Have you ever wondered how different variables might work for your campaigns? In Google Ads, you can actually create experiments to do A/B Testing with all sorts of variables. You can test different landing pages, ad copies, bidding strategies, keywords, keyword match types, and other variables to see which one will work better. With this easy step-by-step guide, you will quickly learn how to create experiments in Google Ads.
Login to your Google Ads account
Navigate to Campaigns (Left-hand-side menu)
Open the drop-down menu and select Experiments
Click on the "+" symbol to create a new Experiment
Select the type of Experiment you'd like to create (Our suggestion is to create a custom experiment)
Select the Campaign type
Name the Experiment
Select the Base Campaign
Click "Save and continue"
Make changes to the Experiment campaign and click "Apply"
📝 Note: The system will create a duplicate campaign that mirrors the base campaign to which we can make any changes. Click on the name of the campaign to make the changes that you like, such as keywords, bidding strategy, landing pages, and others.
Click on "Schedule"
Determine the Experiment Goals
➡️ In this step, we will be determining what is our experiment goals, i.e. what are the metrics that we want to move. You can click on the dropdown on the left to select the metrics, and the dropdown on the right to select increase/decrease. In our example, we would like to see an increase in conversions and a decrease in cost per conversion.
Determine the Experiment Split
➡️ Experiment split determines how you want to split the traffic and budget. For instance, if you set the split by 50-50, it means that one time the system will serve your original campaign and the next time it will serve your trial campaign. You can adjust the percentage as you see fit.
Under advanced options, you can have two options:
Search-based: when somebody types in a keyword/phrase on Google, the first search will show the base campaign and the next search will show the trial campaign.
Cookie-based: every user will only see one version of the campaign. For example, if I type the same keywords multiple times on Google, I will only see only one version of the campaign.
Determine the Experiment Dates
Enable Sync
📝 Note: When the Enable sync is on, it means that any changes you make on the base campaign (the original campaign) will sync automatically to the trial campaign. If we keep it off, any changes we made on the base campaign will need to be applied manually to the trial campaign. Our suggestion is to keep this on.
Click on "Create Experiment"
How to analyze Google Ads Experiment results
Navigate to Experiments
Click on the Experiment you'd like to analyse. The system will show you the Experiment Summary, along with the results according to the goals you have set. In our example, we will be able to see the cost per conversion and conversion data from the experiment campaign.
📝 Note: If the experiment works/doesn't work according to your expectations, then you can cancel the experiment or apply the experiment (once it has gained results). This is an example from a different experiment that is currently in progress, where you can see the results and apply or cancel the experiment accordingly.
Frequently Asked Questions on Google Ads Experiments
What are Google Ads Experiments?
Google Ads Experiments (formerly known as Campaign Drafts & Experiments) allow advertisers to test changes to their campaigns in a controlled environment. You can test different bidding strategies, ad copies, audience targeting, or other campaign elements while measuring the impact on performance metrics.
Why should I run experiments?
Experiments help you make data-driven decisions about your advertising strategy by:
Testing campaign changes before implementing them fully
Measuring the impact of specific modifications in isolation
Comparing performance between different approaches
Reducing risk when trying new strategies
Identifying opportunities for campaign optimization
How do experiments work?
When you create an experiment, Google Ads splits your traffic between a control group (your original campaign) and a test group (your experiment). Both groups run simultaneously, allowing you to compare their performance directly. You can specify how much traffic to allocate to each group and how long the experiment should run.
What traffic split should I use?
The recommended traffic split is 50/50 between control and experiment groups. This provides the most statistically significant results in the shortest time. However, you can adjust the split based on your risk tolerance and testing needs. Use a smaller percentage (e.g., 20% experiment/80% control) if you're testing risky changes.
How long should experiments run?
Experiment duration depends on several factors:
Your campaign's regular traffic volume
Statistical significance requirements
Seasonal considerations
Performance cycle of your business
Generally, aim for at least 2-4 weeks to gather sufficient data. High-traffic campaigns may need less time, while low-traffic campaigns may need longer.
Intresting post