Doldur Music data story

Are songs getting shorter?
Here, yes.

8,778 tracks · CC0

The typical 2020–2025 release here runs about 40 seconds shorter than a 2000s release. That is a striking pattern in this catalog—not proof of a Spotify-wide shift.

View the Kaggle source
3:44typical track length · 2000s
3:36typical track length · 2010s
3:04typical track length · 2020–2025
40 secshorter from the 2000s to 2020–2025

01 · The short answer

Newer releases run
about 40 seconds shorter.

The typical song in this collection falls from 3:44 in the 2000s to 3:04 in 2020–2025. We grouped songs by release year so the change is easy to compare.

Track length, decade by decade

8,776 songs with a usable release year · length shown as minutes:seconds

Typical track length, explicit labels, and current popularity by release period
Release yearsSongsTypical lengthMarked explicitCurrent popularity
1950s72:420.0%58.0
1960s502:520.0%57.5
1970s804:132.5%72.0
1980s1004:124.0%64.0
1990s4064:0121.2%58.0
2000s9373:4414.4%57.0
2010s3,8273:3624.5%58.0
2020–20253,3693:0430.4%57.0

The change is modest between the 2000s and 2010s, then becomes much sharper after 2020. Explicit labels also rise—from 14.4% in the 2000s to 30.4% in 2020–2025. Current popularity barely moves across those three groups, but that score was captured once, not when each song came out. Our average song length by decade guide explains why the median is the safer comparison.

02 · What the file leaves out

A clear pattern.
With an asterisk.

The numbers are usable, but this collection is not a complete history of Spotify. Here is what we checked before deciding how far the headline could go.

Checks that define what this song-length comparison can support
What we checkedWhat we foundWhy it matters
The same songs twiceThe second file repeats all 8,582 songs already found in the original.We counted each song once.
Unexplained editsIt changes 198 release dates and 123 titles without explaining why.We used the untouched original.
How the songs were chosenKaggle does not explain how the artists and tracks were selected.This collection cannot stand in for all of Spotify.
Where listeners heard themThere are no country, chart, stream-count, or market fields.This cannot show a global trend.
When popularity was measuredEach song carries a popularity score captured in the same present-day snapshot.Those scores are not a history of what was popular.

The collection contains no platform incentives or short-form-video exposure, so it cannot show whether Spotify favors shorter songs or whether TikTok made songs shorter. Those explanations need different evidence.

How we checked it

How we counted
every second.

We used the original track_data_final.csv from version 1 of Spotify Global Music Dataset (2009–2025), updated 11 November 2025 and released under CC0: Public Domain.

We read the year from each album release date, left out 2 unusable dates, grouped the other songs by decade, and found the middle track length in each group. That middle value keeps a few unusually long or short songs from pulling the result around.

The collection reaches back to 1952 despite its 2009–2025 title. Those older groups stay visible, but they are too small and selectively assembled to drive the headline.

This comparison shows what changed inside one collection. It cannot tell us why songs are shorter, whether listeners prefer them, or whether the same pattern holds across Spotify. Read why listeners skip songs before treating every early exit as dislike, and use our song-length decision guide instead of forcing one duration target.

Continue exploring

What shorter songs
do—and do not—mean.

These five guides separate the measured duration pattern from claims about platforms, TikTok, skipping, and artistic decisions.

For independent artists

Turn your release into
a story people can find.

Bring in a Spotify track or album, add the story and credits, then export editable release metadata.

Open the Release SEO Kit