For an industry currently preoccupied with artificial intelligence, publishing may be overlooking a more immediate and structural risk: it has already surrendered much of its control over how books reach readers (and how to understand readers).

Amazon determines visibility at the point of sale. TikTok drives breakout success through BookTok. Google shapes discoverability through search. Meta influences audience targeting and advertising reach. Together, these systems form the infrastructure through which books are now found and bought.

This is not simply a question of tools. It is a question of dependency, and it is a pattern the wider media industry has seen before. Over the past decade, newspapers built their audiences on social media platforms, only to lose control over distribution, reader relationships and, crucially, data. When platforms changed their algorithms or deprioritised news, publishers saw traffic collapse, revenues shrink and editorial strategies distorted to fit platform logic.

The result was not just economic damage, but a structural loss of independence. News organisations had become dependent on systems they did not control – and when those systems shifted, they had little recourse.

Book publishing is not identical to journalism as it does not rely on daily traffic cycles, and its revenue model is more diversified. But the direction of travel is strikingly similar.

Today, success in publishing is increasingly tied to platform visibility. A title can languish for months until a viral TikTok video transforms it into a bestseller overnight. Amazon rankings can make or break a book’s commercial trajectory. Search optimisation determines whether readers encounter a title at all.

Publishers are not just selling books, they are optimising for algorithms — and this has consequences that extend beyond marketing, since platforms are not neutral intermediaries.

Certain genres perform well and are amplified. Others struggle for visibility and fade. Topics that are complex, politically sensitive or less easily commodified may underperform not because of a lack of reader interest, but because they are less compatible with platform dynamics. The result is a form of market-driven self-censorship, mediated not by editors or governments, but by opaque technological systems.

When discoverability depends on the algorithm, the boundary between editorial judgement and platform optimisation becomes increasingly blurred.

This is where the parallel with journalism becomes particularly instructive. As newsrooms adapted to platform incentives, editorial priorities shifted – often unconsciously. Stories were selected, framed and promoted in ways that aligned with what platforms rewarded.

Publishing is beginning to face a similar dynamic, albeit more gradually. When discoverability depends on the algorithm, the boundary between editorial judgement and platform optimisation becomes increasingly blurred.

The risk is not simply that publishers become dependent on platforms for sales. It is that platforms begin to shape what kinds of books are commercially viable in the first place. None of this suggests that publishers can simply abandon these ecosystems. Readers are there, and the commercial reality is unavoidable.

But the absence of a clean exit does not mean the absence of strategy. If the lesson from the news industry is that total dependency is dangerous, the logical response is to reduce it while there is still time to do so.

One of the clearest lessons from news media is that losing first-party data was catastrophic. For publishers, this means investing more seriously in newsletters, direct-to-consumer models, subscription ecosystems and owned communities – not as side projects, but as core infrastructure.

A second step is conceptual. Platforms should be treated as volatile channels, not stable foundations. Diversification, in this sense, is not simply about being present on multiple platforms, but about ensuring that no single system becomes indispensable.

There is also a role for collective action. Individually, publishers have limited leverage over global tech companies. But as an industry (also with other affected sectors), there may be greater scope to push for transparency, particularly around discoverability, ranking systems and data access.

Finally, publishing has a stake in broader regulatory debates with which it has, so far, only marginally engaged. In Europe, initiatives such as the Digital Markets Act are beginning to address platform power. Questions of data access, interoperability and algorithmic transparency are not abstract policy concerns. If publishers are not part of that conversation, they risk being shaped by its outcomes rather than influencing them.

Longer term, discussions around digital sovereignty (including the development of alternative technological infrastructures) also matter. A more plural digital ecosystem would not eliminate platform dependency, but it could reduce its concentration.

None of these measures offers a quick fix but they point toward a different trajectory and one in which publishers retain some degree of control over how their books reach readers – because the real risk is not that platforms exist, but that they become unavoidable and dictate outcomes.