GA4 can collect events while still being poorly prepared for the questions your team needs to answer. A working tag is only one part of the setup.
Before collecting production data, decide what belongs in the property, which interactions matter, and how long you need detailed history. Some decisions affect what you can investigate later. Others can permanently remove incoming data.
Define the reporting boundary first
An account organises access and properties. A property is the reporting boundary. A data stream supplies data from a website or app. Choose the boundary from the business question before creating streams for individual pages or teams.
Consider a hypothetical training business with one website, two course categories, and a separate development site. Its question is: which course category receives more completed enquiries? For this example, one production web stream can cover both categories. A category is a reporting field, not a reason to create another stream.
Keep development activity separate from production reporting. Record the production hostname, property ID, stream name, and measurement ID. Give someone responsibility for access and future changes. This makes a wrong destination easier to identify when a tag appears to work.
Choose the reporting time zone and currency with the reporting owner. A visit near midnight can fall on different dates in different time zones. Google’s setup guide explains the property and stream steps, including the effects of a time-zone change.
Choose events from the question you need to answer
For the training business, define a completed enquiry as a successful form submission accepted by the website. A click on the submit button is insufficient: validation or the server can still reject it. Record that distinction before implementation.
