Definition
Event-based analytics records each interaction as an event with a name and a set of name and value pairs.
Those pairs are the event parameters. A generate_lead event might carry a form name, the page it was on and
an estimated value; a purchase event carries a transaction identifier, a value, a currency and a list of
items. In GA4Google's event-based analytics platform. they are called event parameters; other analytics tools often call the same idea
event properties.
A parameter belongs to the event it travels with. It describes that one occurrence and nothing else, which is what lets a single event name, such as a form submission, stand for many different forms across a site while remaining separable in analysis.
Automatic, recommended and custom parameters
- Automatic parameters arrive with every event without configuration: the page location, the page title, the referrer, the language and similar context. They answer where and in what setting an event occurred.
- Recommended parameters belong to the events Google publishes as recommended, such as sign-ups, leads and e-commerce steps. Each comes with documented parameter names, for example a value and a currency, and using those names exactly is what lets built-in reports and integrations read them.
- Custom parameters are whatever else the business decides to send: the course a form concerned, the plan a visitor chose, the calculator a visitor used. They carry the detail that makes the data specific to the business.
The guide to GA4 events and parameters sets out how the three kinds fit into one schema. For e-commerce, product-level detail travels in a separate item list inside the event, and the e-commerce tracking tutorial covers how those item fields relate to the event's own parameters.
In practice: sending parameters through the tag or data layer
On a website, parameter values usually start in the page or the application: the server knows which product is shown and what the basket is worth. The site writes those values into the Data LayerA structured object on a web page that passes values to a tag manager for tags to use. at the moment of the interaction, and the tag reads them from there and attaches them to the event. Reading values from the data layer, rather than scraping them from page text, keeps them stable when designers change the layout. The tutorial on sending GA4 through GTM builds event tags in exactly this way, with each parameter mapped to a data layer variable.
Apps follow the same pattern through their software development kit, with one addition: campaign detail from the link that opened the app has to be carried into events as parameters, or behaviour cannot be split by source. The guide to deep links and app analytics events treats that hand-over as a common place for data to go missing.
Whatever the route, a parameter is only proven when it is seen arriving with the expected value. The tracking verification checklist checks each event in the analytics debug view with its parameters attached, before anyone builds a report on them. Collected parameters are kept even when no report shows them, and the BigQuery export basics guide explains how the raw export holds every parameter for analysis the interface cannot do.
Naming and consistency
Parameter names are case-sensitive and permanent in practice: once reports, audiences and exports depend on a
name, changing it splits history in two. A convention set in the tracking plan, usually lower-case words
joined by underscores to match the recommended names, prevents the same idea arriving as formName on one
page and form_name on another. Values need the same care. A form called "Contact" on one template and
"contact-us" on another becomes two values in every report.
What not to send, and the limits
Never put personal data in a parameter. An email address, a name, a phone number or the text a visitor typed into a free-text field can reach the analytics property through a careless form tag or a page address that carries it. Google Analytics' terms prohibit sending personally identifiable information to it, and doing so can also breach privacy law and the consent the visitor gave, so the tracking plan should say which values may be sent and the verification step should look for stray personal data in the debug view. If a record has to be tied to a person, send an internal identifier that means nothing outside the business's own systems, and have the privacy owner confirm it is acceptable.
Parameters are also limited. GA4 caps how many parameters one event can carry and how many custom dimensions a property can register, and names and values have length limits; the exact numbers are in Google's documentation and can change, so check the platform's current documentation before designing a schema. Register a parameter as a custom dimension only when its values come from a bounded list, such as the cottage name or form name in the example below. A parameter with a very large set of distinct values, such as a search phrase, a page address with identifiers in it or a timestamp, can push the less common values into an "(other)" row in reports, which hides the detail it was registered to show.
Parameters versus user properties
A parameter describes an event; a user property describes a person. Customer tier, sign-up month or logged-in status belong to the user and should persist across that person's events, so they are set as user properties rather than repeated as parameters on every hit. Sending person-level facts as event parameters makes user-level questions hard to answer, and sending event-level facts as user properties overwrites them with each new event.
A cottage agency learns which cottages drive bookings
Hearthwold Holiday Cottages let 18 cottages and tracked each booking-form submission as one event,
booking_request, with no parameters beyond the page. In a year it recorded 2,150 requests but could not say
which cottages or seasons they were for, because the form sat on a shared booking page. The agency added two
parameters, cottage_name and arrival_month, pushed to the data layer when the form was sent, with names
fixed in the tracking plan and values taken from the booking system's own list. Within a season the data
showed that 4 cottages drew 41% of requests and that requests for the three smallest cottages came mostly for
spring and autumn arrivals, which changed where the agency spent its advertising.
Related terms
GA4Google's event-based analytics platform. is built on events and their parameters. A Custom DimensionA GA4 reporting field created by registering an event, user or item parameter so it appears in reports and explorations. is the registration that makes a custom parameter usable in GA4's reports. The Data LayerA structured object on a web page that passes values to a tag manager for tags to use. is where a site normally supplies parameter values to the tag. A Key EventA GA4 event marked as important to the business, the property's measure of success. usually needs the richest parameters, because its value, source and type are what reports on outcomes are built from.
Common confusions
- The event name should carry the detail. Separate event names for every form or product multiply the schema; one event with a parameter is easier to analyse.
- Parameters and user properties are interchangeable. One describes an action, the other a person.
- A parameter in the debug view is in the reports. Custom parameters need registering before reports show them.
- Names can be tidied later. Renaming a parameter breaks continuity with everything collected before.
- More parameters are always better. Each one needs a defined meaning and consistent values; unused ones add maintenance without adding insight.
- Any value can be sent as a parameter. Personal data does not belong in one, and a registered dimension needs values drawn from a bounded list.
