In streaming charts across the EU, only 14 per cent of songs are by artists from another EU country. On the global digital platforms that now dominate book sales, EU titles typically account for less than 10 per cent of the most popular titles. In 2023, out of 184 million tracks available on music streaming services, nearly 80 million were streamed fewer than ten times.
These are among the central findings of the Study on the Discoverability of Diverse European Cultural Content in the Digital Environment, a major research project commissioned by the European Commission under the EU Work Plan for Culture 2023–2026, completed in March 2026.
Compiled by a consortium of six European research institutions — including KEA European Affairs, the Vrije Universiteit Brussel, the University of Warsaw's DELab, and Erasmus University Rotterdam — and peer-reviewed by a Scientific Board of five independent academics, the study is the most comprehensive examination the Commission has produced of how digital platforms are reshaping cultural access across the continent.
Its conclusions are urgent, and its implications for the EU's creative sector run far deeper than any single policy instrument has yet addressed.
Policy has long focused on the production and availability of diverse cultural content, on funding creation, on digitising archives, on supporting translation. What policymakers have been slower to recognise is that in today's digital environment, being available is no longer sufficient.
Discoverability — whether a work is actually noticed, surfaced, and engaged with by audiences — has become the decisive factor.
The researchers define discoverability as a systemic condition shaped by what they call a "discovery ecosystem" of six interconnected elements: the personal and social circumstances that make audiences open to exploration; the ways platforms organise and present content; the human and automated systems that act as gatekeepers; the behavioural data that platforms collect and feed back into recommendations; the actual experience users have as they navigate platforms; and the broader cultural and economic consequences of how all of the above interact.
The central measure of whether discoverability is genuinely happening, the study argues, is not how much content exists — but whether audiences are actually exposed to a diverse range of it. By that measure, Europe's digital cultural landscape is in structural difficulty.
The study's most striking quantitative findings concern the music sector. Streaming now dominates revenues and acts, the researchers write, as "the main gateway to discovery."
Streaming platforms are the primary route through which 89 per cent of European listeners discover music, ahead of recommendations from friends and family (85 per cent) and live concerts (73 per cent). The platform is, functionally, the market.
Yet the dynamics inside that market are heavily skewed. In 2023, 184 million tracks were available on streaming platforms. Nearly 80 million of them — close to half — were streamed fewer than ten times.
At the same time, AI-generated music already constitutes close to one fifth of all uploads, adding further pressure on visibility for human-created work.
The effect of this saturation is not neutral: recommendation systems and user attention are distributed across an ever-expanding pool of content, and algorithms optimise for engagement, which functionally means they privilege what is already popular.
The data on European content in streaming charts illustrates the consequences. European music performs reasonably well in national contexts — 38 per cent of songs in top streaming charts are European, largely driven by domestic artists who account for 24 per cent of chart entries.
But cross-border circulation collapses: non-national EU works account for just 14 per cent. Europe's linguistic and cultural diversity, which ought to be a competitive asset, becomes a structural barrier the moment an artist tries to reach audiences beyond their home country.
The study identifies a particular disconnect in how users perceive their own discovery experience. While 62 per cent of surveyed users report that algorithms help them discover new music, far fewer say this has increased their exposure to national, EU, or non-English-language works. Users feel discovery is happening. The data suggests it is mostly happening within familiar territory.
The book sector presents a different but equally sobering picture. Unlike music, books remain strongly anchored in offline discovery: roughly 75 per cent of sales still take place in physical bookstores, which provide what the study describes as "trusted curation, community engagement, and personalised recommendations." Word-of-mouth remains the most trusted discovery channel.
The physical infrastructure of bookselling is, for now, holding. But that infrastructure is not where the future is being built. On the global digital platforms that are rapidly expanding their reach, EU titles are structurally underrepresented. Based on the study's data analysis of platforms targeting global audiences, non-EU titles frequently dominate, with EU titles typically accounting for less than 10 per cent of the most popular books.
National EU titles achieve substantially stronger visibility on local EU platforms and national storefronts — the researchers note this is where European literary culture largely survives online — but non-national EU titles, the cross-border circulation that would constitute a genuine European digital market for literature, represent a consistently smaller share. The exceptions arise mainly in smaller markets or linguistically shared ones.
Translation is the sector's most acute structural constraint. The study identifies translation gaps and high costs as major barriers to cross-border circulation, compounded by the competitive advantage of English-language editions, which are both more affordable and more widely distributed.
The relative dominance of English online is self-reinforcing: platforms optimise for global audiences, global audiences skew towards English, and recommendation systems learn from that pattern and reproduce it.
Across both sectors, and across the performing arts, visual arts, and cultural heritage the study examines more selectively, a common structural problem emerges: the quality of metadata. Incomplete or inconsistent metadata — incorrect International Standard Recording Codes in music, inconsistent cataloguing standards in books, limited support for non-Latin scripts — undermines visibility at the most basic level.
A work that cannot be accurately described cannot be accurately recommended. The study flags this not as a technical inconvenience but as a systemic equity issue, since the actors least likely to have the resources for rigorous metadata management are independent artists, smaller publishers, and institutions serving minority-language audiences.
The opacity of algorithmic curation compounds this problem. The study documents a persistent lack of transparency around how ranking, promotion, and recommendation decisions are made. Creators, cultural organisations, and researchers operate, as the authors put it, "in the dark about how visibility is generated and how it might be improved."
This opacity is not incidental — it reflects the commercial logic of platforms whose recommendation systems are proprietary assets — but its cultural consequences are significant. It means that structural bias favouring mainstream, English-language, and already-popular content is difficult to identify, difficult to challenge, and difficult to correct.
The study proposes six strategic action areas, directed at both policymakers and cultural sector actors. The first is structural: establishing a dedicated EU multi-stakeholder forum for discoverability, integrating it into EU and national cultural strategies, and creating the cross-sector partnerships needed to address transparency, metadata, and data access in a coordinated way.
The second addresses the data deficit that makes the problem so difficult to measure. The researchers call for strengthening cultural data collection through a forthcoming EU Cultural Data Hub, developing consistent definitions of what constitutes a "European work" for platform purposes, improving metadata standards, and conducting regular consumer surveys on cultural consumption and discovery. Without this evidence base, the study argues, monitoring the visibility of European works remains structurally impossible.
The remaining recommendations span digital capacity-building for creators and smaller organisations that currently lack the skills to optimise their visibility online; audience-focused interventions around digital literacy and linguistic inclusion; continued research into fair and diverse recommendation systems; and, on the supply side, expanded funding for translation and localisation, curated cultural showcases, and mechanisms to improve the prominence of European works on major platforms.
The study does not call for algorithmic quotas — that is not within its scope — but its findings make the logic of intervention on recommendation systems difficult to resist. The Audiovisual Media Services Directive already requires prominence obligations for European works in the audiovisual sector. The study's evidence base now makes a comparable policy conversation for music and books far harder to avoid.
The study does not address AI and copyright directly — that falls to other ongoing EU processes — but it identifies the rise of AI-generated content as a compounding challenge for discoverability specifically. AI-generated music already constitutes close to a fifth of uploads on streaming platforms.
AI-generated books are flooding digital marketplaces. In both cases, the effect is to increase content saturation while raising unresolved questions about attribution, metadata integrity, and cultural value.
The study frames this as a risk of "cultural homogenisation" — not through explicit design but through the logic of scale: if recommendation systems reward engagement, and AI can generate engagement-optimised content at near-zero marginal cost, the structural pressure on human-created European works intensifies regardless of any individual platform's intentions.
This is the discoverability problem in its most acute form. Europe's creative sector is not failing to produce. It is being outcompeted — by algorithmic logic, by content saturation, and by the structural advantages of global platforms operating in English — in the contest for attention that now determines cultural survival.
The Commission has produced the diagnosis. The question, as with the Copyright Directive review now a few weeks away, is how it will act on it.
