Tag PilotThe basics of tracking
Events
When a user lands on your website, they trigger a number of events. Some of them match user actions one to one, such as:
- page view,
- form submission,
- add to cart.
Other events are generated automatically when user behaviour matches predefined criteria, such as:
- session start,
- user engagement,
- user data.
The exact rules for how these events are defined and captured depend on the tracking solution. There are three main ways of capturing events:
- pixel auto-capture,
- tag managers (such as Google Tag Manager),
- the dataLayer.
These events become the data source and the foundation for analytics and conversion tracking. Each event can carry additional information about the user’s action: besides its name, it can carry data about the products the user interacted with or the form fields they filled in. These data points are called event parameters, and they are essential for more advanced tracking scenarios.
Events happen in the user’s browser. To analyse them, or to optimise advertising campaigns based on them, they need to be sent to the tracking platform. That’s why pixels were created.
Pixels
The name “pixel” is historical. In the early days of internet advertising, when third-party cookies were common practice, most platforms used tiny, transparent images (1 pixel big… or small) that loaded within the tracked page and allowed third-party cookies to be created. Other solutions used iframes, but the idea was the same.
Over the years, pixels evolved from simple images into small JavaScript libraries that can do more, such as capturing common events or parameters from the browser.
As third-party cookies are blocked by more and more browsers and ad blockers, 1-pixel images are rarely used anymore, but the name stuck.
Installing a pixel, or more precisely its JavaScript code, on a website creates a general connection between every visitor’s browser and the tracking platform. It involves creating or reading identifiers stored in cookies, performing basic tracking (page views) or enabling auto-capture of common events.
Pixels are very generic pieces of software that need to work on virtually any website, so even with auto-capture the event coverage is very limited. Installing a pixel won’t capture more complex events, like ecommerce actions with detailed product information.
Additionally, each pixel works slightly differently, and the tracked data can vary a lot between platforms, so not all events will be tracked.
Tags
To recap: an event is something that happens on a website. A pixel is a piece of code that connects browser actions with the advertising platform’s systems and databases. Tags are the bridge between the two.
In other words, a tag is an additional, small piece of code dedicated to a specific website and a specific action, such as add to cart. It defines precisely which user actions are tracked.
Unlike pixels, which are generic and work on all websites, tags are custom to one website, so they can track events and parameters more precisely.
Think about all the possible events and content types tags need to handle on a modern website, and managing them becomes challenging, especially because every change to a tag needs to be deployed to the page. That’s why tag managers such as Google Tag Manager were created. Read about them in the next chapter.
Tracking vs tagging
Tracking and tagging are often used interchangeably because their meanings are so close, but there is a distinct difference.
Tracking simply means that information about a user action was sent to a tracking system. That can be Google Analytics 4, Google Ads, Meta or any other analytics or advertising platform that accepts a stream of data. Tracking can happen through built-in pixel capabilities or through dedicated tags.
Tagging, on the other hand, means that all relevant tags were placed on a given website. Those tags, if implemented correctly, should ensure good tracking coverage.
In other words, tracking is the end goal of capturing data about user behaviour, and tagging is a necessary step to get that data.