Google Ads has officially unveiled a new feature called ‘Campaign Guidance,’ designed to provide advertisers with an ‘Experiment Power’ score. This analytical tool aims to help marketers predict the likelihood of achieving statistically significant results before they commit to full-scale testing, offering actionable insights to improve experiment design.
Understanding the Experiment Power Score
The ‘Experiment Power’ score is a new metric that evaluates the reliability of a proposed ad experiment. By analyzing specific campaign attributes, Google provides a score ranging from 0% to 99%, categorized into three tiers:
- Low (0–49%): Likely to produce inconclusive results.
- Medium (50–79%): Better reliability, but may still lack statistical significance.
- High (80–99%): Strongest potential for conclusive, data-driven decision-making.
Key Factors Influencing the Score
Google determines these scores based on several critical parameters to ensure that experiments are set up for success. These include:
- Campaign Volume: The amount of historical data (e.g., conversions and spend) associated with the campaign.
- Campaign Variability: Consistency in past performance, which is vital for stable results.
- Traffic Split: Ensuring the experiment arm has enough volume to draw meaningful conclusions.
- Experiment Duration: The required timeframe to gather enough data, which varies based on campaign type.
- Experiment Type: Different tests, such as those for headlines or bidding strategies, require different volume thresholds.
How to Improve Your Experiment Power
If a campaign receives a low score, Google provides specific recommendations to ‘level up’ the experiment. These include increasing the experiment duration to collect more data, adjusting traffic splits (often recommending a 50/50 split), or selecting different campaigns that have higher baseline traffic and conversion volumes.
Strategic Benefits
The integration of Campaign Guidance is set to simplify the experimentation process for PPC professionals. By catching potential issues—such as insufficient duration or low traffic—early, advertisers can save budget and time that would otherwise be lost on inconclusive tests. This feature is currently available for Performance Max experiments and Broad Match experiments within Google Search campaigns.