The discursive flexibility of changecraft: Platform change discourse in Meta, TikTok, YouTube, and X

Summary

De and Cotter introduce “changecraft” as a conceptual vocabulary for analyzing how major social media platforms — Meta, TikTok, YouTube, and X — discursively construct, justify, and normalize the constant flux of their features, policies, and governance. Drawing on qualitative analysis of 301 public-facing communications from 2022–2025, the authors argue that platforms exhibit a “discursive flexibility” that lets them stage transformation as neutral, responsive, and community-serving while in fact consolidating power and deepening user dependence. The paper reframes platform change not as a hidden technical process but as a politically charged communicative practice that warrants its own analytic toolkit.

Key Contributions

• Coins changecraft as a novel concept for studying platform transformation as a discursive and political practice rather than a purely technical or policy matter.
• Provides a shared vocabulary — infrastructural visibility, ideological recalibration, and patchworking — for parsing how change is communicated.
• Shifts the analytic frame from concealment (à la Gillespie, Plantin, Helmond) to strategic visibility and discursive management of change.
• Offers a comparative empirical mapping across four platforms, surfacing recurring legitimation tropes: creator benevolence, democratization/participation, and connectivity.
• Bridges platform studies with infrastructure studies by showing how platforms invert infrastructure’s typical invisibility, making change aesthetically conspicuous as a governance tactic.

Methods

Qualitative content analysis of 301 platform-authored documents — Meta (88), TikTok (70), X (70), YouTube (58) — drawn from official newsrooms and creator/policy blogs between April 2022 and January 2025. The authors performed iterative inductive thematic coding in ATLAS.ti following Braun and Clarke, beginning with five sampled documents per platform and refining categories across rounds, supplemented by discourse-analytic attention (Wood and Kroger) to framing, positioning, and silences. Data were reduced to platform updates, company communications, user-directed messages, and management statements, with memoing and team discussion generating themes of material change, ideological change, and rationalization.

Findings

• Material changes are foregrounded visually: interface redesigns, screenshots, and feature demos (Meta AI Backgrounds, YouTube Shorts on TV) emphasize how over why, producing an aesthetic of progress.
• These visible updates often mask deeper monetization and control shifts — e.g., Reels ranking/payout changes or YouTube’s likeness-management tools framed as creator protection.
• Ideological shifts are reframed as continuity, using phrases like “updating our approach,” “long-standing efforts,” and “original intention” (Meta’s retreat from third-party fact-checking; X’s post-acquisition transparency rhetoric).
• Creator benevolence framing casts platforms as supportive mentors (TikTok Creator Studio, YouTube Hype, Instagram Best Practices), legitimizing infrastructural and monetization changes.
• Democratization rhetoric (“for fans, by fans”; “collaborative” teen supervision) simulates participatory governance without ceding decision-making.
• Connectivity discourse binds new features (Views metric, stickers, subscriptions) to community-building while quietly embedding surveillance and engagement-maximization logics.
• Patchworking — many small, cumulative updates — reorients platform logics under the radar of any single change.

Connections

This paper sits closest to work on platform infrastructural power and discursive self-presentation: Helmond2026-ll on platformization/infrastructuralization is a direct theoretical interlocutor, and Rieder2026-pp and Rieder2025-ju on platform governance and accountability complement its account of how change is staged rather than hidden. It also speaks to studies of platform policy drift and feature churn such as Bastos2025-ya, Bastos2025-ol, and Murtfeldt2025-wu, and to the broader research-access and transparency concerns raised by Freelon2024-sc and Tornberg2026-lc, which together map the regulatory and epistemic terrain in which changecraft operates.

Podcast

A research-radio episode discusses this paper: Listen