Google Ads is introducing Campaign Guidance, a new experiment support feature that gives advertisers an “Experiment Power” score before they run a test.
A Google Ads Campaign Guidance With Experiment Power Score
The score estimates the likelihood that an experiment will produce statistically significant results. It is available for Performance Max experiments and broad match experiments in Search campaigns.
This is a small update on the surface, but it solves a real problem: too many Google Ads experiments end without a clear winner.
What changed
Campaign Guidance evaluates an experiment setup before launch and gives it a power score. Google groups the score into three ranges:
- Low: 0–49% power
- Medium: 50–79% power
- High: 80–99% power
The score considers factors like campaign volume, campaign variability, traffic split, experiment duration, and experiment type.
In practical terms, Google is warning advertisers when an experiment is likely to be inconclusive before budget is spent.
Why this matters
A/B testing in Google Ads sounds simple until you run it in a low-volume account.
Many B2B SaaS advertisers test bidding strategies, broad match, Performance Max, or creative assets without enough conversion volume to learn anything reliable. After three weeks, the test ends with mixed results, no clear winner, and a wasted decision cycle.
Campaign Guidance should reduce that.
It gives advertisers a better sense of whether the test has enough data behind it. That is especially useful for teams testing Performance Max or broad match, where algorithmic learning already depends on sufficient conversion and spend volume.
We have covered this issue before in our Performance Max guide.
What affects Experiment Power
Google says Experiment Power can be affected by campaign selection, campaign consistency, traffic split, experiment duration, and the type of experiment being run.
That means advertisers can improve the score by:
- Choosing campaigns with more conversion volume
- Running experiments for longer
- Using a cleaner traffic split, often 50/50
- Avoiding unstable campaigns with erratic performance
- Selecting a test type that matches the available data
This does not mean every advertiser should chase a high score at all costs. Sometimes a low-volume campaign still needs directional testing. But it does mean teams should stop treating weak tests as final answers.
OneMetrik’s take
This update is useful, but it should not replace judgment.
For B2B SaaS advertisers, the most common mistake is not running too few experiments. It is running experiments that the account cannot support statistically. A campaign with five conversions a month is not a reliable test environment for major bidding or match type changes.
Before using Campaign Guidance, advertisers should audit the basics:
- Is conversion tracking clean?
- Are primary and secondary conversions separated?
- Is offline conversion data flowing back into Google Ads?
- Are campaigns structured by intent, funnel stage, or audience quality?
- Is there enough budget to let the experiment complete?
Our Google Ads audit guide is a good starting point for that foundation.
What advertisers should do now
Do not launch every recommendation blindly. Use Campaign Guidance as a filter.
Before starting your next Performance Max or broad match test, check the Experiment Power score. If the score is low, ask why. Is the campaign too small? Is the test too short? Is the split too conservative? Is performance too volatile?
Then decide whether the test needs to be redesigned or whether the answer should come from a different campaign.
For teams using Google Ads automation, this update reinforces a bigger point: automation works best when the inputs are clean. Google can guide the experiment, but it cannot fix poor campaign structure, weak conversion data, or unclear business goals.
We explain that in more detail here.
Final thought
Campaign Guidance is not a flashy Google Ads feature. But it may become one of the more useful ones for serious advertisers.
It helps teams avoid the worst kind of experiment: the one that consumes budget, delays decisions, and still tells you nothing.
For B2B SaaS and performance marketers, that is a welcome shift.