Examining Platformization in Cultural Production: A Comparative Computational Analysis of Hit Songs on TikTok and Spotify

This study employs a comparative computational analysis of a two-year longitudinal dataset to demonstrate how the distinct platformization practices of TikTok and Spotify shape the characteristics of hit songs while revealing evidence of cross-platform influence where Spotify popularity may precede trends on TikTok.

The Gist

Imagine the music industry as a massive, bustling marketplace where songs are the goods being sold. For decades, this marketplace was run by a few big, established shopkeepers (record labels) who decided what got on the shelves. But today, two new, very different "super-malls" have opened up: TikTok and Spotify.

This paper is like a detective story where researchers went into both malls to see how they run their businesses, what kind of music becomes a "hit" in each, and whether one mall influences the other.

Here is the breakdown of their findings using some everyday analogies:

1. The Two Different Malls

Think of Spotify as a high-end, traditional record store.

• How it works: You go there to listen to full songs from start to finish. It's curated by experts (and algorithms) who act like knowledgeable clerks recommending albums.
• The Vibe: It's about the music itself. The focus is on the full experience, the lyrics, and the artist's story.
• The "Big Shops": Because it's a traditional distribution channel, the big record labels (the "Mega-Corporations") still hold the most power here. They have the money to get their songs on the best shelves.

Now, think of TikTok as a wild, chaotic street fair or a talent show.

• How it works: You don't just listen; you do. People use 15-second clips of songs as background music for their own videos, dances, or jokes. The music is just a prop for the video.
• The Vibe: It's about the vibe and the trend. If a song makes you want to dance or make a video, it wins.
• The "Small Shops": Here, the big corporations aren't the only ones winning. Independent artists (the "local street performers") can go viral just because their song fits a funny dance challenge perfectly.

2. What Makes a Song a "Hit"? (The Ingredients)

The researchers looked at what makes a song popular in each place and found they need different ingredients:

• The "Label" Factor:

  • On Spotify: The "Big Shops" (Major Labels) dominate. It's like a high-end restaurant where only the famous chefs get to cook.
  • On TikTok: It's a mix. Interestingly, the study found that songs from independent artists (the "street performers") actually did better than the big labels on TikTok. The algorithm loves a fresh, viral sound, regardless of who owns the copyright.

• The "Theme" Factor:

  • On Spotify: People listen to full songs, so they care about the story. Songs about Love and Relationships are huge hits. Songs about anger, politics, or violence tend to get less play. It's like people want a romantic dinner, not a political debate, while listening to music.
  • On TikTok: The theme of the song doesn't matter as much as the energy. Whether it's about love or partying, if it makes you want to move, it works. The lyrics are often chopped up, so the message gets lost, but the beat remains.

• The "Genre" Factor:

  • On Spotify: It's a melting pot. You hear everything from Pop to Rock to R&B.
  • On TikTok: Dance is king. If a song has a beat you can move to, it's a winner. The study found that "Dance" music is the only genre that specifically predicts a hit on TikTok.

3. The "Domino Effect" (Who Influences Whom?)

This is the most fascinating part of the study. The researchers asked: Does a song get popular on TikTok first and then go to Spotify? Or does it start on Spotify and then blow up on TikTok?

They found a clear pattern: Spotify usually starts the fire, and TikTok fans the flames.

• The Analogy: Imagine a song is a new toy.
  1. Spotify (The Launchpad): The big record labels release the toy on Spotify first. They put it on the "New Releases" shelf. People start listening to it.
  2. TikTok (The Viral Engine): Once people hear it on Spotify, they take that sound to TikTok. They start making dances, memes, and videos with it. Suddenly, the toy is everywhere.
  3. The Result: The song often hits the Spotify charts before it hits the TikTok charts. Once it explodes on TikTok, it might actually drop in popularity on Spotify because the "novelty" has worn off, or the song has moved on to the next phase of its life.

The Big Takeaway

The world of music isn't just one big pool anymore; it's two connected but very different ecosystems.

• Spotify is the Distribution Center: It's where music is delivered, curated, and consumed as a complete product. It favors the established giants.
• TikTok is the Creative Lab: It's where music is remixed, repurposed, and turned into a cultural moment. It favors the creative, the viral, and the independent.

The study concludes that while they are different, they are deeply connected. A song usually gets its "official" start on the distribution side (Spotify) and then gets its "cultural" explosion on the creative side (TikTok). They are partners in the dance, even if they are dancing to different beats.

Technical Summary

Here is a detailed technical summary of the paper "Examining Platformization in Cultural Production: A Comparative Computational Analysis of Hit Songs on TikTok and Spotify."

1. Problem Statement

The (re)creation and distribution of cultural products, particularly music, are increasingly governed by digital platforms through a process termed platformization. While existing literature has extensively studied platformization within single platforms (e.g., focusing on algorithmic power or business models), there is a significant gap in understanding:

• Comparative Dynamics: How different platforms with distinct governance structures, user bases, and algorithmic priorities shape cultural production differently.
• Cross-Platform Interactions: How influence flows between platforms within a multi-platform ecosystem. Most studies treat platforms in isolation, failing to capture the interconnected nature of the digital music market.

This study addresses these gaps by systematically comparing TikTok (a user-generated short-video platform) and Spotify (a music streaming platform) to understand their distinct platformization mechanisms and their mutual influence on hit song popularity.

2. Methodology

The study employs a longitudinal computational analysis using a two-year dataset (June 1, 2020 – May 31, 2022).

• Data Collection:

  • Source: Chartmetric (music analytics service).
  • Scope: Daily Top-100 hit song charts from both TikTok and Spotify.
  • Sample Size: After cleaning and deduplication, the final dataset included 321 unique TikTok hit songs and 1,707 unique Spotify hit songs (2,028 total records, with 68 songs appearing on both platforms).
  • Data Enrichment: Lyrics were retrieved via the Genius API; genre data was supplemented via Wikipedia and Chosis; record labels were coded based on ownership by major groups (Universal, Sony, Warner) vs. independent labels.

• Measures:

  • Dependent Variable: Music Popularity, measured by on_chart_days (cumulative days on the Top-100) and daily popularity (ranked 1–100).
  • Independent Variables:
    • Record Label: Major vs. Independent.
    • Genre: Categorized into Pop, Rock, Hip-hop/Rap/Trap, R&B/Soul, and Dance.
    • Theme: Coded via content analysis of lyrics into five categories (e.g., Love/Relationship, Social/Politics, Identity/Depression, etc.).
  • Control Variables: Song duration, peak daily popularity, and singer popularity (follower count).

• Analytical Techniques:

  1. Mixed-Effects Linear Regression: To test associations between label, theme, genre, and popularity within each platform (controlling for song-level variance via random intercepts).
  2. Independent Samples t-tests: To compare the distribution of labels, genres, and themes between the two platforms.
  3. Vector Autoregression (VAR): To analyze cross-platform influence. The study used weekly time-series data for songs appearing on both platforms to perform Granger causality tests, Cumulative Impulse Response Functions (CIRFs), and Forecast Error Vector Decomposition (FEVD) to determine the direction and magnitude of influence flows.

3. Key Contributions

1. Comparative Platformization: It highlights the distinct strategic priorities of TikTok (user-oriented, creative production) vs. Spotify (producer-oriented, distribution/consumption) and how these manifest in the features of hit songs.
2. Cross-Platform Interconnectedness: It provides empirical evidence of the "ecosystem" nature of platforms, moving beyond single-platform analysis to examine how popularity spills over between different media types.
3. Methodological Innovation: It is among the first studies to leverage longitudinal computational analysis across two major platforms, complementing previous qualitative and cross-sectional research with robust quantitative data.

4. Key Results

A. Platform-Specific Differences (RQ1–RQ3)

• Record Labels:
  • Spotify: Dominated by Major Labels (75.6% of hits) compared to TikTok (50.3%).
  • Popularity Trend: Counter-intuitively, songs by Independent Labels were more popular (ranked higher) than those by Major Labels on both platforms.
• Themes:
  • Spotify: Themes significantly predicted popularity. Love/Relationship themes were positively associated with success, while Social/Politics/Violence themes were negatively associated.
  • TikTok: No specific lyrical theme significantly predicted popularity, likely because music on TikTok serves as a backdrop for video content rather than being consumed for its lyrics.
• Genres:
  • TikTok: Dance genre was positively associated with popularity, reflecting the platform's "dance challenge" culture.
  • Spotify: No single genre significantly predicted popularity, though the platform showed higher proportions of Pop and Hip-hop/Rap/Trap overall.
• Centralization & Longevity:
  • TikTok exhibited higher centralization (fewer unique songs dominating the charts) and longevity (songs stayed on charts longer) compared to Spotify.

B. Cross-Platform Influence (RQ4)

• Temporal Sequence: Analysis of the 68 songs appearing on both platforms revealed a strong unidirectional flow: 73.8% of songs entered the Spotify charts before TikTok, and 82.4% exited Spotify before TikTok.
• Granger Causality:
  • Spotify $\to$ TikTok: Limited evidence of Spotify popularity causing subsequent TikTok popularity (impulse responses approached zero).
  • TikTok $\to$ Spotify: For a specific subset of major-label songs, increased TikTok popularity negatively predicted subsequent Spotify popularity (suggesting a substitution effect or that TikTok popularity peaks after Spotify popularity has already waned).
• Conclusion on Flow: The evidence suggests that Spotify popularity generally precedes TikTok popularity. Major labels likely launch songs on Spotify for initial exposure, which are then repurposed by users on TikTok for viral challenges.

5. Significance and Implications

• Theoretical: The study refines the concept of platformization by demonstrating that it is not a monolithic process but varies based on the platform's position in the multi-sided market (user-centric vs. producer-centric). It confirms that platforms act as distinct "infrastructures" that reorganize cultural production differently.
• Industry:
  • For Labels: Major labels retain dominance on Spotify but face a more level playing field on TikTok, where independent artists can achieve viral success.
  • For Strategy: The findings suggest a "Spotify-first, TikTok-second" strategy for major releases, where Spotify serves as the distribution engine and TikTok serves as the viral amplification engine.
• Cultural Production: The study illustrates how algorithmic logic shapes content: Spotify curates full-length, lyric-driven experiences, while TikTok curates short-form, participatory, and dance-oriented experiences.

Limitations: The study notes that the data period coincided with the early COVID-19 pandemic, which may have skewed consumption patterns. Additionally, the complexity of platform algorithms makes it difficult to isolate specific technical mechanisms solely through chart data.