Google Analytics 4 implementation for websites and apps

Google Analytics records every interaction on a website or app as an event. We implement it so the numbers can be trusted: a clean tag setup, events and key events that match your goals, consent handled properly, and reporting that your team can use.

Talk through your analytics setup Request an estimate

Tagging with the Google tag or Google Tag Manager
Events and key events that match your goals
Consent Mode set up in the right order
Reporting through the Data API and BigQuery

When Google Analytics is the right choice

Google Analytics suits products that need event-level measurement:

You need to measure a website and apps together, with custom events and parameters.
You want to feed conversions and attribution into Google Ads. Events marked as key events can become Google Ads conversions.
You want raw event data in BigQuery for your own analysis. The daily export is unsampled and you own the exported data.
You want programmatic reporting through the Data API, which returns the same data as the interface.

When Google Analytics is not the right choice

Know these limits before you rely on it:

You need long-range exploration of user-level data. For standard properties, retention is limited to 2 or 14 months, so export to BigQuery if you need longer.
Your audience is small. Thresholding can hide rows in some reports, and you cannot change it.
Your volume is very high on a standard property. The BigQuery daily export has a cap there.
Server-side or offline data would be your only source. The Measurement Protocol supplements tag collection and may give partial reporting on its own.

What we do with Google Analytics

Typical work, described in general terms.

Tag setup

The Google tag or Google Tag Manager installed once, with a data layer that carries what your pages know.

Events and key events

Recommended events where they exist, custom events where they do not, and the right ones marked as key events.

Consent

Consent Mode defaults and updates in the right order, with the visitor's choice stored by your own code.

Cross-domain measurement

One measurement ID across your domains, with the linking settings and tests that keep sessions whole.

BigQuery and the Data API

Exports and reports so your own dashboards match the interface and your history lasts longer.

Audits and fixes

Finding duplicate tags, missing events, internal traffic in the data and consent that does not work.

What Google Analytics gives you

Google Analytics records interactions as events across web, Android and iOS. Google's documentation separates automatically collected events, enhanced measurement events that you switch on, recommended events that you implement with set names and parameters, and custom events for what no other type covers. Any event can be marked as a key event, which feeds attribution and Google Ads. Custom dimensions register your own event parameters for reports. The Measurement Protocol sends server-side or offline events, the Data API returns report data, and BigQuery export provides event-level data. Universal Analytics stopped processing data on 1 July 2023, and access to its data has ended.

What an implementation needs beyond the tag

Consent: Consent Mode defaults must be set before any configuration or event command, or they do not work. Updates must be sent before page transitions, and your code must store the visitor's choice, because Consent Mode does not.
Limits: there are documented limits on parameters per event and on the length of event and parameter names, and a key event whose name is too long is not reported as one. We design event names to fit.
Retention and thresholds: user-level data is kept for a limited time and some rows are hidden for small audiences. We tell you what the reports cannot show.
Cross-domain: use the same tag ID on every domain and list the domains. Redirects and some JavaScript navigation can strip the linking parameter.
Internal traffic: filters apply to web data only. An active exclusion is permanent and the data never reaches BigQuery, so we test the filter first.
Custom dimensions: a registered parameter appears in reports after a delay, and its scope and name cannot be changed after saving.

How we work

Agree what you want to measure: the decisions each number should support, before any tag is added
Name events once: a written event plan, so names and parameters stay consistent
Set consent before data flows: defaults, updates and storage tested in a browser
Hand over properly: the event plan, tag container and settings documented, with notes on how to test a change

Frequently asked questions

Can we recover old Universal Analytics data?

Google says Universal Analytics data is no longer available in the interface or API. Only data exported before the shutdown, for example into BigQuery, remains.

Will numbers match the interface if we build our own dashboard?

The Data API returns the same data as the interface and follows the property's reporting identity setting. Thresholding can still hide some rows.

How long can we keep user-level data?

For standard properties, 2 or 14 months. Longer periods apply only to Analytics 360 properties. Exporting to BigQuery is the route for older data.

Does a consent banner alone control tracking?

No. The site must pass the visitor's consent state to the Google tag through Consent Mode, and must store the choice itself. Whether your obligations are met is a legal question the documentation does not answer.