What “tracking” means

Tracking is the process of collecting and using information about online activity so a system can recognize patterns, measure behavior, or support features like personalization, analytics, or advertising. In practice, tracking often relies on signals that persist across time (for example, identifiers stored in a browser) or signals that can be correlated (for example, request metadata).

Tracking is not one single technique. It can be done for different purposes, such as:

  • Analytics and performance measurement (understanding what people do)
  • Advertising and targeting (showing relevant content)
  • Security and fraud prevention (detecting suspicious behavior)

How tracking works in the browser and apps

Many tracking flows start when your device makes a request to a website or an app. The system may then use one or more of the following:

  1. Stored identifiers
  • Browser cookies and similar storage can contain identifiers that let a service recognize returning visits.
  • Some identifiers are first-party (set by the site you’re visiting) and others can be third-party (set by a different domain).
  1. Request and device signals
  • Network requests include metadata such as headers, URL paths, and timestamps.
  • Device and browser characteristics can help systems differentiate clients, even when identifiers are limited.
  1. Consent and configuration
  • Many tracking components depend on whether you accepted certain categories (e.g., marketing or analytics).
  • If consent is denied or tracking is restricted, the system may fall back to less-identifying measurement or stop storing persistent identifiers.
  1. Combining signals
  • Even when direct identifiers are removed, some tracking may still infer relationships from multiple signals. The exact outcome depends on the environment and the tracking design, so you may see partial tracking or reduced accuracy.

Key limitations and what tracking cannot reliably guarantee

Tracking is constrained by both user controls and technical realities:

  • Privacy settings and blocking: Browser settings, extensions, or “strict” tracking protections can reduce persistence of identifiers.
  • Consent choices: Opting out can limit which purposes receive identifiers or event data.
  • Incomplete coverage: Not every request is trackable; some environments restrict storage or cross-site recognition.
  • Changing identifiers: Identifiers can be rotated or cleared, which breaks long-term continuity.

So while tracking can often describe trends and interactions, it does not necessarily produce perfect, long-lasting identification. If you observe less stable behavior (for example, resets after browser changes), that’s often a sign of these limitations rather than a complete absence of tracking.

Practical checks you can run

You can validate what’s happening without needing special tools beyond your browser:

  • Check stored data: Open your browser’s site data/storage settings and look for cookies or identifiers set by the domains you visited.
  • Inspect network calls: Use your browser’s developer tools to view outgoing requests and note which third-party domains are contacted.
  • Review consent prompts: If a site offers toggles for analytics or marketing, compare what changes after you allow vs. deny.
  • Compare behavior across sessions: If identifiers persist, you may see returning behavior; if they reset, it indicates restricted persistence.
  • Test with tracking protection enabled: Temporarily enable your browser’s privacy controls and observe whether request patterns and stored data decrease.

If you’re trying to distinguish tracking types, look at the purpose-labeled scripts or endpoints (analytics vs. advertising) and whether they set persistent storage.

To place tracking in context, it helps to separate similar terms:

  • Analytics: Often focuses on measurement of usage rather than targeting.
  • Advertising targeting: Often uses tracking data to tailor what is shown.
  • Fingerprinting: Uses combinations of attributes to distinguish clients when identifiers are limited.
  • Attribution: Connects events to campaigns or sources, which can involve tracking plus additional logic.

These concepts overlap, but they differ in goals and in how directly they rely on persistent identifiers.