Skip to content
accs-net.com

Press Esc to close

Property

Glossary · 7 min read ·

A property is the container in Google Analytics where data for one user base lives — the level of the account hierarchy that collects events from your site or app, stores them, and holds the reports, key events, audiences and settings built on top. Google’s account-structure documentation defines a property as representing “data for one user base,” sitting inside an account and fed by one or more data streams. Almost every important GA4 configuration decision — retention, attribution, identity, integrations — happens at property level, which is why understanding the container matters more than it sounds. This entry covers the hierarchy, the property/data stream split, the difference between a property ID and a measurement ID, and how many properties a business actually needs.

Where a Property Sits in the Hierarchy

The GA4 hierarchy runs four levels deep:

Level What it represents Typical scale
Organization Your company across Google’s marketing products One
Account Ownership and terms of service — one legal entity One per company
Property One user base: its events, reports and settings One per brand or business unit
Data stream One source feeding the property: a website, an iOS app, an Android app Up to 3 per property in practice

An account can hold up to 2,000 properties; a property belongs to exactly one account. Google’s structural recommendation is short: one account per company, one property per brand or business unit, and within a property “a maximum of 3 data streams: 1 single web data stream to measure the web user journey and 1 app data stream each for iOS and Android.”

Property vs Data Stream

The pair is the most common point of confusion in GA4 setup. The property is the database and reporting layer; a data stream is a pipe into it. Events from all streams of a property land in one shared pool — a purchase from the iOS app and a purchase from the website end up in the same reports, distinguishable by the stream dimension but not separated by default. Two consequences follow:

  • Cross-platform measurement is the point. If your web and app users are the same people, one property with multiple streams (plus User-ID) is how GA4 stitches the journey. Splitting web and app into separate properties makes cross-platform tracking impossible after the fact.
  • Separate audiences deserve separate properties. If two sites serve genuinely different user bases, one property would blend their metrics into averages describing neither. The “one user base” phrase in Google’s definition is the test.

Property ID vs Measurement ID

Two identifiers get mixed up constantly:

  • The property ID is a plain number (like 328420000) identifying the property itself. You use it in API calls, BigQuery configuration and support conversations. Visible under Admin → Property details.
  • The measurement ID looks like G-XXXXXXX and identifies a specific web data stream, not the property. It is what goes into your tag so events reach the right pipe. Our measurement ID entry covers it in detail.

If a tool asks for an ID, the format tells you which one it wants: numeric → property, G- prefix → stream.

What Lives at the Property Level

Nearly all GA4 configuration is per property, which makes the property the unit you administer, audit and export:

  • Data retention — the 2-month or 14-month cap on event-level data for standard properties, set per property.
  • Attribution settings — the attribution model and lookback windows.
  • Reporting identity — how users are deduplicated (User-ID, device ID, modeling).
  • Key events, audiences, custom dimensions — all defined per property.
  • Product links — Google Ads, Search Console and BigQuery connections are per property; the BigQuery export in particular is configured property-by-property.
  • Access management — user permissions can be granted at account or property level.

GA4 vs Universal Analytics: No More Views

Universal Analytics had a third layer — views — under each property, and teams leaned on them for filtered slices: a test view, a “no internal traffic” view, a per-country view. GA4 properties have no views. Their jobs moved elsewhere: internal-traffic exclusion became a data filter at property level, ad-hoc slicing moved into report filters and Explorations, and the sandbox role is served by a separate test property. Teams migrating from UA most often stumble here, recreating per-view structures as a sprawl of properties instead of using filters within one.

Subproperties and Roll-Up Properties (360)

Google Analytics 360 — the paid enterprise tier — adds two structural tools. A subproperty is a filtered subset of a source property (for instance, one region’s traffic exposed to one team, for data governance). A roll-up property combines several source properties into one aggregate view for a portfolio or holding-company perspective. On the standard (free) tier, neither exists: the property you collect into is the property you report from, and combining sites into one picture means either sharing a property or joining exports downstream in BigQuery.

Day-One Property Settings That Are Hard to Fix Later

A handful of property-level choices are cheap at creation time and expensive afterwards:

  • Reporting time zone. It defines where “a day” begins and ends for every report. Changing it later applies going forward only — data already processed keeps its old day boundaries, so the switch day and trend comparisons around it get messy. Pick the zone your business actually operates in, once.
  • Reporting currency. Revenue arriving in other currencies is converted at the daily rate into the property currency. Change it later and historical revenue is not restated — you get a seam in every money chart.
  • Data retention. Standard properties choose 2 or 14 months of event-level data for Explorations. The default is 2 months, and it silently limits how far back ad-hoc analysis can reach — most properties want it moved to 14 immediately, because the extension is not retroactive.
  • Industry category and property name are freely editable and cosmetic — rename any time without data impact.

Standard-Tier Property Limits Worth Knowing

Limit Standard property
Properties per account 2,000
Event-scoped custom dimensions 50
Custom metrics 50
User-scoped custom dimensions 25
Audiences 100
Key events 30
Event-level data retention up to 14 months

Google Analytics 360 raises most of these substantially. On the free tier the limits are generous for one property doing one job — and another argument against sprawl, since every extra property carries its own copies of definitions to configure and keep consistent.

How Many Properties Should You Create?

Fewer than you think. The recurring failure mode is property sprawl — one per country, one per subdomain, one per team — which fragments users who actually belong to one base and multiplies configuration to keep in sync. The working rule from Google’s own guidance: one property per user base, one web stream inside it, app streams alongside when the apps serve the same users. Create an additional property for a genuinely separate audience, a sandbox for testing, and little else. If the motivation for a new property is “different people want different reports,” that is a filters-and-permissions problem, not a structural one.

Frequently Asked Questions

What is the difference between an account and a property?

The account is the ownership shell — one legal entity, terms of service, top-level access. The property is where measurement happens: events, reports and settings for one user base. A small company typically needs exactly one of each.

Can I merge or move properties?

A property can be moved to a different account (with the right permissions), but two properties cannot be merged — historical event data cannot be combined inside GA4 after the fact. The practical merge path is exporting both to BigQuery and joining there, which is why the structure deserves getting right early.

Should each subdomain get its own property?

Usually no. Subdomains sharing one audience belong in one property — one web stream covers them, since the GA4 tag reports the full hostname with every event and reports can filter by it. A separate property per subdomain only makes sense when the subdomain genuinely serves a different user base.

Can I change a property’s time zone or currency later?

You can, but both changes apply going forward only — historical data keeps the old day boundaries and the old currency conversion. Expect a visible seam in reports around the change date, and avoid comparing periods across it.