Why Internet Advertising Inevitably Changes Organic Exposure

Internet advertising rests on a basic commercial logic. Why are advertisers willing to pay? Because a message that relies solely on ordinary searches, follows, reading and sharing may not attract enough attention. What advertisers purchase is precisely the opportunity for exposure they might otherwise never receive.

Even without paid promotion, information does not circulate evenly. Users have their own interests and networks of people they follow. Platforms need mechanisms for searching, ranking and recommending content, and some content naturally receives more exposure than others. Organic circulation therefore does not mean circulation without algorithms, nor does it imply equal or fair exposure. Advertising changes this arrangement by introducing the ability to pay into the existing allocation of visibility.

A simple mathematical model can illustrate the process.

Suppose that, over a given period, a platform can provide a total of V effective impressions, and individual messages originally receive x₁, x₂, …, xₙ impressions. Then:

x₁ + x₂ + … + xₙ = V

An advertiser now pays for one message to receive Δ more impressions than it would have received without that payment. If V remains constant, the total impressions available to all other messages must fall by Δ. In other words, the additional visibility purchased for the promoted message takes up attention that could otherwise have gone to other content.

In practice, platforms can increase V by attracting more users, extending the time they spend on the service, and similar means. Yet users’ attention remains limited within any particular period. What this model establishes is that, whenever paid promotion actually increases a particular message’s exposure, advertising changes the allocation that would otherwise have occurred. It does not establish that advertising necessarily widens overall disparities in exposure, still less that all advertisements lack informational value. We need to distinguish the claim that an intervention changes the allocation from the claim that it makes the allocation more unequal.

There is a further issue. Advertisers usually purchase more than a chance appearance: they want predictable, measurable commercial results. They need to know how many people an advertisement can reach, how many clicks it can attract, and whether it can generate transactions. To meet these expectations, platforms must continually analyse user behaviour and adjust where, when and to whom content is shown, improving the efficiency of advertising delivery.

This gives platforms a continuing commercial incentive to intervene more deeply in the circulation of information. An algorithm may initially help users organise content from the accounts they follow. Later, it begins to predict their interests, and then identifies audiences more likely to respond to advertisements. The algorithm itself is a technical means. The business model that monetises exposure and attention drives the change.

Sohu Video’s “Following Feed” offers a concrete example. Charles Zhang hopes users will obtain more of their information through accounts they actively choose to follow and content shared by real people, reducing the platform’s forceful direction of attention. There are practical grounds for this aim: excessive recommendation can indeed weaken users’ initiative in choosing what information to consume.

But if the Following Feed ultimately continues to depend mainly on advertising, the tension remains. A user may neither follow a particular business nor encounter its content through a friend’s repost, yet the business may still want to pay for more exposure. If the platform strictly respects existing follow relationships, advertising reach is constrained. If it promises additional reach, it must intervene in content allocation. This is more than a choice between two algorithms: it is a structural conflict between organic follow relationships and paid exposure.

Sohu has spent years exploring social products of this kind, and the Following Feed continues to emphasise cultivating its ecosystem. According to Sohu’s 2025 financial results, online games accounted for about 87% of the group’s total revenue, while the Sohu segment, which includes media and other businesses, continued to report an operating loss. This indicates that Sohu has other businesses and financial resources to support long-term exploration. It does not allow us to calculate the Following Feed’s own profit or loss directly, or attribute the entire segment’s losses to it.

The Following Feed could, of course, keep advertising to a lower proportion of its content, or develop subscriptions and professional services. These paths are possible, but their sustainability still has to be demonstrated through revenue and costs. If the service ultimately relies on advertising at scale, it will face the same economic constraints as other internet platforms: higher advertising revenue requires enough exposure to sell and an effective way of allocating it.

The fundamental issue with internet advertising, then, is that commercial payment becomes one factor determining information’s visibility. The use of algorithms alone does not explain the problem. Advertisements can provide useful information, and they can be constrained by clear labelling, frequency limits and users’ freedom of choice. Yet as long as a platform sells additional exposure, what users actually see cannot be determined entirely by the organic relationships through which information would otherwise circulate.

This is the tension between an advertising-funded business model and the organic circulation of information that cannot be eliminated altogether.