The Practical AB Testing Guide for Google Ads Experiments

Improving your creative & targeting in Google Ads can take a lot of gradual trial & error. Fortunately there are tools like Google Ads Experiments that allow you to A/B test different ad creative, audiences, landing pages and more. 

A/B testing & Google Ad experiments have been around for a while. If you’re not using these tactics, then you could be missing out on key insights to improve your results.

This article will review what A/B testing & Google Ads experiments are, how to run A/B tests in Google Ads, and best practices to get great results from your A/B tests.

What are A/B Tests

A lot of you likely know this, in which case, feel free to skim to the next section. For those of you that want a better understanding of A/B testing, this could be very useful.

A/B testing in marketing is the practice of testing one control vs one variation to see which one improves vs your goal metrics. Another word for this is split testing. For example, you can show one part of your audience landing page A vs the second part of your audience will see landing page B.

A lot of marketers like to set up tests for the sake of. Without determining what success looks like, you will be setting yourself up for failure.

A/B testing is valuable because you can see what resonates with your potential customers more. This allows you to keep up to date with their ever changing preferences while improving metrics such as your CTR, CPCs, and Conv. Rate.

The image above illustrates the potential value of A/B testing. Hypothetically speaking, you would be able to generate 3 additional conversions with a test landing page vs the original. That could be a huge increase in ecommerce sales or leads in Google Ads without increasing your budget.

To run A/B tests in Google, it is best to use Google Ads Experiments.

What are Google Ads Experiments

Google Ad Experiments are a helpful way to continuously test different campaign settings to improve your performance. There are four different types of experiments that you should consider using to run A/B tests.

Ad Variations – this is the main Google Ads experiment type. You can test different components of responsive search ads including your headlines, descriptions, URL paths, & Final URL (landing page). This allows you to test very specific components of your ads while also controlling how much traffic you want to use for your test variation.

Custom Experiments – these Google Ad experiments run on search or display campaigns. You can select one or multiple of your current campaigns to then test different components of your campaigns. A classic test marketers like to run in Google Ads is CPC bidding vs automated bidding. You can see which type of bidding gets better results if you’re concerned about switching to a different bid strategy. There are several other components of the campaign like the keywords, audiences, settings, and so forth. Then schedule the experiment.

Video Experiments – allow you to determine which of your video ads are better performing on YouTube. In Google Experiments, you can set up A/B tests in 2 to 4 different ad groups, choose the campaigns to include in your experiment, and select a particular success metric. This is great for anyone that has a significant brand awareness budget or gets sales/leads from YouTube.

App Asset Experiments – this Google Ads experiment type is great for measuring the performance of adding assets to your app campaigns. App advertisers can test different headline, description, image, and HTML5 assets to see what performs best. There are more details in this article. This Google Ads experiment type is great for app advertisers and is equivalent to ad variation testing in this space.

What’s great about all of these experiments is that you control the percent of traffic used, which A/B test variations are being tested, you get to see all the results compared between the control vs variation, and…’s free to run an experiment.

How to Run a Google Ads Experiment

This can seem tricky at first. Following these basic steps can help you run each experiment type. We also added a few extra steps for your benefit.

  1. Determine your testing goals – figure out what you want to achieve before you run a test in Google Ads experiments. Here are a few good examples to help you get started:
    1. Increase the CTR of your search ads by 10% in the next 90 days.
    2. Lower your CPCs by 5% with automated bidding in 30 days.
    3. Increase your conv. rate by A/B testing different landing pages over the next 180 days.
    4. Test different headlines to improve your lead quality.
  1. Figure out what to test first – consider where you are likely to get the best results with the least amount of effort. We like to create testing lists with our clients and brainstorm different ad creatives/audiences/campaign settings to test. Happy to help you with this process as well. For now, it is good to consider this to determine what to test first.
  1. Test setup in Google Ads Experiments:
    1. Log into Google Ads
    2. Go to the left hand side of your screen and click on Experiments
    3. Select the experiment you want to run
      1. Either ad variations for responsive ads
      2. Video experiments for video creative
      3. All Experiments/Custom experiments for campaigns & apps
    4. Follow the prompts in Google and you will be well on your way to completing your first experiment!

Some important features to consider when you set up your Google Ads experiment.

  • Make sure you select the right campaign or asset. Sounds easy, but it’s definitely worth double checking.
  • Determine your experiment split for your A/B test. All 4 experiment types have this in some form. It’s usually good to run a 50/50 test or 60/40. Anything more than that needs more time to run or might not be statistically significant.
  • Make sure you get the right start/end dates in place before running your campaign. Not getting enough data or conclusive data means you will need to run your experiment again. 

A/B Testing Best Practices in Google Ads Experiments

Here are best practices to consider for Google Ads Experiments in addition to what we’ve already written about:

  • Look at how many conversions/clicks/impressions your ad creative/asset/campaign generates. This will determine how long you need to run your experiment to get more results.
  • If you believe an experiment will most likely work based upon prior tests, you may want to give the variation more traffic to start with.
  • Remember that after the experiment, you can apply the winning variation to your campaign so it can get all of the traffic going forward. It wouldn’t be ideal to run an experiment, get better results, then not realize those results going forward just because changes didn’t apply.
  • If you have a smaller account, then running tests for statistical practicality might be best. You can always revisit tests later on if you see different results over time.
  • Running multiple tests for a similar concept can be helpful to prove its validity. Sometimes things like seasonality, other account changes, promotions, or smaller data sets can skew results. Running a similar test at different times can provide more conclusive results.

Here are some A/B test ideas that you could run as a new Google Ads Experiment:

  • Tests for CTR:
    • Change the headlines to mirror keyword intent or the literal keyword
    • Add CTAs vs benefit callouts
    • Compare different benefit/feature callouts
    • Test different ways of saying the same thing (ex. 20% off vs save 20%)
    • Test different qualifying text to lower the CTR but get more qualified traffic
    • Test different descriptions to see if you can get more qualified people or increase the CTR
  • Tests for CPCs
    • The tests for CTR indirectly impact CPC tests. This is good to keep in mind as you run your tests. More engagement can increase quality score which can help your CPCs or you can get more of your impressions to turn into clicks.
    • Test different match types
    • Test different bid strategies
    • Test different campaign settings (location, dayparting, device, etc)
  • Tests for CVR
    • Test different landing pages such as:
      • Home page vs PDP page
      • Home page vs service page
      • Original page vs contact us page
      • Original page vs DIY information
      • Product/Service category pages vs Individual Service/PDP pages
    • Test different headlines/descriptions to see if the quality improves or more people convert. You could test variations like different promotions, using job titles in the ad copy, using locations in the ad copy, adding reviews, adding different benefits, and so much more

Final Thoughts

This article should give you everything you need to get started with A/B testing using Google Ads experiments. If you need more help or additional insights on Google Ads experiments or A/B testing, we’re here for you! Book a free consultation with us today and we’ll go over how to get started on Google and how to get better results.

Centaur Consulting Group

Want a FREE Digital Marketing Consultation?

  • 30+ years of experience
  • Analysis, insight and strategic advice
  • Find Out How We CAN Help You
Book FREE Consult