GA4 Setup
GA4, configured properly — not just installed.
Property structure, data streams, enhanced measurement, custom events, conversions, audiences, and the Google Ads + BigQuery links that unlock actual decision-making data. Most GA4 installs are pasted tags and nothing more — that's not analytics.
GDPR-compliant setup
Consent Mode v2 ready
Google Analytics 4 · Configuration
Collecting
Create GA4 property
Configure data streams
Enhanced measurement
Custom events
Mark conversions
Build audiences
Link Google Ads + BQ
Create GA4 property
Property name
trailkit-production
Time zone
(GMT+01:00) Dublin
Currency
EUR (€)
Why "GA4 is installed" isn't the finish line
Pasting the tag gets you 10% of the value.
A typical unmanaged install
✕ Gtag copied from Google, no customisation
✕ Default channel groupings — "Unassigned" everywhere
✕ No custom events — can't answer "how many people X?"
✕ No conversions defined — no goal tracking
✕ Internal traffic not filtered — your team inflates numbers
✕ Referral spam not excluded
✕ No Ads link, no BigQuery, no Search Console
✕ Reports unusable, trust erodes, team stops opening it
A properly configured property
✓ Custom data layer + event schema documented
✓ Enhanced measurement tuned, not default-on
✓ 15–30 custom events matched to real user actions
✓ Conversions marked + attribution model chosen
✓ Internal and dev traffic filtered
✓ Spam and bot filters applied
✓ Ads + BigQuery + GSC + Looker Studio linked
✓ Reports trusted, team acts on data
What GA4 changes from UA
Six things that tripped most teams up on the migration.
Universal Analytics ended July 2024. If your setup was built for the UA mindset and just ported, you're likely reading broken data. Here's where it goes wrong.
Sessions aren't the primary unit anymore
GA4 is event-based. Sessions are a derived concept. Many UA reports that 'worked' don't exist in GA4 — new reports think in users and events first.
Bounce rate is redefined
GA4's engagement rate is NOT the inverse of UA's bounce rate. Direct comparisons will mislead — we teach teams to read engagement rate as its own metric.
Attribution defaults shifted
GA4 defaults to data-driven attribution where UA used last-click. Same campaign can look very different. Knowing which model is active matters.
Data thresholding hides small numbers
GA4 suppresses low-volume rows to protect privacy. Can look like data is missing. Turning on Google Signals controls this trade-off.
Custom dimensions are now event-scoped
UA had session/user/hit-scoped dimensions. GA4 mostly event-scoped. Old dashboards need rethinking, not just re-mapping.
Reports are configurable — but not by default
Explore hub has power, but standard reports are thin. You need to build the reports your team needs, not rely on defaults.
Irish/EU compliance built in
Consent Mode v2 — not optional.
Google requires Consent Mode v2 signals for advertisers in EEA, UK, and Switzerland as of 2024. Without it, Google Ads audiences lose match rates and conversion modelling stops working.
We set up Consent Mode with your CMP (OneTrust, Cookiebot, Axeptio, or similar), validate signals are firing, and implement server-side via Google Tag Manager for the signals that matter most.
Consent Mode v2 signals
ad_storage, ad_user_data, ad_personalization, analytics_storage — all configured and validated.
CMP integration
Works with major EU CMPs. We wire the consent flow into GTM's consent state properly.
Conversion modelling
Consent-denied users' conversions still partially attributed via Google's modelling — works only if v2 is properly wired.
Server-side tagging (optional)
For clients wanting more control over data sent to Google, we deploy sGTM via Google Cloud Run.
The reports that actually get used
Reports we build as part of a setup engagement.
Acquisition deep-dive
Traffic by source, medium, campaign with engagement + conversion rate. The 'where should we spend more' report.
Product / landing page performance
Sessions per page, engagement rate, conversion rate, revenue. The 'which pages work' report.
Funnel exploration
Homepage → Product → Cart → Checkout → Purchase drop-offs. The 'where are we losing people' report.
User cohort retention
How users first seen in month X return in subsequent months. The 'does our site create habit' report.
Campaign ROI view
Ad spend (from Ads link) ÷ GA4 revenue, segmented by campaign. The 'what actually earns' report.
Executive KPI one-pager
Monthly dashboard in Looker Studio. 6–8 metrics, YoY + MoM, narrative annotations for anomalies.
Frequently asked
GA4 setup questions.
Typical timeline: 2–4 weeks. Week 1 is discovery (what decisions do you need to make?) + planning the event schema. Week 2 is implementation via GTM. Week 3 is QA, validation, and report building. Week 4 is handover and training. Shorter if your site is simple; longer if you have multi-property or app tracking needs.
Related services
Works together with
Get a free GA4 audit.
Give us read-only access and we'll return a 20-point scorecard and the top 5 fixes that'll make your GA4 data actually trustworthy. Free, no commitment.