When the music industry moved from CDs to streaming, it did not just change how people listened. It changed the entire economics of getting paid. On platforms such as Spotify and Apple Music, artists earn a fraction of a cent each time someone presses play (roughly US$0.006 per stream). That means a song needs to be heard thousands of times just to generate a modest return. With more than 120,000 tracks uploaded to streaming services every day, the old marketing playbook of pricing, promotions, and physical distribution has become largely irrelevant.

In this landscape of overwhelming choice, playlists (curated collections that helped listeners navigate millions of songs) have emerged as one of the most consequential marketing tools in the business. Record labels have taken notice, with the three major labels – Universal, Sony, and Warner – each building their own playlist brands (Digster, Filtr, and Topsify, respectively) and deploying dedicated teams to pitch songs for inclusion on playlists run by the platform or third parties.

The stakes are high for both artists and the music industry. Two-thirds of global recorded music revenue, valued at US$29.5 billion, comes from streaming. While the industry has poured resources into playlist marketing, a fundamental question remained unanswered: which playlists work, and why? Until now, the evidence base was thin.

What 200,000 playlist experiments revealed

A study published in the Journal of Marketing addresses this gap. The paper, Driving Music Demand in the Age of Streaming: Understanding the Heterogeneity in Curated Playlist Effectiveness, was authored by Professor Nils Wlömert of WU Vienna University of Economics and Business, Professor Dominik Papies of the University of Tuebingen, and Professor Harald van Heerde from the School of Marketing at UNSW Business School.

The researchers analysed more than three years of daily global Spotify data, covering 23,721 playlists, 368,274 songs from 158,320 artists, and over 123 billion streams. They identified 200,455 instances in which a song was added to a single playlist and later removed, and used a difference-in-differences approach (comparing each listed song’s streaming trajectory with a matched control) to measure the effect. On average, playlist inclusion lifted streams by 8.5%. After removal, streams remained about 4% higher – a carry-over that accounted for roughly 32% of the total uplift.

Subscribe to BusinessThink for the latest research, analysis and insights from UNSW Business School

“This is an economically meaningful finding, especially because streaming revenue depends on the number of streams a song gets,” said Prof. van Heerde, who noted that, for income, the exact amount depends on baseline streams, platform payout rates, and the artist-label contract.

Given this, Prof. van Heerde said the research reports revenue implications as simulations. In those simulations (with decay), he said the average listing-period impact is about 7% of baseline revenue over 90 days, and carry-over adds about 2.1% over the following 90 days.

Why not all playlist placements delivered equal results

Those averages, however, masked a wide variation. The most influential factor was playlist popularity. Songs placed on playlists in the top 20% by follower count saw a 21.6% increase in streams; those on bottom-quintile playlists saw no uplift. In revenue terms, a top-quintile placement yielded about US$1997 over 90 days, compared with US$77 for a bottom-quintile listing.

A song’s prior exposure mattered almost as much. Songs with the least prior playlist exposure received a 14.6% uplift – more than four times the 3.4% for the most widely exposed tracks. Position on the playlist also shaped results: songs in the top 10 slots received a 13.4% lift versus 8% further down, a finding the researchers likened to shelf placement in physical retail.

Prof. van Heerde said it is important to keep these numbers in mind. “First, in a streaming world, traditional marketing instruments that would typically be used to boost streams (e.g., price promotions or record stores) are no longer available,” he said.

“Second, playlist listings are for free, i.e., artists do not pay for them. Third, especially for artists and songs that were not on the radar before, these effects can be much larger, so playlists are an economically very important tool for artists, both in terms of income and in terms of visibility.”

The trade-off between smooth listening and music discovery

The study’s most consequential finding for playlist strategy concerned two types of “fit”. Attribute-based fit captured how closely a song’s musical style matched the playlist. Co-listing-based fit measured how often a song had previously appeared alongside its new companions on other playlists.

Both types of fir lifted streams during listing, with co-listing fit producing the larger effect (12.4% for the top quintile versus 3.9% for the bottom). But after removal, the relationship reversed: songs with lower fit retained larger gains, with the carry-over roughly twice as large. Less predictable combinations fostered the kind of discovery that turned playlist listeners into lasting fans.

Similarly, Prof. van Heerde said there is also a trade-off between short-term consumption and longer-term discovery. Songs placed on highly fitting playlists (such as a well-known 80s rock song on an 80s rock playlist) increase short-term consumption, but they reduce carry-over effects because there is less scope for discovery.

Learn more: 4 lessons in digital transformation from the lucrative K-pop industry

Curation approach and the right song for the right context

Context-based playlists – built around activities such as running, studying, or driving – outperformed content-based playlists (organised by genre or artist) on both listing and carry-over effects. The researchers linked this to the prolonged and recurring nature of contextual listening. Who curated the playlist also mattered: platform-curated playlists produced the largest lift during listing, but label-curated playlists generated the strongest carry-over.

“So, the key message is: playlist listings alone are not what drives demand, but it must be the right song on the right playlist,” said Prof. van Heerde.

What playlists mean for independent vs major labels

The research found that independent labels saw larger uplifts than major labels, consistent with the idea that playlist exposure was more valuable for artists who lacked the promotional reach of the majors.

“Our results do not support the idea that playlists only help major-label acts,” Prof. van Heerde said. In relative terms, the revenue uplift is somewhat larger for independent-label songs than for major-label songs (8.6% vs 7.5% during listing, and 4.0% vs 3.6% carry-over), but he noted that this difference is relatively small.

“That said, majors still have structural advantages in reach and curation infrastructure. The paper shows that major-label playlist brands are very large in aggregate follower count (e.g., Filtr, Digster, Topsify, which are owned by the major labels, ranked just behind Spotify),” he said. “Playlists can open doors for independent artists, but the system is still uneven because major labels dominate in terms of distribution power (through playlists) even in a streaming world.”

The power of playlists and gatekeeping risks

“Playlists are a powerful marketing channel and curator power matters,” said Prof. van Heerde, who observed that, overall, playlist effectiveness is high relative to reported advertising and radio-play effectiveness in prior recorded-music work.

He also highlighted concerns about “gatekeeping risk” and noted the strongest effects are tied to larger playlists, which can concentrate influence. “We do recommend close observation of curator power by producers and policymakers. If Spotify, as the largest curator of playlists, decides to drop an artist from its playlists, this would have severe monetary consequences for that artist,” he said.

On streaming services, revenue is usage-based, payout rates vary across platforms, and artists’ income can benefit from playlist listings. How much depends on the platform, so Prof. van Heerde said platform choice can matter (while noting that TikTok does not pay artists for the use of their music). “In any case, if listeners care about the artists, they should select a platform that pays artists, e.g., Apple Music, Spotify, or Deezer, or any other paying platform,” he concluded.