Running a geo holdout or exclusion-based lift test can be a powerful way to measure the true incremental impact of your campaigns. But will excluding a region (or multiple) trigger a re-learning phase for your campaigns? The impact of running geo lift tests can have different effects depending on the advertising platform you're using.

Meta Ads explicitly include audience changes (such as location exclusions) as a trigger for resetting the learning phase.

Any time you change the targeting settings of an ad set (even just by removing one region), the algorithm starts re-learning, as it has to optimize delivery based on a new audience mix.

Google Ads does not typically re-enter the learning phase when you exclude or hold out a geo location—unless that region accounted for a large volume of conversions or traffic.

However, if you’re using automated bidding strategies (like Target CPA or Maximize Conversions), removing a high-performing location can cause temporary performance fluctuations as the algorithm reallocates budget across the remaining areas.

TikTok Ads does not display a visible learning phase indicator like Meta, but location exclusions can trigger re-optimization, especially if the held-out region had significant engagement or conversion volume.

Like Meta, TikTok relies on audience signals for delivery. Changing geo targeting alters the available audience pool, which can cause the system to adjust how it bids and delivers.