A Fine, Sunny Outlook Despite the Google Optimize Sunset

The Google Optimize sunset happens on September 30, 2023. Your experiments and personalizations can continue to run until then, but any experiments and personalizations still active on that date will end. Here’s an expert POV on what to do now.

Unless you’ve been hiding under a rock, you’ve probably heard the news about Google sunsetting their very popular (and free) experimentation platform, Google Optimize (as well as the paid version Google Optimize 360). Here are considerations for experimentation teams to keep in mind.

Google Optimize was an experimentation gateway platform

Many of us in the industry have very mixed feelings about Google Optimize. On the one hand, the free platform got companies who otherwise may not have gotten into experimentation in the first place obsessed with data-driven decision-making. Those companies were able, for the first time, to get a small taste of what’s possible with online experimentation. And now they’re hungry for more.

We can only hope they won’t give up just because the platform they relied on goes away.

Google Optimize sunset creates opportunity

In fact, we can hope that brands will actually use this sunsetting as an opportunity to level-up their game with a better tool or improved practice. Because you see, Google Optimize had some problems—problems many may not have been aware of.

And because we were so happy to see so many testing in the first place, some of us in the community may have been less vocal than we could have been (though others were never shy to voice their concerns in multiple posts).

Regardless of the benefits or drawbacks of the platform, it’s going away.

What comes next

The official Google announcement states that Google Optimize and the paid version, Google Optimize 360, will no longer be available after September 30, 2023. Google plans to focus on GA4 and integrations with optimization platforms rather than building a new experimentation platform.

Does this solve your problem? Can you just use Google Tag Manager to “split” your traffic, developers to manually create different content, and GA4 to analyze your test? Well, if you’ll recall the links posted above related to concerns about Google Optimize…those are specifically about the statistical validity of GA4. So, in short, no. Please don’t do that.

And Google’s solution of integrating with other platforms—please don’t do that either. Same problem there, unfortunately (the TL;DR involves sampling of data, which is a no-no with testing stats.)

It’s time to choose a new a/b test tool!

Where does that leave you? Well, you can pick a new tool. If you have sufficient traffic, conversions, and the budget to support a platform, please do that. There are wonderful platforms with solid statistical engines, easy-to-use interfaces, and all the capabilities you need—often with amazing support, all within your budget. Don’t assume they’re all as expensive as that one estimate you got from a leading vendor in the market.

Need Help Choosing a New Tool?

We've Got You Covered

Important considerations for your experimentation team

Got sufficient traffic/conversions?

What if you don’t have sufficient traffic? More critically, what if you don’t have enough conversions? Because success is not just about traffic. You can have all the traffic in the world, but if that traffic isn’t converting (on whatever metric you’re hoping to influence), you still won’t be able to easily/effectively measure how well your test is impacting it.

Remember back in middle school lab class when you got to play with the microscope? Think of your conversion rate as the object on that slide, and the “power” of your test as your ability to see the object on your slide. That “power”—the ability for you to be able to “see” if something is happening when you introduce a change agent such as a new feature or offer or product—is so much easier if the object under that microscope starts out large (i.e., a lot of traffic or high conversions).

Got sufficient time/sample size?

If you’re starting with a very tiny object (low traffic and low conversion), then you need to up the power of your microscope to be able “see” the change. You can do this by making a bigger intervention or by taking a much larger sample (in the case of online testing, that means a much longer runtime, which we often don’t have the luxury to do).

A/B testing is one tool among many

Does that mean we may not be able to make better decisions with data if we don’t have sufficient data? Of course not! A/B testing is just one tool in our amazing tool box. We can open our Structured Ideation toolbox and begin the process of continuous improvement cycles that are just as important.


In short, we should look to the sunsetting of Google Optimize and wish it a fond farewell, while also looking forward to a hopeful new tomorrow. What was started with Google Optimize does not end—it only gets better from here.

Reach out if you’d like help figuring out where to go from here. We’d love to help.


Google Optimize is an experimentation platform created by Google a little over 5 years ago that provides basic A/B, MVT, and redirect testing capabilities to test and personalize different variants of a web page to see how they perform against one another.

Google has announced their intentions to sunset Google Optimize and Google Optimize 360. Test and personalization campaigns will no longer be available after September 30, 2023. After that date, Google will only support experimentation and personalization via integrations with GA4. 

Google has stated in the same announcement that Optimize did not have the features and capabilities their customers required, and they have chosen to focus their efforts on improving GA4 and the integrations with existing market experimentation and personalization platforms that already have those features and capabilities the market is seeking.

The decision to continue using Optimize up to the sunset date is a business decision for each organization to make. At minimum, we recommend immediately making plans to download and store your historical data from Optimize before that data is lost, as your historical record for your experimentation program is an important part of building and maintaining a learning library and culture of experimentation. Once a go-forward plan has been aligned (new platform or plan to switch to Structured Ideation research toolbox), then you will be in a better position to decide if you want to continue using Optimize until the sunset date or to begin transitioning to your new “tool”.

Keep testing! Pick a new tool (vendor evaluation or if the traffic and conversions on your website don’t support – or the cost is prohibitive – consider a tool from the Structured Ideation toolbox).

Yes. Google has already announced 3 vendors (AB Tasty, Optimizely, and VWO in alphabetical order) they’re providing easy point & click integrations with, they have API integrations with many other vendors already, and they’re adding additional integrations between now and the sunset.

You will want to make a plan to download and store that data locally or in the cloud before the sunset. Your historical testing record is an important part of building and maintaining a learning library and culture of experimentation. 

Related Posts

Search Discovery
Education Community

Join Search Discovery’s new education community and keep up with the latest tools, technologies, and trends in analytics.




Scroll to Top


Catch the latest industry trends we’re watching and get new insights from our thought leaders delivered directly to your inbox each month.