Universal Analytics is gone. Not deprecated — gone. On July 1, 2023 it stopped collecting new data. On July 1, 2024, Google deleted all historical UA data from the interface. If you didn't migrate or export before that date, your old data is unrecoverable.
That's the headline. The longer story is that Google Analytics 4 isn't just a rebranded UA. It's a different product built on a different data model, with renamed metrics, a redesigned interface, and a learning curve most marketers underestimated. The reports you knew don't exist in the same place. Bounce rate works differently. Sessions count differently. Pageviews aren't called pageviews anymore — they're "views," tracked as one event type among dozens.
Google didn't force this switch arbitrarily. UA was built in 2012 around a session-and-pageview model that made sense for desktop websites and predictable user journeys. By 2020 most users moved between mobile apps, web, multiple devices, and privacy-restricted browsers — UA couldn't model that cleanly. GA4's event-based architecture handles cross-platform tracking, integrates machine learning for predictive metrics, and complies with stricter consent rules out of the box.
That's why this guide exists. If you've already migrated, you need to understand why your GA4 numbers don't match your UA numbers (different counting methods — see below). If you haven't migrated, you need to know what's salvageable and what to do next. The full picture starts with a clear-eyed look at what is google analytics 4 as a product separate from its predecessor — same name, very different tool. For the practical day-to-day, our how to use google analytics walkthrough covers the GA4 reporting interface from setup to first reports.
Short answer for the impatient: GA4 is more powerful, harder to learn, and your old data is probably gone. Keep reading for the eight differences that matter most.
Google announced the UA deprecation in March 2022, gave users 16 months to migrate, then stopped data collection on the original July 1, 2023 deadline. Most properties got auto-migrated to a basic GA4 setup by Google's automated tool — but "basic" meant tracking ID only. Custom events, conversion goals, audiences, and ad-account linking didn't carry over. If you didn't do the manual migration work, your GA4 property is missing the configuration that made your UA reports useful.
Google announces UA deprecation. Notifies all property owners via email and Analytics interface banner. 16-month migration window opens.
Google auto-creates GA4 properties for properties that haven't manually set them up. Basic config only — no event mapping or goal carryover.
UA standard properties stop collecting data. Historical data still viewable in interface. New hits sent to UA tracking IDs are discarded.
Historical UA data deleted from interface. UA 360 properties extended through July 2024 ended too. Data is gone from Google's servers unless exported.
GA4 is the only option. UA data unrecoverable unless previously exported to BigQuery (GA 360 only) or downloaded as CSV reports.
Google's official documentation lists dozens of changes. Most are cosmetic. These eight are the ones that change how you do your job.
UA tracked sessions and pageviews as the core unit. A user visited, viewed pages, those pageviews aggregated into a session, sessions rolled into users. Simple, predictable, somewhat limited.
GA4 tracks events. Everything is an event — a page load fires a page_view event, a scroll fires scroll, a file download fires file_download, a video play fires video_start. GA4 automatically collects around 90 events without any configuration. Custom events you define get the same treatment. This is more flexible but completely changes how you query data.
The metric you trusted in UA either disappeared or means something different in GA4. Bounce rate in UA = sessions with a single pageview. GA4 originally killed bounce rate, replaced it with engagement rate (sessions lasting 10+ seconds, OR with a conversion event, OR with 2+ pageviews). After backlash, Google added bounce rate back as the inverse of engagement rate — but the calculation is different from UA's, so the numbers don't match.
Sessions count differently too. UA started a new session whenever a campaign source changed mid-visit (utm_source parameter rewrite). GA4 doesn't do that. One visit, one session. Users went from a 30-day window in UA to 28-day active users in GA4. Pageviews became simply "views" — the same metric, renamed and tracked as a page_view event count.
UA's sidebar had Audience, Acquisition, Behavior, Conversions, Real-Time — roughly 50 prebuilt reports across those sections. GA4 has User, Acquisition, Engagement, Monetization, Retention — about 20-25 prebuilt reports. The rest moved to the Explore module, where you build custom analyses using Free-Form, Funnel, Path, Segment Overlap, and Cohort templates. Explore is more powerful than UA's custom reports but takes longer to learn.
UA used tracking IDs formatted UA-XXXXXXX-X. GA4 uses Measurement IDs formatted G-XXXXXXXX. If your site still has only a UA tag in the header, it's collecting nothing. Check your tag manager or site source. For most marketers, the smoothest path involves working with google analytics consultant or a contractor who's done this migration before — the configuration translation is mechanical but easy to mess up.
UA was web-only. If you had a mobile app, you used Firebase. GA4 unifies web and app tracking under one property — multiple data streams (web, iOS, Android) flow into the same reports. For businesses with apps, this is a major upgrade. For web-only businesses, it doesn't matter.
In UA you defined goals (destination URL, duration, pages per session, event). GA4 doesn't have goals. You mark an event as a "conversion" in admin settings. Up to 30 events can be marked as conversions per property. The migration tool didn't carry these over — you had to re-create them manually in GA4.
UA audiences were rule-based — visited page X, spent more than Y seconds, came from source Z. GA4 audiences add machine learning predictions: "likely to purchase in the next 7 days," "likely to churn," "predicted revenue." These predictive audiences require enough conversion data to train the model (Google requires 1,000+ returning users with conversion events in the previous 28 days).
This is the change nobody saw coming. In UA, BigQuery export was a GA360 enterprise-only feature costing $150K+/year. In GA4, BigQuery export is free for all properties — up to 1 million events per day on the free tier, with a 1GB daily streaming limit. If you've avoided GA's interface limitations, BigQuery gives you raw event data you can query with SQL. Pair this with your existing google analytics for seo tracking and you can do attribution analysis UA couldn't.
If you've been hunting for reports that used to be right there in UA's sidebar, here's the rough translation. UA's Audience Overview = GA4's User → User Acquisition. UA's Behavior Flow visualization is gone entirely — closest replacement is GA4's Explore module with the Funnel exploration or Path exploration templates, both of which need manual setup. UA Real-Time is still called Real-Time. UA's E-commerce reports moved to Monetization. UA's Site Search became a custom event you have to configure yourself.
UA: Session-based. Hits roll up to sessions, sessions to users. Predictable but limited to web-style page journeys.
GA4: Event-based. Every interaction is an event with parameters. Page view, scroll, click, video play — all events with different parameters. 90+ events auto-collected without setup. Custom events take parameters up to 25 per event.
Impact: SQL-querying your data in BigQuery looks completely different. Existing dashboards and tools built on UA's session model won't work without rewrites.
Bounce Rate: UA = single-page sessions. GA4 = inverse of engagement rate (different calculation, different numbers).
Sessions: UA = 30 min inactivity timeout. GA4 = same default but session_start event is logged differently. Numbers don't match.
Users: UA = 30-day lookback for unique users. GA4 = 28-day active users by default.
Pageviews: UA called it Pageviews. GA4 calls it Views (and it's just a page_view event count).
Conversions: UA = Goals. GA4 = Conversions (marked events).
UA: ~50 prebuilt reports (Audience, Acquisition, Behavior, Conversions, Real-Time). Custom Reports module for ad-hoc analysis.
GA4: ~20-25 prebuilt reports under Reports section (User, Acquisition, Engagement, Monetization, Retention). Most ad-hoc analysis moved to Explore module with Free-Form, Funnel, Path, Segment Overlap, and Cohort exploration templates.
Impact: Learning curve. Reports you used daily in UA may take 15 minutes to recreate in GA4 Explore the first time.
Auto-migrated: Tracking pixel installed, basic data collection started. Done by Google's setup assistant for most properties in late 2022.
Manual required: Conversion events, custom dimensions/metrics, audiences, Google Ads links, BigQuery export setup, e-commerce events, custom event mappings.
What didn't carry over: Historical UA data (still viewable until July 2024, then deleted), Goals configuration, View filters (replaced with data filters), Custom Reports (must recreate in Explore).
Migration always means trade-offs. Here's the honest accounting of where GA4 is worse than UA, and where it's substantially better.
Historical data, primarily. Unless you exported to BigQuery (GA 360 only) or downloaded CSV reports before July 2024, your UA data is gone. Year-over-year comparisons reset to zero in GA4.
Behavior Flow visualization is gone. The squiggly path diagram showing where users moved between pages? Doesn't exist in GA4. The closest equivalent is Path exploration, which is more cumbersome to set up and harder to read.
View filters — gone. UA let you create multiple Views of the same property with different filters applied (internal traffic excluded, specific countries only, test versus production). GA4 has Data Filters at the property level only, with limited options (Internal Traffic, Developer Traffic, no IP-based exclusion the way UA had).
Custom Reports as you knew them — gone. Build them in Explore now. Functional but different. Most analysts need a few hours to rebuild reports they used daily in UA.
Real-Time depth — reduced. UA's Real-Time reports showed users, pages, events, conversions, and locations across multiple tabs. GA4's Real-Time is a single dashboard with less drill-down.
Event-level data. Every interaction is captured as an event with parameters. UA forced you to track everything through custom events with limited parameters (event category, action, label, value). GA4 lets you define up to 25 parameters per event. For complex analytics, this is transformative.
Free BigQuery export. The biggest gift Google gave non-enterprise users. You can pipe every event into BigQuery, query with SQL, build dashboards in Looker Studio or any BI tool. UA charged $150K+/year for this through GA 360.
Predictive analytics. Purchase probability, churn probability, predicted revenue — Google's ML estimates these for individual users based on conversion data. UA had nothing comparable.
App + web unification. If you have a mobile app, GA4 tracks it alongside your website in one property. Cross-platform user journeys are visible without the gymnastics UA required.
Cleaner privacy story. Consent Mode v2 is built in. Cookie-less measurement using modeled data fills gaps from users who decline consent. UA was a constant scramble with cookie banners and consent plugins.
Better integration with the broader Google stack. Google Ads, Google Tag Manager, Search Console, BigQuery, Looker Studio — everything connects more cleanly in GA4. For SEO work specifically, our google analytics website traffic guide covers the reports SEO teams use most. To go deeper, google analytics course reviews catch you up on the GA4-specific training that didn't exist in UA's curriculum.
If you're reading this and haven't migrated, the situation depends on what's on your site right now. Most properties got auto-migrated by Google's setup assistant in late 2022 — your GA4 property exists with basic tag tracking. The work that remains is configuration: conversions, audiences, custom dimensions, e-commerce events.
Log into analytics.google.com. Look at the property selector. You should see two properties: your old UA property (data viewable until July 2024, empty after) and a new GA4 property. If you only see UA, Google's auto-migration didn't run — you need to create a GA4 property manually via Admin → Create Property.
Open Realtime → Overview. Visit your site in another tab. You should see your visit appear within 10 seconds. If nothing shows up, the GA4 tag isn't installed. Check Tag Manager or your site's <head> for the G-XXXXXX measurement ID.
Admin → Events → mark important events as conversions. Common ones: purchase (auto-tracked for e-commerce), generate_lead, sign_up, contact. You can have up to 30 conversion events.
GA4 auto-tracks scrolls, file downloads, outbound clicks, video engagement, site search — but it might miss things specific to your site. Custom events go through Google Tag Manager or direct gtag.js calls.
Admin → Product Links → Google Ads → connect your account. Conversions sync automatically. Audiences become available for remarketing campaigns. Get the right google analytics account permissions before this step — you need Editor or Administrator role on the GA4 property and the Ads account.
Admin → BigQuery Linking → link a GCP project. Free up to 1M events/day. This single step is the biggest reason to migrate — you get raw data access UA never gave non-enterprise users. To use the data for SEO analysis, including not-provided keyword workarounds, our guide on google analytics not provided keywords shows the BigQuery queries that surface what Search Console hides.
Real talk: your UA data is gone. Google deleted it. No amount of escalating with Google support will recover it. The only path forward is to set up GA4 fresh and start building new data.
The good news — that 12-18 months of fresh GA4 data will eventually replace what you lost, with better granularity than UA ever provided. The bad news — your year-over-year analysis is broken until at least mid-2025.
What to check before assuming everything is lost: third-party tools you used alongside UA may have your traffic data archived independently. Marketing dashboards built on BI tools often have UA data snapshots in their data warehouses. Old screenshot reports, monthly emails sent to stakeholders, and shared Google Slides decks may have traffic numbers you can rebuild metrics from. None of this replaces clean data, but it can fill historical gaps.
This trips up almost everyone. Your GA4 sessions show 50% fewer than UA reported for the same period. Your conversion count is half. Bounce rate looks dramatically different. Stakeholders accuse the new data of being wrong.
It isn't wrong. The math is different.
UA counted sessions when a user arrived AND the campaign source changed mid-visit (utm_source parameter rewrite started a new session). GA4 doesn't do that — one visit equals one session regardless of mid-visit URL parameter changes. This alone makes GA4 sessions appear lower than UA sessions for sites running paid ads.
UA's bounce rate was strict — single pageview, period. GA4's bounce rate inverts engagement rate. A user who scrolls past 90% of a page and stays 30 seconds but views only one page counts as engaged in GA4 (not a bounce). In UA, that was a bounce. Same user, opposite metric value.
Users count differently too. UA used a 30-day window for unique users. GA4 uses 28-day active users. Tiny difference, but compounded over months it adds up.
Conversions — GA4 counts every conversion event, not just the first per session. If a user submits a form twice in one session, GA4 logs 2 conversions; UA logged 1. For lead-gen sites this inflates conversion counts dramatically.
Bottom line: stop comparing GA4 numbers to UA numbers. The math is different by design. Use GA4 as your new baseline. Year-over-year comparisons start fresh from your GA4 launch date.
The Google Analytics Individual Qualification (GAIQ) certification was retired by Google in 2023 along with UA. The skills it tested still matter though — and many job postings still reference it. If you want to test your GA4 knowledge, run through our practice question banks. The format mirrors the original GAIQ exam structure but updated for GA4 metrics, reports, and data model. Try the google analytics iq exam for a full exam simulation, or the google analytics individual qualification exam questions bank for category-by-category practice. Both cover GA4-relevant material.
No. Universal Analytics stopped collecting new data on July 1, 2023. Historical UA data remained viewable in the interface until July 1, 2024, when Google permanently deleted it. There is no way to access UA data anymore unless you previously exported it to BigQuery (GA 360 customers only) or downloaded reports as CSV.
Most properties were auto-migrated by Google in late 2022. Log into analytics.google.com — you should see a GA4 property next to your old UA one. If you only see UA, create a GA4 property manually via Admin → Create Property. The basic data collection works, but you still need to manually recreate conversions, audiences, custom events, and integrations like Google Ads links.
The two products count differently by design. GA4 sessions don't restart when a campaign parameter changes mid-visit (UA did). GA4 bounce rate inverts engagement rate (different math than UA's single-pageview definition). GA4 counts every conversion event, not just first per session. The numbers aren't wrong — they're measured differently. Use GA4 as your new baseline rather than comparing directly.
The data model. UA tracked sessions and pageviews as the core unit. GA4 tracks events — everything is an event with parameters. Page views, scrolls, clicks, file downloads, video plays are all events. This makes GA4 more flexible for tracking complex user interactions but completely changes how you query and report on data.
Yes — for standard GA4 properties, BigQuery export is free up to 1 million events per day with a 1GB daily streaming limit. This was previously a GA 360 enterprise feature costing $150K+/year in UA. For most websites this is the single biggest upgrade in GA4 — raw event data accessible via SQL queries.
For analysts familiar with UA, plan 2-4 weeks to become comfortable with GA4's interface and event model. The Reports section is straightforward — Explore module is where most of the relearning happens. Most analysts can do basic GA4 work within a few hours but need a couple of weeks before they're as productive as they were in UA.
Yes. UA used tracking IDs formatted UA-XXXXX-X. GA4 uses Measurement IDs formatted G-XXXXXXXX. The two are incompatible. If your site still has only a UA tag, it's collecting nothing. Add the GA4 Measurement ID through Google Tag Manager or directly in your site's head tag. Both can coexist during migration.
Goals don't exist in GA4. You mark events as conversions instead — Admin → Events → toggle "Mark as conversion." You can have up to 30 conversion events per property. The migration tool didn't carry UA Goals over, so this is one of the most common manual steps people forget after auto-migration.
GA4 is free for standard use, same as UA was. GA 360 is the paid enterprise version with higher data limits, advanced features, and SLAs. For most websites the free tier handles 10M+ events per month comfortably. Free includes the BigQuery export that used to require GA 360 — a major change from UA's pricing model.
The original GAIQ certification was retired in 2023 with UA. Google now offers the Google Analytics certification through Skillshop, focused on GA4. The exam structure changed — fewer questions, GA4-specific content, free to take. Our google analytics certification guide covers what's on the current exam and how to prepare.