Guide
Why GA4 configuration is hard to hold in your head
The GA4 admin section is organised by object type, not by task. Data streams live in one place, key events in another, custom definitions in a third, data retention behind a settings link, filters somewhere else again, and product links in their own section. Every one of them is two or three clicks from the others.
The result is that almost nobody can answer basic questions about a property from memory. What is the retention window? Are there active filters, or just ones stuck in Testing? How many custom dimension slots are left? Each answer is a small expedition, and the questions come up constantly.
One screen, several jobs
- Configuration snapshot
- Everything that defines how the property behaves, rendered together. This is the view you want when you inherit a property, when a client asks what is set up, or when you are trying to explain why a report looks the way it does.
- Inline editing
- Writable settings change in place. Fixing data retention, a timezone, or a currency stops being a navigation exercise. Read-only fields stay visibly read-only, so it is always clear what the API will and will not let you change.
- Key reports
- The handful of reports you actually check when validating a property — traffic, key events, and their trends — without switching to the GA4 interface and losing the configuration context you were just reading.
- Annotations
- Dated notes attached to the property: launches, replatforms, tracking changes, incidents, campaign starts. This is the institutional memory that otherwise lives in one person head, and it is the difference between "traffic dropped in April" and "traffic dropped in April because we replatformed on the 14th".
Annotations are the underrated part
Six months from now, someone will look at a chart and ask what happened in April. Without annotations the answer is a conversation with whoever was there at the time, if they still work with you. With annotations it is a line on the property.
The events worth recording are the ones that change what the data means rather than what the business did:
- Site launches, replatforms, and major redesigns — the single biggest source of unexplained metric shifts
- Tracking changes: a new container version, a renamed event, a consent banner update
- Outages and incidents, including the ones where tracking broke rather than the site
- Campaign starts and stops, particularly large paid pushes
- GA4 configuration changes: a new filter going active, retention extended, attribution model changed
The last category matters more than it sounds. A filter switching from Testing to Active changes the numbers on the day it happens, and it looks exactly like a traffic drop to anyone reading the chart later.
Manager, audit, or detector?
These three tools cover different questions about the same property, and it is worth being clear which one answers what.
The property manager answers
- What is this property configured to do?
- Which settings can I change, and can I change them here?
- What happened on this date that explains the data?
- What do the key reports look like right now?
The other tools answer
- Is the configuration correct? — the GA4 audit
- Did something break just now? — the anomaly detector
- Is data being discarded? — the dimension checker
- Do the tags actually fire? — the tag debugger