How we rank
We count how many people really listen to and follow an artist — on YouTube, Spotify and other platforms we have honest data from. These numbers give a popularity score that sets the place — live, with every data update.
What counts is listening NOW: every view loses weight over time — after a year it counts half, after two years a quarter. Each platform weighs as much as its numbers agree with the rest. One viral hit, an old hit, hundreds of tiny uploads or bought numbers do not decide the place.
What helps
- People listening to your best tracks now — fresh views weigh the most.
- New tracks, and older ones people keep coming back to.
- Spotify listeners and YouTube subscribers, when they go together with listening.
- Several strong tracks rather than one.
What does not help
- The number of uploads. Each additional track counts less than the previous one.
- One viral hit. The best track counts at most 2× the rest of the catalogue — viral hits get their own Trending list.
- Bought views. Views without likes and comments only count up to a normal level.
- An old hit with no new listeners. A view from a year ago counts half, from two years ago a quarter.
- No account on a platform never lowers the place — we count what is there.
From numbers to a place
Four examples from a simulation
Each was inserted into the real population of artists (576 artists, as of 8 Oct 2026, Ranking 2.1). Higher on the axis means a better place.
- Steady artist #5440 tracks with 50–300K views, still being played
An even catalogue and loyal listeners — 72 places ahead of the viral hit.
- Old hit #60a 10M hit + 9 tracks from 2 years ago, nothing new
Previously #47. Old views lose weight: of 11M, about 5.6M count today — and half as much each year if nobody comes back.
- Viral #1261 track, 10M views in 30 days
Fresh views count in full, but the best track counts up to 2× the rest. On the Trending list it is near the top.
- Spammer #4601,500 uploads with 100 views each
Quantity gives almost nothing: 1,500 tiny tracks count like one track with a few hundred views.
Why #1 is #1
Today's top three, from real data — what sets them apart, without technical points.
-
59.4M YouTube track views87.6K Spotify monthly listeners94.8K YouTube subscribers
- Strongest side: YouTube — listening to the best tracks. Here the artist does better than 99% of ranked artists who have this source.
- Ahead of #2 Antartlife mainly thanks to YouTube subscribers.
-
64M YouTube track views71.3K Spotify monthly listeners53.3K YouTube subscribers
- Strongest side: YouTube — listening to the best tracks. Here the artist does better than 99% of ranked artists who have this source.
- Ahead of #3 Polo Vibes mainly thanks to listening to the best tracks on YouTube.
- Gap to #1 MusicAiRecords: mainly YouTube subscribers.
-
58.7M YouTube track views174.2K Spotify monthly listeners51.6K YouTube subscribers
- Strongest side: YouTube — listening to the best tracks. Here the artist does better than 99% of ranked artists who have this source.
- Ahead of #4 Disco Fala mainly thanks to listening on YouTube Music.
- Gap to #2 Antartlife: mainly listening to the best tracks on YouTube.
Where the data comes from
- YouTube
- subscribers, channel views, track views, likes and comments — YouTube Data API
- Spotify
- monthly listeners — the artist's public page, once a day
- followers — official API (business and creator accounts)
- Deezer
- fans — the artist's public page, once a day
- YouTube Music
- views of distributed tracks (the “Artist - Topic” channel) — YouTube Data API, once a day; separate from YouTube
- TikTok
- followers and likes — the artist's public profile page (the link on their profile), once a day; followers count as audience (from 30 shared artists), likes are only shown
The ranking place is our own calculation based on YouTube API data and public platform pages.
Glossary of numbers
Every number and label you see in the ranking, on an artist profile, on a track card and on the lists. For each: what it means, where it comes from, how often it changes and whether it affects the ranking place.
Ranking place
- #N “of 402 ranked artists”
-
The artist's place in the main ranking. The number after “of” is all public artists with a place in the main ranking today — without Debuts and without profiles whose tracks we have not measured yet.
- Most likely #a–#b
-
The range of places the artist most likely falls into once data uncertainty is taken into account. How: 400 times we draw each artist's score within its uncertainty (fewer sources, or sources that disagree, mean a wider spread), re-sort the ranking and record the place. The range runs from the place of the best 10% of draws to the worst 10% (P10–P90) and always includes today's place. Narrow = a certain place, wide = the neighbours are close.
- ▲ / ▼ / “no change since yesterday”
-
How many places the artist moved up (▲) or down (▼) compared with yesterday's ranking, renumbered among public profiles only.
- NEW (ranking row)
-
The artist had no place in the main ranking yesterday and has one today.
- tied
-
Two or more artists have the same (rounded) score and share the place.
- #1 in genre
-
The artist's place among main-ranking artists who share the genre (the profile's headline genre; anyone who has it among their genres counts). Same order as the main ranking.
- Live ranking · updated
-
The date (on the home and ranking pages also the time) of the last recalculation from measurements. “Live” means no weekly charts — the ranking is recalculated with every data update.
- data under review
-
Something in the data stands out (views without likes, one track making all the growth, unusually many followers, an uncertain platform account) and we are checking it. The label itself never removes places; only a disputed platform account is left out of the score until confirmed.
Where the place comes from
- Weights % (e.g. “YouTube — listening 82%”)
-
How much each source weighs in this artist's score. The data sets the weights, not us: the better a platform's numbers agree with the rest across hundreds of artists, the more reliable and the heavier they are. That is why artists with the same set of sources (e.g. YouTube + Spotify + Deezer) have identical weights. Weights differ when an artist is missing from a platform — the other sources share its part (no account never lowers the place) — or when subscribers or fans exceed the cap of 3× above listening.
- YouTube — listening to the best tracks
-
The main source. We take the views of each catalogued track — but fresh views weigh more than old ones: a view from a year ago counts half, from two years ago a quarter. Then we sort the tracks from best and count the 1st track in full, the 2nd at one half, the 3rd at one third and so on — so the best tracks people play now count, not the number of uploads or past fame. The best track counts at most 2× the rest of the catalogue (at least 250K), and views without likes and comments only up to a normal level.
- Age of views (old hits)
-
Ranking 2.1 measures listening NOW. Every view loses weight with age: after a year it counts half, after two years a quarter. Views gained between our measurements get the measurement date. We do not know day by day the views from before our first measurement, so we spread them evenly from the track's release to that first measurement. A hit people still play loses nothing — fresh views keep coming. An old hit with no new listeners slowly gives up its place.
- YouTube — subscriptions
-
Channel subscribers as a measure of a steady audience, counted at most 3× above how much the artist is actually listened to.
- Spotify, Deezer and Instagram in the score
-
They count only for a confirmed account and only once at least 30 artists with confirmed accounts are available for calibration. Deezer and Instagram are audience — capped at 3× above listening. A reading older than 30 days never counts.
- Why #N
-
On a profile: the artist's strongest side (the source where they rank best against other ranked artists) and a comparison with the neighbours — what keeps them ahead of the place below and what they lack most for the place above. Computed from the same components as the place. “Minimal difference” means the neighbours are so close that the 7-day average decides. Freshness is already part of “YouTube listening”, so there is no separate “form” reason.
Platform numbers on the profile
- subscribers (YouTube)
-
Channel subscribers as YouTube reports them (YouTube rounds large numbers, e.g. to 43,200).
- channel views
-
All views of all videos on the channel — including ones we do not count as tracks (Shorts, teasers, covers, non-music videos).
- views of catalogued tracks
-
Views of only the videos we qualified as the artist's music tracks; the same track uploaded several times counts once (the version with most views). That is why it is lower than channel views.
- monthly listeners (Spotify)
-
How many different people listened on Spotify in the last 28 days — from the artist's public page.
- fans (Deezer)
-
Fans from the artist's public Deezer page, always shown with a link to it.
- track views (YouTube Music)
-
Total views of tracks released through a distributor (DistroKid etc.) that YouTube Music publishes on the automatic “Artist - Topic” channel. A separate source: these views are never added to YouTube. When the same track also has a music video, the profile shows one track with two numbers. They count towards the score like YouTube listening (older views weigh less) once 30 artists with a Topic channel are available.
- followers (Instagram)
-
Followers of a business or creator Instagram account linked on the profile. They join the score once 30 confirmed accounts are available.
- TikTok
-
Followers and likes from the artist's public profile page (the link on their profile here). Followers count as audience only from 30 artists shared with YouTube; likes are only shown.
- read (date)
-
The day we read that number from the platform.
Artist profile and track card
- N tracks in the catalogue
-
How many of the artist's tracks we show publicly: original music tracks that passed our checks (no covers, compilations, Shorts or manually excluded items).
- AI music ✓
-
Confirmation that this is music made with AI. It links to the confirming source (channel or video description, the artist's statement); if there is none, to the YouTube channel.
- Channel views — last 30 days
-
Growth of channel views between the first and the last reading of the last 30 days (number and %).
- “growing faster than X% of profiles”
-
How fast the channel grew over 30 days compared with other profiles: X% grew more slowly. Shown only from 50% up, with a full 30 days of readings and no drop in the channel counter.
- Views, likes, release date
-
Views and likes from the latest measurement of the YouTube video; release = the video's publication date. “+X% since (date)” is the change since the track's first measurement.
- NEW / RISING (track)
-
NEW = published in the last 30 days. RISING = one of the artist's 3 tracks with the largest % growth since the first measurement (when at least 3 tracks have several measurements).
- “too early”
-
The track was first measured right after release, so a % growth since the first measurement would mislead — we do not show it.
- measurement / N measurements
-
How many times we read this track's numbers.
Lists: Trending, Rising, Debuts
- Trending — “growth in 90 days”
-
Who gained the most track views in the last 90 days, counted like listening: track gains from the largest, each next one with less weight; a new track (last 90 days) counts in full. No hit cap here, so a viral hit can be first.
- Rising
-
Main-ranking artists growing faster over 90 days than a typical artist of the same size, ordered by the biggest pace advantage. The list itself does not change the main-ranking place (since Ranking 2.1 fresh growth works through the age of views).
- Debuts
-
Artists with fewer than 3 measured recordings or whose first recording is younger than 30 days. Recordings from YouTube and YouTube Music count together, the same song on both only once. They move to the main ranking once they have at least 3 measured recordings and the first is older than 30 days.
- Numbers in a ranking row
-
Views of catalogued YouTube tracks plus Spotify monthly listeners (without Spotify: YouTube subscribers). The thin coloured bar is each platform's share of the score.
The formula, for the curious
- Score (a technical value, not shown in the ranking) = 100 + 20 · (θ − 6). θ is popularity on a log scale: +20 = 10× more popular.
- Age of a view: weight = 0.5^(age in days / 365) — half after a year, a quarter after two. Growth between measurements gets the measurement date; views from before the first measurement are spread evenly from release to that measurement.
- YouTube listening: tracks from best to worst (after ageing), track no. i counts with weight 1/i (Zipf).
- Hit cap: the best track counts at most 2× the rest of the catalogue (always at least 250K).
- Accepted track views ≤ K × (likes + 3 × comments); today K = 415.
- Platform calibration on shared artists (≥30 pairs; YouTube Music — the same measure as YouTube — from 10 pairs), weight = reliability from the data. Today: YouTube listening 73%, ytmusic 9%, YT subscriptions 8%, Spotify 7%, Deezer 2% (recalculated 2026-10-25).
- Audience cap: subscribers and followers at most 3× above how much the artist is listened to.
- Place: 7-day average of measured days without extremes; a neighbour overtakes with a score lead > 0.5.