Google has introduced a new tool that makes it easier for advertisers to test their video ads. This new tool simplifies setting up experiments to measure how different creative elements impact campaign performance. It helps marketers make data-driven decisions and optimize their ads for the best results. Let’s dive into the key features and how to use this powerful tool.
What is the New Google Video Experiment Tool?
The new Google Video Experiment tool allows advertisers to test their video ads simply and efficiently. This tool allows you to quickly set up experiments to measure the performance of your ads across key metrics. It is designed to give clear insights into which creative elements work best for your campaign goals, like increasing conversions or lowering cost-per-view.
Why is Video Ad Testing Important?
Video ad testing helps marketers understand what resonates with their audience. By testing different versions of video ads, you can determine which ones generate more views, clicks, or conversions. It also allows you to optimize your campaigns by focusing on what works, ensuring you spend your advertising budget wisely.
Streamlined Experiment Setup
Google has made setting up video experiments easier than ever. Whether working with existing campaigns or starting fresh, the setup process is straightforward. Here’s a breakdown of how it works:
- Choose a control arm campaign from your existing video campaigns.
- Google automatically creates a duplicate of your campaign for the treatment arm.
- You can then modify your treatment campaign by adding or removing video assets.
This streamlined setup saves time and allows you to focus on analyzing the results.
Measure the Impact of Creative Elements
Once your experiment is set up, you can test how different creative variations affect important metrics such as:
- Cost-per-view (CPV)
- Cost-per-conversion
- Brand Lift
- Video view rate
This allows you to see how changes to your video ads, such as adding or removing assets, influence the success of your campaign.
Flexible Confidence Intervals
The tool offers flexible confidence intervals, meaning you can start seeing directional results within days. Over time, your results will gain statistical significance, giving you greater confidence in your findings. If you allow the experiment to run longer, you can reach up to 95% statistical significance.
How to Ensure Accurate Results
Google’s Video Experiment tool helps avoid setup errors, ensuring accurate results. This means that you can trust the data you gather from your experiment. To further improve accuracy, ensure your video assets and campaign settings are consistent across your control and treatment arms.
A/B Test Video Assets
For simpler experiments, Google offers a basic A/B testing option. In an A/B test, you compare two versions of your video ad (one in each campaign). This type of experiment is perfect if you only want to test one variable at a time, such as the creative asset.
How A/B Testing Works
- Select a success metric to test (such as Clickthrough Rate or Cost-per-Conversion).
- Choose a Video Reach or Video View campaign to serve as your control.
- Google will duplicate the campaign and allow you to modify the treatment arm.
- Run the experiment and make a data-driven decision based on the results.
Custom Experiments for More Advanced Testing
For advertisers who need more control, Google provides a custom option. Custom experiments allow you to test any variable (or multiple variables) and compare up to 10 experimental arms.
How Custom Experiments Work
- Select a success metric to test, such as Brand Lift or Conversion Rate.
- Choose one or more campaigns to serve as control and treatment arms.
- Adjust variables such as audiences, bidding strategies, or creative assets.
- Use the results to optimize your campaigns.
Setting Up a Hypothesis
Before you start an experiment, it’s important to define a clear hypothesis. This is a question or statement that your experiment will answer. For example, you might ask: “Which of these video ads drives a higher view-through rate: a 2-minute tutorial or a 15-second direct offer?”
Your hypothesis should be tied to your campaign goal. By having a clear hypothesis, you can focus on the key metric that matters most to your campaign.
Example: Car Manufacturer Video Experiment
A car manufacturer wants to test which video ad leads to more conversions for their new SUV. They set up an experiment with two video ads and chose “Conversions” as the success metric. After running the experiment, they found that one ad led to 8 times more conversions than the other.
Steps to Create a Video Experiment
Setting up a video experiment is simple. Here’s how you can do it:
- Go to your Google Ads account and click on the Campaigns icon.
- Select Experiments from the dropdown menu.
- Choose Video Experiments.
- If you’re running an A/B test, select A/B Test Video Assets. For more complex experiments, select Custom.
- Choose a success metric, such as Clickthrough Rate or Cost-per-View.
- Select a campaign to use as your control.
- Google will duplicate the campaign and create the treatment arm.
- Modify the treatment arm by adding or removing video assets.
- Review the budget and set experiment dates.
- Save and start your experiment.
Viewing Results
Once your experiment runs, Google will split traffic evenly between the control and treatment campaigns. This ensures that you’re comparing apples to apples when analyzing the results. You can monitor the performance of both campaigns in real time and even end the experiment early if necessary.
Interpreting Results: Directional vs. Conclusive
While your experiment is running, you may notice “directional” results. These preliminary insights can guide your decision-making, but they are not conclusive. You’ll need to wait until the experiment reaches 95% confidence for conclusive results.
What is a Confidence Interval?
A confidence interval represents the likelihood that your results are accurate. For example, a 70% confidence interval means that if you repeated the experiment, you would get the same result 70% of the time. A 95% confidence interval means you can trust the result with much higher certainty.
End Your Video Experiment
When you’re ready to end the experiment, Google provides two options:
- If your treatment campaign was more successful, you can replace the control campaign with the assets from the treatment.
- If the control campaign performed better, you can end the experiment without applying the changes.
Advanced Custom Experiments
For advertisers who want more flexibility, the custom experiment option allows you to test up to 10 arms. This is ideal for campaigns that must test multiple variables simultaneously, such as different audience segments or ad formats.
Testing Non-Creative Variables
In addition to testing creative assets, you can use custom experiments to test variables like:
- Audience targeting
- Bidding strategies
- Ad formats
You can fine-tune your campaigns by experimenting with different settings for better performance.
Key Terms to Know
Here are some important terms to understand when running a video experiment:
- Hypothesis: The question or statement your experiment is testing.
- Experiment Arm: A campaign or group of campaigns in the experiment.
- Traffic Split: How you divide traffic between experiment arms.
- Statistical Significance: Confidence that the results are not due to chance.
Conclusion
Google’s new Video Experiment tool makes it easier than ever to optimize your video ads. Whether testing creative assets or targeting the audience, this tool provides actionable insights to help you make smarter decisions. By simplifying the setup process and offering both A/B and custom options, Google has made it easy for advertisers to run experiments and maximize the impact of their video campaigns.
