Google Ads has a new feature named Campaign Guidance that will give you an "Experiment Power" score that will show advertisers "the likelihood of achieving statistically significant results, along with actionable recommendations to improve your experiments and make informed decisions," Google wrote.
This should give advertisers more confidence in the experiments they run and if they will succeed or not. Hana Kobzová, who spotted this first and wrote about it on PPC News Feed wrote, " feature offers an ‘Experiment Power’ score to predict the likelihood of achieving statistically significant results and provides actionable recommendations for improvements."
The "Experiment Power" scores are interesting to me, Google explained the factors involved in this score and they include:
- Campaign selection: Campaigns with sufficient volumes are necessary for a healthy experiment setup.
- Campaign variability: Campaigns with consistent performance over time lead to more reliable experiment results.
- Traffic split: Sufficient traffic volume needs to be in the experiment arm to provide conclusive results.
- Experiment duration: Experiments should run for a minimum duration based on the campaign volume and traffic to ensure reliable and conclusive results.
- Experiment type: Test settings like headlines, bidding strategy among others, require variable experiment durations and minimum performance volumes for conclusive results.
The experiment score has three categories of ranges including:
- Low: 0 - 49% power
- Medium: 50 - 79% power
- High: 80% - 99% power
How do you level-up (sorry, had to) your Experiment power score? Google says:
- Choosing campaigns with sufficient volume: Running experiments with campaigns that have a large enough volume of data (for example, conversions, spend, among others) can significantly increase the power of your experiment.
- Adjusting the experiment duration: Increasing the duration of your experiment will allow you to gather more data, resulting in higher power.
- Optimizing the traffic split: For faster results, consider using a 50/50 split between your control and treatment groups.
- Selecting alternative campaigns: Explore other campaigns with higher volume for your experiment.
Here are the listed benefits of Campaign guidance with experiment power scores:
- Run more effective experiments: Avoid inconclusive results or unnecessarily long experiment durations.
- Get actionable insights faster: Make data-driven optimizations and improve campaign performance more quickly.
- Get personalized recommendations: Adjust budgets, experiment durations, or campaign selections to increase the likelihood of conclusive results.
- Simplify experiment setup: Get clear, easy-to-understand explanations and actionable steps to run successful experiments.
This feature is available for all Performance Max experiments and Broad Match experiments in Google Search campaigns.
For more details, so this Google help document.
Forum discussion at X.
