Advertising Identifiers: How Digital Advertising Knows What to Show You

In digital advertising, there is a paradox that is rarely discussed publicly, yet it defines almost everything. The market may not know who you are. But it is critically important for it to understand that you are the one again.

It is precisely this ability to recognize a user a second time that forms the foundation of virtually all the mechanisms we have become accustomed to: retargeting, personalization, frequency control, and attribution. Without it, advertising does not disappear, but it very quickly returns to a state of mass, almost blind reach - expensive and not particularly effective.

In this sense, identification is the fuel that powers the entire system.

TL;DR

• Advertising identifiers allow systems to understand that they are dealing with the same user or device, rather than a new random impression;
• On the web, this role has long been performed by cookies, but their stability is rapidly declining due to browser restrictions and privacy policies;
• In mobile environments, advertising IDs (IDFA, GAID) are used, which for a long time were significantly more reliable thanks to their ecosystem-level binding;
• No alternative model (fingerprinting, universal IDs) has become a full-fledged replacement;
• Any form of identification operates with a margin of error: one user can look like several, and several can look like one;
• The market is shifting from total tracking to models where there are fewer signals and more uncertainty;
• The future lies in the transformation and increasing complexity of identification.

An advertising identifier is a method that helps a system “recognize” a user or device. Thanks to it, ads can be delivered more precisely, avoid being shown too frequently, and help measure how effectively a campaign has performed.

What Advertising Identifiers Really Are

It’s important to clear up one of the biggest misconceptions right away: an identifier is not a passport. It’s simply a mechanism for repeated recognition.

A system does not necessarily need to know who exactly opened an app. What matters is recognizing that this is the same smartphone or browser where an ad was shown, for example, yesterday. That is essentially how the modern advertising market works.

How Digital Advertising Evolved Toward This Model

To understand the role of advertising identifiers, it helps to remember how the early web worked.

Internet advertising used to be far simpler but also far cruder. Banner ads were purchased not to reach a specific audience, but simply to occupy visible space on a website. The problem quickly became obvious: the system could barely tell whether it was showing an ad to a new person or to the same user for the tenth time.

Cookies became the first major answer to this challenge. They gave the market a basic form of memory.

Thanks to cookies, advertising platforms gained the ability to:

• limit ad frequency;
• build behavioral audience segments;
• launch retargeting campaigns;
• measure effectiveness.

In practice, cookies transformed advertising from a mass-broadcast phenomenon into a system capable of “remembering.”

Why Cookies Are No Longer Enough

The main limitation of cookies is simple: they do not remember you; they remember the browser you use. As a result, for an advertising system, you using Chrome on a laptop, Safari on a phone, or Firefox on a work computer may look like completely different people.

The situation is made even more fragile by the fact that cookies are easily lost: they can be deleted, blocked, or simply lost due to browser settings. For a long time, the industry largely ignored these weaknesses because even an imperfect system still allowed digital advertising to function.

The reality also varies dramatically across browsers. In Chrome, identification remains relatively stable thanks to cookies, while in Safari it is almost nonexistent, and in many other browsers it is significantly weaker. In other words, part of the audience is already cut off from traditional targeting mechanisms — not because of user behavior, but because of technical limitations.

This leads us to a clear conclusion: cookies can no longer serve as a universal identifier:

1. In Chrome, most advertising requests still contain a persistent identifier — at roughly 70–80%.
2. In Firefox, that figure drops to around 30–35%.
3. In Opera, it ranges between 50–60%.
4. Safari represents a fundamentally different situation: stable cookie-based identification is practically nonexistent.

This means that part of the audience effectively “disappears” for advertising systems — not because of user behavior, but because of the technical environment in which users operate.

Why Apps Couldn’t Rely on Cookies

When digital audiences began shifting not just to mobile devices, but specifically into app ecosystems, it became obvious that the web-based identification model would no longer work because an app is not a browser.

Cookies were originally designed for the web environment, where users move between websites through a browser. Inside apps, however, this model simply could not scale with the same level of efficiency.

At that point, the industry faced a challenge it had already encountered before: user behavior had evolved faster than the advertising infrastructure built to support it.

As a result, apps gradually developed their own identification framework: mobile advertising IDs such as IDFA and GAID.

The underlying principle remained the same: the system does not necessarily need to know who the person behind the smartphone is. What matters is recognizing the relevant user or device to:

• control ad frequency;
• avoid chaotic ad duplication;
• build audience segments;
• measure campaign effectiveness.

In the app ecosystem, advertising identifiers ceased to be just a useful addition. Without them, in-app advertising simply could not function as a полноценная system.

How Apple and Google Control Advertising Identifiers in Apps

The moment a user downloads an app through the App Store or Google Play, they enter an ecosystem where access to advertising data is controlled by the platform itself.

That is why IDFA (Identifier for Advertisers) and GAID (Google Advertising ID) were introduced as special identifiers that Apple and Google designed as a kind of compromise, since they operate independently of both device hardware and browsers. They provide the advertising industry with the minimum level of data needed to function, while still keeping ultimate control in the hands of the platforms.

Put simply: an app cannot freely access arbitrary user data — everything operates within the rules defined by the operating system.

The process works roughly like this:

1. The user opens an app;
2. An SDK (a built-in technical module responsible for communication with advertising systems) generates an ad request;
3. The operating system determines which advertising ID is available;
4. This signal is passed into the advertising ecosystem;
5. DSPs and SSPs then decide how to use it for bidding, frequency control, or attribution.

As a result, Apple and Google have become far more than just app stores — they are now the effective gatekeepers of advertising identity.

For the industry, this became a turning point.

Control over identity is not just a privacy issue; it is a question of market power. Companies that have direct relationships with users through logins, email addresses, or phone numbers gain an entirely different level of control over data. That is precisely why platforms like Google and Meta occupy a much stronger position compared to independent players in the advertising ecosystem.

Why First-Party Data Doesn’t Fully Solve the Problem

At first glance, it may seem that the industry has already found the answer: first-party data. If a brand has a user’s email address or phone number, then identity resolution should no longer be an issue. But in practice, a much more complex problem emerges: how to connect that data to a specific ad impression.

On the web, systems usually do not know exactly which user has opened a page. What they see is a browser, a cookie, or another technical signal: not an email address or a phone number.

That is precisely why most market participants cannot effectively activate their own data within advertising systems. Only platforms that control the user entry point, such as Google or Meta, where users log in and provide their personal information, can do this at scale.

Everyone else is forced to rely on far weaker signals.

The challenge becomes even more severe when trying to match users across the web and mobile apps. On the web, systems see cookies or browser-level identifiers; in mobile environments, they see advertising IDs. But these are fundamentally different identifiers, and in most cases, they are not linked to one another.

As a result, the same person may appear as two completely different users — one in a browser and another inside a mobile game or app. Independent players have no reliable way to “stitch” these identities together.

True cross-device matching at scale is realistically available only to Big Tech platforms that can rely on logins, email addresses, or phone numbers. The rest of the market simply does not operate under the same conditions.

Did the Industry Try to Bypass These Limitations?

When it became clear that cookies and traditional identifiers were losing effectiveness, the industry began searching for alternative solutions.

One of the most radical approaches was a technique known as fingerprinting.

Its logic is fundamentally different. Instead of relying on a single identifier, the system attempts to assemble a unique combination of device characteristics: language settings, browser type, installed fonts, extensions, IP address, operating system, and even the specific way a graphics card renders images. Together, these signals form a so-called “device fingerprint” — a profile that can, with high probability, recognize a user even without cookies or advertising IDs.

At first glance, this seems like the perfect solution. In practice, however, it turned out to be far more problematic.

1. First, fingerprinting is expensive. Collecting and processing large volumes of device parameters for every advertising request creates significant technical overhead.
2. Second, it is not perfectly accurate. Different users may share very similar device characteristics, causing the system to merge multiple people into a single profile. In other cases, the opposite happens: one user is fragmented into several separate identities.
3. But the biggest issue is privacy.

Fingerprinting effectively bypasses user control mechanisms. It is harder to disable, operates largely invisibly, and fundamentally conflicts with the idea of transparent user consent for tracking.

That is why platforms and regulators have gradually moved to restrict fingerprinting — either directly or indirectly through stricter data-access policies. As a result, the technology never became a universal standard for advertising.

Today, fingerprinting is more commonly used in areas where security matters more than marketing: banking services, e-commerce, and fraud prevention systems.

The reason is simple: fingerprinting was never truly designed to operate within the rules of the ecosystem — it was designed to work around them. And in digital environments, models built around bypassing the rules rarely become sustainable long-term standards.

Why Identification Is Also About Infrastructure — Not Just IDs

In advertising, identification is not only about the identifier itself, but also about the infrastructure built around it. A good example is advertising pixels — snippets of code installed on websites that allow platforms to connect user behavior across different environments.

This is how systems learn not only that a user visited a site, but also which products they viewed, what they added to their cart, and what they ultimately purchased. In this sense, identification becomes far more than a technical marker — it becomes part of a complex system for collecting and processing behavioral signals.

Most retargeting scenarios are built precisely around these pixels:

1. A user interacts with a website;
2. The system records that behavior;
3. The data is then used to serve ads in other environments.

There is, however, an important nuance: this logic only works where the relevant code has been installed. Without pixels or integrations, the system simply cannot “see” the user and therefore cannot use the associated behavioral data.

How Advertising Identifiers Work Today in the App Store and Google Play Ecosystems

Imagine a simple scenario. A user owns both an Android smartphone and an Android tablet. On the smartphone, they sign into Google Play using their Google account, download a game, and register within it. At that moment, the system receives an advertising ID associated with that ecosystem account.

Later, the same user picks up the tablet, signs into Google Play with the same account, and downloads a different app. From the perspective of the advertising system, these are not two unrelated installs on two separate devices. They represent the same user operating within a single account ecosystem.

This is where one of the key advantages of Apple’s and Google’s platform-based approach becomes apparent: they can provide far more accurate user recognition across multiple devices within their own ecosystems.

Put simply, if a person interacts with several devices under the same account, the advertising system understands that this is the same audience. This is the foundation of cross-device logic.

For brands, this capability is critically important because it allows them to build not just isolated ad impressions, but a coherent and continuous advertising experience across devices.

How an App Sends Signals to the Advertising System

When discussing advertising identifiers in the in-app environment, it is important to understand one thing: an app does not “know the user” in the traditional sense. Its role is far more practical — to pass enough signals into the advertising ecosystem so that the system can understand which device or environment it is interacting with.

These signals may include:

• an advertising ID;
• technical device parameters;
• the app type or category;
• basic session-level signals.

From there, the logic closely resembles web programmatic advertising: the request is sent to an SSP, the DSP receives the signal, evaluates it in the context of active campaigns, and decides whether the impression is worth bidding on.

The key point is that the identifier itself is not a “portrait of a person.” It is a technical mechanism for recognizing a user or device within the advertising ecosystem.

What Happens When Identification Is No Longer Available

One of the biggest myths in digital advertising is the assumption that advertising systems always know exactly who the user is. In reality, that is often not the case.

In environments such as Safari or anywhere identifiers are unavailable, advertising platforms are forced to work with whatever signals remain: IP addresses, device type, behavioral context, or the content of the page itself.

Naturally, this model is less precise. But it does not stop the system from functioning: it simply shifts it from deterministic identification to probabilistic prediction.

Put simply, when a user cannot be recognized directly, algorithms attempt to “infer” who they might be. In that sense, a significant part of modern advertising no longer operates on exact identity, but on intelligent assumptions.

Privacy-First: The Moment the Rules Began to Change

The past few years have become a turning point for the advertising industry. GDPR, Apple’s App Tracking Transparency (ATT), and the decline of third-party cookies have fundamentally changed the core principle of digital identification.

Where identification was once the default, it increasingly now depends on explicit user consent.

This does not mean the death of advertising identifiers — but it does mean they are evolving. The market is gradually moving away from direct, often aggressive tracking toward more sophisticated models built around first-party data, clean rooms, and other privacy-oriented approaches.

In practice, the industry is learning how to operate more intelligently with fewer available signals.

For agencies and brands, this means one important thing: the era of extremely simple targeting is gradually coming to an end.

It is no longer enough to simply “find the user.” Companies now need to understand:

• which signals are actually available;
• how reliable those signals are;
• and how to balance effectiveness with evolving privacy expectations.

That is why identity strategy is increasingly becoming not just a technical issue, but a business one.

And this is where another important shift emerges — one that is not always immediately obvious. In the past, the main challenge was finding the user. Today, the challenge is understanding what to do when you can only see them partially.

That changes the very logic of modern marketing.

Conclusion

Advertising identifiers are not just a technical detail or a backend concern reserved for specialists. They shape how accurately brands can work with audiences, how efficiently advertising budgets are spent, and how personalized communication can become.

And although the system is now evolving under the pressure of a privacy-first reality, the core principle remains unchanged: for advertising to be effective, the market must still be able ( in some form) to remember who it has already interacted with.

If you want to make sure your advertising truly reaches the right audience, contact Fusify.