Want to run A/B tests and actually see what’s working? You need Google Optimize wired up to Google Analytics. It’s not rocket science, but it is easy to mess up if you’re not careful. This guide is for marketers, analysts, or anyone who wants real data from experiments—not just “it felt better” guesses.
If you’re tired of vague results and want reporting you can trust, keep reading. We’ll walk through exactly how to connect Google Optimize with Google Analytics, what you can skip, and how to make sure it’s all working. No fluff, no hype.
Why bother connecting Optimize to Analytics?
Let’s get real: Google Optimize alone gives you the basics (like which variant won), but not much else. If you want to slice your test data by device, traffic source, or user segment, you need to send that data to Google Analytics.
By connecting the two, you'll be able to:
- See experiment data alongside your regular site reports
- Break down results by geography, device, or custom segments
- Avoid the “black box” feeling of trusting Optimize’s summary stats
If you care about understanding why a test worked (or didn’t), this is the way.
Before you start: What you actually need
Don’t get tripped up by missing prerequisites. You’ll need:
- A Google Analytics property set up (Universal Analytics or GA4)
- Edit permissions in both Optimize and Analytics
- The ability to change your website’s code (or a dev who can help)
- Google Optimize container already created
Pro tip: If you’re using Google Tag Manager, your life is a little easier, but you still need to double-check everything.
Step 1: Double-check your Google Analytics setup
Before you touch Optimize, make sure Analytics is working and collecting data. Go to your site → open Analytics in another tab → check Real-Time reports. If you see yourself as a visitor, you’re good.
Skip this step if: You already use Analytics daily and know it’s working.
Step 2: Create or open your Google Optimize container
- Head to Google Optimize and sign in.
- If you haven’t set up a container yet, create one (it’s basically a workspace for your experiments).
- Give it a name you’ll remember. (“Homepage tests” beats “Test123” every time.)
Pro tip: One container per website is the rule of thumb. Don’t overthink it.
Step 3: Link Google Optimize to your Google Analytics property
This is where most people mess up—so go slow.
- In Optimize, inside your container, click the experiment you want to connect.
- In the experiment setup, find the “Measurement and objectives” section.
- You’ll see an option to “Link to Google Analytics.” Click it.
- Pick the right Analytics property and view (for Universal Analytics) or data stream (for GA4).
- Save.
What can go wrong? - Picking the wrong Analytics property (happens all the time). - Linking to a test or dev view instead of your real production data.
Double-check: Open Analytics, go to Admin > Property > Tracking Info, and confirm it’s the same property you just linked.
Step 4: Add the Optimize snippet to your website
There are two main ways to do this: directly in your site’s code, or via Google Tag Manager (GTM). Here’s what matters:
If you’re manually editing your code:
- Paste your Analytics tracking code first.
- Immediately after, add the small Optimize snippet.
- Use the “anti-flicker” snippet if you care about hiding changes while the experiment loads (trust me, you should).
Example: html
Replace UA-XXXXXXX and OPT-XXXXX with your actual codes.
If you’re using Google Tag Manager:
- Add the GA tag as normal.
- Add the Optimize tag, link it to your container ID.
- Make sure your triggers fire on all pages you want the experiment to run on.
- Don’t forget the anti-flicker snippet—GTM can’t inject CSS before the page loads, so this needs to go in your site’s
<head>directly.
Heads up: GTM is great, but if your page loads slowly or scripts load out of order, you can get weird results (like visitors seeing a flicker or the wrong variant). Test on a real device, not just your desktop.
Step 5: Set up your experiment objectives in Optimize
Don’t just rely on “pageviews” or “sessions.” Get specific.
- Add objectives tied to actual business outcomes: form submissions, purchases, or engagement events.
- You can select goals already set up in Google Analytics, or set custom objectives.
- If you’re using GA4, make sure your “events” are actually showing up in Analytics before you use them as test objectives.
What’s worth tracking: - Micro-conversions (button clicks, email signups) - Revenue events (for e-commerce) - Bounce rate or time on page (if you must, but don’t obsess over these)
What to ignore: - Vanity metrics (“scroll depth” is mostly useless unless you know what to do with it) - Goals you never check
Step 6: Test, preview, and QA everything
Don’t launch blind. Save yourself some headaches:
- Use Optimize’s preview mode to check variants.
- Open Google Analytics Real-Time and see if your own test visits show up with the right experiment data.
- Try incognito/private browsing and mobile devices—sometimes tracking breaks in the real world.
- If you’re using ad blockers, know that lots of users are too—and they can block Optimize or Analytics. Always expect some data loss.
Look for: - The “experiment” dimension or “variant” shows up in Analytics reports - Your experiment objectives actually register conversions
Step 7: Where to find your experiment results in Google Analytics
In Universal Analytics:
- Go to Behavior > Experiments (not available in GA4)
- Or, use a custom report: add “Experiment Name” and “Variant” as secondary dimensions
In GA4:
- There’s no “Experiments” report—yeah, it’s annoying.
- Go to Reports > Engagement > Events, filter by your experiment/event names.
- Or, build an Exploration with experiment dimensions.
Pro tip: Don’t expect GA4 to make this easy. If you care about A/B test reporting, you might need to do some extra slicing and dicing in Explorations or export data to a spreadsheet.
What actually works (and what doesn’t)
What works: - Simple tests with clear objectives - Double-checking your tracking setup before running a “real” experiment - Keeping one source of truth—don’t compare Optimize’s built-in results to Analytics, just pick one - Using Google Tag Manager if it’s already part of your stack
What doesn’t: - Overcomplicating with multiple containers or properties - Setting dozens of goals and then ignoring them - Trusting Optimize’s stats blindly (always sanity-check in Analytics)
Stuff to ignore: - “Advanced integrations” unless you have a real need (and developer help) - Chasing every micro-metric—focus on what actually moves the needle
Keep it simple and iterate
Connecting Google Optimize and Google Analytics shouldn’t feel like open-heart surgery. Stick to the basics: clear objectives, clean tracking, and a quick test before you launch. Most “advanced” stuff just adds confusion and more ways for things to break.
If something’s not working, go back and check your links, container IDs, and code placement. Nine times out of ten, it’s a copy-paste mistake or a trigger firing at the wrong time.
Don’t chase perfection. Get the connection working, run a small test, and build from there. The best experiment is the one you actually ship—and can report on with confidence.