Audience Segments. How Digital Advertising Shows Us What We Want

It seems that online advertising knows everything about us: you've just been looking for flights, and now you see banners with tours, and you're thinking about buying sneakers - advertising reads minds. In fact, there is no magic here. Advertising technologies do not know a specific person, they work with groups.

An audience segment is a generalized portrait of a group of users united by similar characteristics. Advertising technologies don't work with specific people, they work with probabilities and models. The segments answer a key marketing question: who should I show the message to right now in order for it to be relevant.

Audience segmentation is based on functions that help describe user behavior and interests. In practice, demographic segments (age, gender, marital status), geographical (country, region, city), behavioral (sites visited, actions, interest in product categories), interests, and technical parameters such as device type, operating system, smartphone or TV model are more often used.

It is important to understand that a segment is not a list of people or a database of personal data. This is a statistical group formed on the basis of recurring patterns. The advertising system doesn't "know" who is on the other side of the screen, but it most likely understands which audience group it belongs to.

Segment formation is a multi—stage technological process. The main signals come from websites and apps: page views, clicks, and time spent interacting with content. The technical parameters of the devices and the environment complete the picture. This information is processed in the software infrastructure and converted into segments ready for use in advertising. Data management platforms, DMP, are responsible for this.

DMPs accumulate disparate signals from various sources and lead to a single representation: behavioral data, device specifications, as well as data from advertisers and platforms received from top officials. All data is anonymized, cleaned, normalized, and combined using advertising identifiers.

As a result, DMP does not store "users", but forms stable behavioral profiles and segments. These segments are then used in the DSP to set up targeting. In fact, DMP acts as a link between the chaotic flow of data and a managed advertising strategy.

After the segments are formed, they are transferred to the DSP platforms through which the purchase of advertising is carried out. Here, the advertiser selects the audience, controls the bids, frequency of impressions, and scale of campaigns.

The modern market is moving towards an ecosystem approach, in which data collection, segmentation, procurement and analytics are combined into a single technological chain. This reduces information loss and allows you to adapt faster to changes in the advertising environment.

Modern DSPs also try to create their own DMP, but it is usually more reliable to diversify data sources. For example, UMG uDSP uses data from UMG DMP & AmberData DMP.

Walled Gardens, which are closed advertising environments, occupy a separate place in the ecosystem. In such systems, audience data, segmentation logic, and ad display algorithms remain inside the platform and are not conservatively displayed outside.

For advertisers, this means access to a large audience, but with limited transparency. Segments are managed according to the rules of the ecosystem, and moving the audience outside of it is usually impossible.

Therefore, there is a growing interest in independent AdTech ecosystems and software solutions that flexibly work with segments, combine data sources and are not tied to a single closed environment.

Smart TV is becoming the fastest growing digital advertising channel around the world, and the segmentation logic here differs from web and mobile applications. There are practically no cookies in the TV environment, and identification is based on the device and the context of its use.

Segments are formed based on the TV model, operating system, region, viewing time, content genres, installed applications, connected devices, network environment, and screen usage scenarios. In fact, Smart TV does not work with an individual user, but with a household. For example, the same TV can be included in the family viewing segment in the evening and in the background content segment in the afternoon.

Most of the data transmitted through this channel is concentrated in the ecosystems of device manufacturers and OTT platforms, which brings Smart TV closer to the Walled Gardens model. More and more TV manufacturers are stopping producing TVs without Smart TVs. At the same time, the development of artificial intelligence has reduced the cost of their support and production.

Already, most TVs are sold at a big discount, since the terms of use already include permission for situational analysis of the screen content in order to obtain advertising data. At the same time, for advertisers, it is a channel with a high level of attention, low banner blindness and steady growth in the coming years.

Audience segments have become the foundation of modern online advertising. They allow brands to more accurately reach their audience, agencies to manage their performance, and the market as a whole to develop their own technological solutions.

The development of independent advertising ecosystems, integration with large data sources, and the expansion of channels such as Smart TV show that segmentation will only become more complex and more intelligent.