Don’t sleep on the change: Migrating from GA3 to GA4

Key takeaways In March 2023 we wrote this article as a countdown to the Universal Analytics shutdown. That deadline has passed, and so has the deadline to export UA data. What remains is the question many companies still face: was our GA4 set up properly? This update explains what changed and how to check. What […]

Key takeaways

  • Universal Analytics (GA3) is gone. It stopped collecting data on 1 July 2023, and from the week of 1 July 2024 its interface, API and all historical data were deleted.
  • Your UA data can’t be recovered unless you exported it before July 2024.
  • Many GA4 setups are still incomplete after rushed 2023 migrations: missing events, 2-month data retention, no Consent Mode v2, broken Looker Studio reports.
  • Use the 10-point checklist below to find and fix the gaps.

In March 2023 we wrote this article as a countdown to the Universal Analytics shutdown. That deadline has passed, and so has the deadline to export UA data. What remains is the question many companies still face: was our GA4 set up properly? This update explains what changed and how to check.

Let your customers find you, wherever they look.

What happened to Universal Analytics and its data?

DateWhat happened
1 July 2023Standard Universal Analytics properties stopped processing new data
1 July 2024UA 360 properties stopped processing data
Week of 1 July 2024UA interface and API shut down; all UA data deleted (also for read-only users)

Source: Google Analytics Help. UA data was never migrated to GA4. If you exported reports (CSV, Google Sheets, or BigQuery for 360 properties) before July 2024, those files are all that’s left. Tools that read UA through the API, such as Looker Studio, no longer show any UA data.

What’s the difference between GA3 and GA4?

AreaUniversal Analytics (GA3)Google Analytics 4
Data modelHits (pageviews, events, transactions) grouped into sessionsEverything is an event with parameters; sessions are calculated from events
EventsCategory, action, labelEvent name + parameters (recommended and custom events)
Goals / conversionsGoalsKey events (renamed from conversions in March 2024)
Bounce rateSingle-page sessionsShare of sessions that weren’t engaged (the inverse of engagement rate)
ViewsMultiple views per propertyNo views; data streams and filters instead
Data retentionUnlimited (aggregated reports)2 or 14 months for user-level data in explorations (standard)
BigQuery export360 onlyFree for all properties (1M events/day daily export limit for standard)
PrivacyCookies-based, IP storedNo IP storage, Consent Mode with modelling

How do you check if your GA4 migration was done right?

These are the gaps we find most often in analytics audits of properties migrated in 2023:

  1. Data retention. The default is 2 months, which limits explorations. Set it to 14 months (Admin → Data collection and modification → Data retention).
  2. Key events. Check that all former goals exist as key events (purchase, generate_lead, sign_up) and use recommended event names where possible.
  3. E-commerce tracking. Compare GA4 purchases and revenue with your store or CRM. A gap above 5–10% usually means missing purchase events, duplicates or consent issues.
  4. Event mapping. Old category/action/label events should be translated into meaningful event names with parameters, not copied one-to-one as generic events.
  5. Custom dimensions. Parameters you report on (e.g. author, product type, logged-in status) must be registered as custom dimensions, or they won’t appear in reports.
  6. Consent Mode v2. Required in the EEA since March 2024 for Google Ads personalization and remarketing. Check that ad_user_data and ad_personalization are sent. See Consent Mode v2.
  7. Cross-domain and referral exclusions. Payment gateways (e.g. PayU, Przelewy24, PayPal) should be listed as unwanted referrals so they don’t take credit for purchases.
  8. Channel grouping. Check UTM consistency, and use the default AI Assistant channel (added in May 2026) to see traffic from ChatGPT, Gemini or Perplexity.
  9. Reporting setup. Looker Studio reports on the GA4 connector are subject to GA4 Data API quotas (tokens per hour and day, 10 concurrent requests for standard properties). Large reports work better on BigQuery.
  10. BigQuery export. Turn it on even if you don’t use it yet: GA4 doesn’t backfill, so the export only contains data from the day you link it. See setup and costs.

Make your brand visible – from Google to AI chatbots.

Should you still compare GA4 data with old UA numbers?

With care. Users, sessions and bounce rate are calculated differently, so a drop or rise after migration doesn’t necessarily mean real change. Compare trends, not absolute values, and use your store or CRM data as the reference for revenue and orders. For attribution differences, see GA4 attribution models.

FAQ

Can I still access my Universal Analytics data?

No. Google deleted all UA properties and data from the week of 1 July 2024. Only data you exported before then (CSV, Sheets or BigQuery) is still available.

Can I import UA data into GA4?

No. GA4 doesn’t import historical UA data. You can combine exported UA data with GA4 data in BigQuery or a BI tool, but metrics aren’t directly comparable.

What replaced goals in GA4?

Key events. In March 2024 Google renamed GA4 conversions to key events; any event can be marked as a key event.

Why do my Looker Studio reports with GA4 show quota errors?

The GA4 connector uses the GA4 Data API, which limits tokens and concurrent requests per property. Simplify reports, use data extracts, or connect Looker Studio to the GA4 BigQuery export.

Sources

About the author: Krzysztof Surowiecki wrote this article as part of the Insightland team. Insightland is a Wrocław-based SEO and GEO agency, part of the Hexe Capital Tech Hub. Contact us.

more

Related blog posts

Monthly updates

Indexation error - Google Update

23 Nov 2020 • Marcin Gaworski

Voice Search Website optimization

How do you optimize your website for voice search?

31 Jul 2019 • Insightland

AI Search

AI search optimization: how to control what artificial intelligence says about your brand

23 Oct 2025 • Natalia Ząber