Google has rolled out a significant update to Performance Max campaigns: advertisers can now run structured asset experiments directly within their campaigns. This update allows running a proper experiment that will help you figure out which creative assets work best for your campaigns rather than trying to guess which headline, image, or video makes the difference.
If you use Google Ads and run Performance Max campaigns – the type that makes up for almost half of all Google Ads conversions by 2026 – the update should be interesting to you.
What Is Performance Max and Why Assets Matter So Much
Let me quickly refresh your memory. As you know, Performance Max is the all-encompassing type of campaigns offered by Google. You feed assets such as headlines, descriptions, images, videos, and logos into the system, set budgets and bidding preferences, and the AI takes care of placing your ads across Search, Shopping, YouTube, Display, Gmail, Maps, and Discover.
Since ads are assembled from your assets automatically, the quality and diversity of your assets matter significantly. An asset group in Performance Max campaigns may consist of 15 headlines, 5 descriptions, 20 images, and 5 videos, and the algorithms decide which combination of these assets to showcase on which ad placement.
The long-lasting problem with Performance Max is lack of clarity about which creative strategies work well for the platform. In order to figure it out, you need to run a proper A/B test, but it has not been possible until now. This is what Google’s new update aims to solve.
What You Will Be Able to Test
According to the company’s news post, with the new update advertisers will finally be able to launch asset experiments within the framework of Performance Max campaigns. Let’s go over all the details.
First of all, you will be able to test how different creative assets perform in relation to each other. This can mean anything from a simple test of whether a particular set of assets performs better compared to another one to a more advanced analysis that allows you to test, for instance, what effect adding new assets to your campaign produces or how holiday assets work compared to evergreen ones. Assets that were created using Google’s Asset Studio are also testable.
Another addition worth noting is the possibility of introducing a second success metric for each experiment. Previously, experiments in Google Ads were based solely on one metric (such as CTR or CVR). You will soon be able to introduce a secondary metric that will give you a broader perspective on how your experiment affects your campaign’s performance.
Lastly, there is a unified experiments management tab on Google Ads dashboard called simply ‘Experiments’. All the tools for measuring conversion lift and conducting experiments are now located in this tab, allowing you to streamline your workflow and compare the results of different experiments easily. The company promises to bring more conversion lift and experiment capabilities to the tab soon.
Support for Google Ads API and Manager Account Coming Soon
Soon, you will be able to manage experiments for multiple accounts via Google Ads API and Manager Account (MCC). The news is important because it allows agencies to automate experiment management.
Why This Update Is Important
While Performance Max has its advantages and is one of the most effective advertising formats by 2026, it has faced criticism due to lack of transparency. Most of the decisions are taken by Google’s algorithms, and users have little insight into what is happening.
While this update does not increase transparency of the campaign process per se, it gives advertisers one more tool to understand how their strategies work. Namely, instead of speculating what will happen after switching to a different visual design, you can run a controlled experiment and figure it out objectively.
If you invest significantly in your creative assets, including taking photos, shooting videos, or creating seasonal campaigns, then this update is particularly important. For ecommerce businesses, being able to experiment with seasonal assets is particularly valuable.
Getting Started
Asset experiment functionality is available in all Performance Max campaigns. You may set up an experiment by using the new ‘Experiments’ tab in Google Ads and following the instructions on-screen.
To conduct an experiment, you first have to define your goal – for instance, you want to test how well a new set of images will work compared to your current asset group. Then, select your primary metric (soon you will be able to choose an additional metric). And wait until the experiment finishes.
As we have mentioned previously, since Performance Max campaigns use a single-campaign structure, your test will take place in one campaign and will leverage its learning data.
The company suggests leaving other settings, such as Final URL expansion, unchanged during the experiment; you should not remove or add any assets once the experiment has started.
Upon completion of the experiment, you will be able to apply the results automatically or leave your campaign untouched.
The Bigger Picture
It looks like this update fits a trend. Over the last year, Google has been working on bringing transparency to Performance Max and giving advertisers better insights into how it works. The additions include campaign-level negative keywords, channel reporting, asset level performance data, and now structured experiments.
While the campaign will continue to work differently from Google Ads campaigns, this update makes it a lot easier to make sure that what is happening is good for your business.
Frequently Asked Questions
Q 1: What are asset experiments in Performance Max?
They are tests that allow you to run comparisons between different groups of assets in a Performance Max campaign. Instead of swapping assets in a campaign and monitoring the effect on live performance, you will be able to structure this process and analyze outcomes.
Q 2: What types of experiments can I run?
Depending on your needs, you can run full tests to pit completely different asset groups against each other, partial tests to see how new assets affect the performance of an existing campaign, or tests between seasonal and evergreen asset sets. In addition, any assets created with Google’s AI-based Asset Studio are eligible to participate in these tests.
Q 3: What is the second success metric feature?
Right now, experiments are measured by one metric. As part of a future update, Google is adding a possibility of measuring a second success metric simultaneously with the first one to show you how this creative change affects the overall performance across multiple dimensions.
Q 4: Where do I access experiments in Google Ads?
All experiments (as well as conversion lift studies) have been moved to a dedicated page in Google Ads – there are no separate locations anymore.
Q 5: When will Google Ads API & MCC Support Be Available?
Google guarantees the availability of API & MCC support for this feature to start rolling out within a couple of weeks of the launch. For agencies working with multiple advertiser accounts, this is a critical functionality.
Q 6: Will running this experiment interfere with the learning stage of a campaign?
As of now, asset experiments will run within the framework of a single campaign, meaning no division into two distinct campaigns. Google advises against changing some settings and modifying existing assets while the experiment runs to prevent negative interference.
Q 7: Whom does the update help the most?
This new update is especially beneficial to brands that put effort into creating high-quality creative assets. Seasonal promotions, development of new brands and ideas, video campaigns will get a proper testing framework.
Q 8: Is it possible to automatically apply assets that performed better in an experiment?
Yes, although in most cases, you will have the choice to use the winning assets, end the experiment without applying anything to your campaign or have the automatic application enabled for this test.
Q 9: Will this feature be accessible to all advertisers?
The update is currently being rolled out across all Performance Max advertisers, meaning most advertisers will be able to use it in about a week from now.
Q 10: How is this different from testing new assets manually?
Every addition to your assets pool becomes a part of the whole campaign and gets lost in the overall statistics. That makes it difficult to determine whether the improvement in performance came from creative or some external factors.
