Enhypen202054,3192.0%
10Illit202453,0332.0%
11Seventeen201547,7521.8%
12Girls' Generation200745,7641.7%
13NewJeans202242,8181.6%
14Babymonster42,3111.6%
15Red Velvet34,2061.3%
16Meovv31,9381.2%
17Ive202130,9261.2%
18NCT 127201630,7721.2%
19Tomorrow X Together201927,6651.0%
20Exo201225,8051.0%
21Rescene202424,0650.9%
22Mamamoo201422,8950.9%
23NCT201622,1070.8%
24Iz*One201820,9950.8%
252PM200820,3890.8%

Founding year is blank where Wikidata does not record one. It is shown because the list is not sorted by it and the reader can see that for themselves — the top of this table spans three decades of debuts.

Two separate counts, and where they disagree

A single month could be a fluke, so we ran the same measurement on a different act set, over a different window, collected separately: 380 acts, act-page reads over twelve months. If the lopsidedness were an artefact of one month, the two would not match.

The most-readOne month, 774 actsTwelve months, 380 actsGap
Top 19.7%9.8%0.0 points
Top 318.0%18.9%0.9 points
Top 522.9%25.1%2.1 points
Top 1033.5%36.4%2.9 points
Top 2047.0%49.9%3.0 points
Top 5064.6%70.0%5.5 points
Top 10078.4%84.9%6.5 points
Top 20089.8%95.8%5.9 points

They agree closest at the very top — top 1 comes out 9.7% against 9.8% — and drift apart further down, by up to 6.5 points at top 100. That drift is expected and we would rather explain it than hide it: the twelve-month set contains fewer acts, so any given count of acts is a larger slice of a shorter list. What does not change is the finding. Both counts put roughly a tenth of everything on one act and about half of everything inside the top twenty.

Stated as a single number, the one-month spread has a Gini of 0.8183 and the twelve-month spread 0.7545. 0 would mean every act is read exactly as much as every other. 1 would mean a single act takes all of it.

🔴 What this does not say

The thread that sent us here was about money, and that is exactly where a page like this can mislead. So, plainly: a Wikipedia read is not a ticket, a stream, a fee or a won. We have no revenue data for any act on this page. An act can be read rarely and tour profitably, and an act can be read constantly and lose money.

What the numbers do support is narrower and still worth having. If you are the median act on this list, the number of times your page was opened in English in a month is 519 — page-opens, not people, since one reader returning three times counts three. Whatever the economics of an event are, they are being attempted at that level of outside recognition, not at BTS's. The gap between those two conditions is what the forum thread was reaching for, and it is the one part of it we can put a number on.

It is not a measure of popularity. A read is somebody looking a name up, which is not the same as liking, buying a ticket, or streaming.

It is not a measure of money. We hold no fee, ticket or revenue data for any act on this page and make no claim about any act's finances.

It is not a ranking of quality, and an act low on this list is not being called unsuccessful. The shape of the whole is the point, not any one row.

Acts whose reading could not be retrieved are left out, not counted as zero. The main count already dropped 73 of 847 acts for that reason.

Where else this leads

The same lopsidedness turns up when we count titles instead of acts: Korean titles do not spread evenly across Netflix country charts either — they arrive nowhere for weeks and then in several places at once. And for what happens to an act's reading after its best month, the median act loses half of it in two months, which is the other half of why the middle of this list is where it is. For where a given act's reading lands once it has any, see the group page or the member pages.

14 seconds, with sound

Fourteen seconds, read aloud: seven hundred seventy four Korean acts drew two point six seven million English Wikipedia reads in the single month of August twenty twenty six, and twenty four of those acts hold half of it. The bottom half of the list — three hundred eighty seven acts together — holds three point one percent, the same share as the top three point one percent. The act in the middle was read five hundred nineteen times in the month. These are encyclopaedia pages being opened, not tickets, not streams and not money, and we hold no revenue data for any act on this page. Free to repost with the address on it.

All our short films — each one built from a table on this site

What we wrote from this data

Sources

Main count — 774 Korean acts, English Wikipedia reads in the single month 2026-08. Re-counted from kcw-group-afterlife.json; no new collection.

Check count — 380 acts, act-page reads over the twelve months 2025-09 ~ 2026-08. Re-counted from kcw-member-vs-group.json; a different act set, a different window, a different collector.

Reads — Wikimedia Pageviews REST API, English Wikipedia, all-access, agent=user (bots excluded). Acts — Wikidata SPARQL. Neither well was drawn again for this page; both counts are recomputed from data we already had, so the two figures cannot have drifted apart through collection. Main count collected 7 September 2026, 10:15 KST, check count 2026-09-07 23:58 KST. Recounted 2026-09-08 11:44 KST.