Setting up a GA4 property and data streams

Part 1 covered why GA4 exists. This part is the actual setup — the real hierarchy every GA4 implementation is built from, and Bright Leaf Coffee's first property, created from scratch.

The real hierarchy: Account, Property, Data Stream

Google Analytics Account: Bright Leaf Coffee
  Property: Bright Leaf Coffee — Web & App
    Data Stream: brightleafcoffee.com (Web)
    Data Stream: Bright Leaf Coffee iOS App (if one existed)

An Account is the top-level container, typically one per business. A Property sits beneath it, representing a specific business unit or product being measured — most businesses need only one. A Data Stream is the actual source of data flowing into that property — a website, an iOS app, an Android app — and a single GA4 property can combine multiple streams into one unified view, a genuinely different structure from Universal Analytics, which required separate, disconnected properties for web and app.

Creating the account and property

Account name: Bright Leaf Coffee
Property name: Bright Leaf Coffee — Web & App
Reporting time zone: match the business's actual primary time zone
Currency: USD (matters directly for e-commerce revenue reporting, part 9)

Time zone and currency aren't cosmetic settings — every report's date boundaries and every revenue figure GA4 displays are calculated using these values, and getting either wrong at setup produces subtly incorrect data that's easy to miss until a report looks inexplicably off later.

Adding the web data stream

Stream type: Web
Website URL: https://brightleafcoffee.com
Stream name: Bright Leaf Coffee — Main Site

Creating this stream generates a unique Measurement ID (in the form G-XXXXXXXXXX) — the specific identifier every tracking snippet installed in part 4 needs to actually send data to this exact property. This ID is the direct connection between the website's code and this specific GA4 property; a mismatched or missing ID is one of the most common reasons a "correctly installed" tracking snippet reports no data at all.

Enhanced measurement: on by default, worth understanding

Enhanced measurement (enabled by default):
  ✓ Page views
  ✓ Scrolls
  ✓ Outbound clicks
  ✓ Site search
  ✓ Video engagement
  ✓ File downloads

GA4 automatically tracks these common interactions with zero custom code required, a genuine improvement over Universal Analytics' need for manual event tracking on most of these — worth knowing they're on by default rather than assuming every single event needs custom implementation, which part 3's event model covers in more depth.

A real second property structure for GreenDesk

Account: GreenDesk
  Property: GreenDesk — Web & App
    Data Stream: greendesk.io (Web)

Even though GreenDesk currently has no app, structuring the property as "Web & App" from the start — rather than naming it narrowly as "GreenDesk Website" — avoids a real, if minor, future headache: adding an app data stream later works cleanly regardless of the property's original name, but a property named too narrowly can create confusion once it's tracking more than what its name implies.

Data retention settings, worth reviewing immediately

Event data retention: 14 months (the maximum available, vs. the 2-month default)

This setting, easy to miss during initial setup, controls how long GA4 retains user-level and event-level data for certain exploration reports (part 12) — the default is genuinely short, and extending it to the maximum immediately at setup avoids losing the ability to look back further than a couple of months later, once that data has already been silently discarded.

Next: the GA4 data model in full — events and parameters, and what "everything is an event" actually means in practice.