Google Ads has rolled out a new way to test broad match keywords, and it’s far more streamlined than before.
Instead of duplicating your campaign, the new setup lets you run experiments within a single search campaign using built-in control and treatment arms. This is available only for campaigns using Smart Bidding, and portfolio bid strategies are not supported.
This update makes it much easier for advertisers to evaluate broad match performance without overhauling campaign structure.
Here’s how it works:
- Google will duplicate your existing exact and phrase match keywords as broad match versions.
- Traffic is automatically split 50/50 between the control (original keywords) and treatment (broad match) groups.
- Campaign-level changes during the test are applied to both arms, reducing setup errors and sync issues.
You can launch this experiment in two ways:
- From the Experiments section in your account.
- From the Recommendations tab by selecting “Apply as experiment” under a relevant suggestion.
Why this matters:
- No need to create separate campaigns.
- Results are more reliable and faster due to reduced “cold start” periods.
- You can track everything under the Experiments tab, with summaries that compare both arms.
Google recommends running the experiment for 6–8 weeks for the best insights. If results are favorable, you can later upgrade your keyword strategy to include broad match.
Read the official help page found by Hana Kobzová here: About new broad match experiments.