Dimensions, metrics and scope

Every table in GA4 is built from dimensions (text attributes: country, page path, event name, session source) and metrics (numbers: users, sessions, event count, revenue). You pick dimensions to break the data down and metrics to measure it.

Each dimension and metric has a scope, meaning which object it describes:

• User scope: attributes of the person, e.g. First user source, Age, Lifetime value, user properties.
• Session scope: attributes of a visit, e.g. Session source / medium, Session default channel group, Landing page, metric Sessions.
• Event scope: attributes of a single event, e.g. Event name, Page location, Link URL, metric Event count; most custom dimensions live here.
• Item scope: attributes of a product in an ecommerce items array, e.g. Item name, Item brand, metric Item revenue.

Combining scopes is allowed but can be meaningless. Breaking Sessions down by Event name asks "how many sessions contained each event", which is fine; breaking Event count for purchase down by Landing page is also fine because the landing page is attached to the session the event belongs to. But a session-scoped dimension against a user-scoped metric, or an item-scoped dimension against an event-scoped metric, produces either (not set) or inflated numbers because the join is not one-to-one.

A practical rule: start from the metric you want, then only add dimensions at the same scope or a wider one (event inside session inside user). GA4's dimension picker in Explore marks incompatible combinations, and standard reports simply refuse to add them.

(not set) in a row usually means the dimension had no value for those events, not that data was lost.