Mariana Mazzucato and her co-authors (Tim O'Reilly and Ilan Strauss) propose examining the power of digital platforms through an unusual "entry point": not through the collection of personal data per se, but through how platforms manage users' attention and turn that attention into a source of revenue. The book consists of two parts: the first describes the idea of "algorithmic attention rent" and how to identify it. Attention rent is the revenue a platform generates when it uses its algorithms to direct user attention in a way that benefits it, potentially making search results worse for users without losing sufficient number of users, sales, or partners to make it unprofitable. The second part provides a broader overview of modern forms of economic rent and attempts to distinguish between "useful" and "harmful" rent income. Rent is an income that, roughly speaking, requires no additional effort, and there are situations where rent becomes a symptom of companies minimizing their efforts to the detriment of users and partners, for example, by exploiting their market power instead of evolving and outperforming their competitors.
The first part of the book focuses on digital platforms and how "attention markets" differ from traditional markets where goods are purchased for money. The authors argue that the key resource of platforms is not data, but user attention, and market power is manifested in the ability to direct attention in ways that benefit the platform, sometimes in ways that make the resulting feed or search results independent of user preferences, the relevance of competitors' content, or even explicit search queries. It's important to note that authors agree that platforms do not collect data or that this collection is not essential to their market power. However, for the types of platforms discussed in the book (social media, search engines, marketplaces), data is not valuable by itself; it allows platforms to understand which ads to place and where to place them on the page to ensure users view and click them, thereby turning users' attention into a source of revenue. However, the authors avoid making a reservation that user data as such can be an indicative commodity for companies; this point remains outside their scope.
"Rent-seeking attention" occurs when a platform intentionally stops showing the most relevant results to a user's query (which is essentially how platforms initially attracted their audience) and can afford to do so without incurring significant losses in sales/users/partners.
The platform can afford this because it acts as a "trusted intermediary": users tend to trust the top search results, and the platform can insert content that is not optimal for the user but is beneficial to the platform, for example, paid content, into the "primary focus zone." The authors attribute this to "positional" and "trust" cognitive biases: people are more likely to click on what is shown "higher" and looks "recommended by the system."
The authors provide several examples of search engines within search systems and marketplaces, demonstrating how advertisements (or rather, paid search results) occupy the positions in search results where, statistically, users look the most. This can cause "organic" results (those that actually match a user's query) to be overlooked, degrading the quality of search results while allowing platforms to both collect advertising fees and spend less on improving search quality.
An important section of the first part is an attempt to measure such practices. The authors propose comparing organic and paid rankings, as well as assessing whether advertising duplicates existing information (reducing diversity), and comparing the quality of organic search results on the dominant platform with those of competitors who are more dependent on competition.
The potential damage is also being discussed: consumers may see more content designed for "poor engagement" (conflicts, unhealthy interest, misinformation), while advertisers may suffer from being placed next to inappropriate content. The primary beneficiary of this is the major platform.
Trying to answer the question "What should be done?", the authors propose, for example, reserving some of the most prominent positions on the screen for "organic" results, and if an ad essentially repeats an existing "organic" link and doesn't add new information, the "organic" result should be shown first. For recommendations (TikTok, X/Twitter, Instagram), the authors propose requiring platforms to provide users with "hard" preference settings that are truly saved (if a user opts out of the imposed preferences, the platform should not quickly and surreptitiously revert them back). However, the authors admit that behavioral restrictions alone may not be enough: platforms quickly change their designs and algorithms, resist rules, and overly strict regulation can accidentally suppress beneficial innovations. Therefore, the most important measure proposed in the book is regular disclosure of the operational metrics by which platforms manage attention: ad load, share of organic and ad clicks, user time spent on services, traffic volume to external sites and the company's own products, sales and revenue data on marketplaces, etc.
In the second part, the authors distinguish between “useful” rents, for example, when a company invents something new and as a result receives increased income for some time, and “harmful” rents, when profit is gained primarily from control over the rules of the game, market access, or other people’s dependencies, rather than from actual improvements to the product or technology.
This section pays close attention to the logic of Schumpeter's "creative destruction": innovations can generate "rents" as a temporary reward for a more efficient technology, but as the technology diffuses, these rents should diminish. Using patents as an example, the authors demonstrate that in practice, the "temporary reward for innovation" can turn into long-term monopoly profit, and that innovation studies often fail to distinguish between these cases.
Considering the book's strengths, the authors shift the conversation from the abstract "they have a lot of data" to a more understandable level, i.e., what exactly the user sees, why, and how they can make money. The authors provide examples of specific companies, drawing on the data available to them, and honestly point out when the information is insufficient for definitive conclusions. The mechanism of "competition for screen space" (between paid and organic) is described clearly and down-to-earth, with clear implications for users and businesses.
Another advantage is that authors are attempting to identify verifiable indicators of "algorithmic rent," suggesting ways to measure this rent and understand when a platform is no longer useful to users and clients.
However, remarks cannot be avoided. The authors claim that platforms are "data-rich" environments where negative impacts can be measured directly, but they also point out that regulators are constrained by the platforms' failure to regularly disclose key performance indicators. This is not a strict logical failure (one could say, "the data exists, but there is no access"), but while the measurement is "direct," without access to the platform's internal metrics and experiments, it may in reality be indirect and incomplete.
It follows that the practical applicability of the tests proposed in the book depends heavily on future disclosure rules, and this is a political issue, which the book presents more as a conclusion than as a well-developed mechanism.
In their measurement methodology, the authors write that if platforms "promised and demonstrated" that the goal of their algorithms is to optimally distribute attention, then a discrepancy between organic and paid search results is a sign of rent-seeking. However, even if a platform once attracted users with the promise of "better search results," in reality, it may consider its goal to be mixed optimization (combining quality and revenue), in which case the discrepancy between organic and paid search results will not necessarily be "abuse" but a result of the chosen business model. To confidently call this a rent, it is necessary to further prove that the user actually expected one thing, received another, and that the quality actually worsened (and did not simply change).
Moreover, the authors link rent to deviations from the "best attention allocations" the platform is capable of and that it supposedly promises. However, in a practical sense, "best" is a slippery concept: it varies for different users, contexts, and tasks. As a result, "best" easily transforms from a measurable criterion into a normative "how things should be." This is important because the entire "low-quality allocations" construct requires a clear way to separate "worse for the user" from "simply different."
Finally, the book presents a well-articulated problem, but its final conclusions serve more as a program for further research than a set of actionable criteria "for today." This is explicitly acknowledged and is perfectly normal for scientific publications, but the average reader may have a feeling of incompleteness: the problem is clear, but the "what to do" remains at the level of directions and principles.
In summary, the book discusses digital platforms as systems that manage attention and can extract "excess revenue" through their role as intermediaries. This topic continues the discussion of how platforms use data, and in this sense, there is a pleasant sense of evolution in the research. In this sense, it will be of interest not only to academics but also to anyone interested in how "free" platforms profit from users, the logic behind platform market power, and how our attention is turned into a "commodity."
At the same time, key assumptions ("what exactly the platform promised," "what constitutes the best search results," "when a discrepancy between organic and advertising is abuse rather than a feature of the model") require stricter criteria, otherwise the conclusions easily become controversial. Those expecting a ready-made set of specific regulatory solutions and clear, one-step tests should consider: a significant portion of the second half is precisely the formulation of the problem and an invitation to further research.