One marketer ran a simple test. And what happened next might change how you manage your ad sets.

If you’ve been managing Meta Ads for any length of time, you’ve probably had this rule drilled into your workflow: When you add a new ad to an existing ad set, the whole thing goes back into learning.

That’s not just something people randomly practice. It’s something Meta itself has always said. Their Help Center article still explains that even a small edit, like adding a new ad, can reset the learning phase. And because of that, a lot of advertisers avoid tweaking high-performing ad sets once they go stable. No one wants to risk performance dropping off just because you added a fresh creative.

But something interesting just came up.

A small test. A surprising result.

Jon Loomer, who’s been one of the most trusted voices in Meta Ads for years, ran a very simple test. He took a live ad set, one that had 22 ads running, and added one more ad to it.

Now normally, you’d expect that to trigger a reset. The ad set would enter learning again, and you’d see delivery become unpredictable for at least a few days.

But that didn’t happen.

The new ad got approved. It started serving impressions. And the ad set stayed right where it was, without returning to learning.You can read the full breakdown of his test on his blog. He walks through the account setup, the number of ads, and what he observed.

PS: Don’t consider this an official feature announcement. But it’s one of those quiet behaviour shifts that could easily go unnoticed if you aren’t paying attention.

So what’s actually happening?

That’s still a bit unclear.

Meta hasn’t updated its documentation. The Help Center still states that edits like adding ads will trigger a reset. But in this test, that doesn’t seem to be happening anymore.

So maybe it depends on how big the change is. For example, if you’re adding five new ads to a tiny ad set, it could still push it back into learning. But if you’re adding one or two ads into an already busy ad set, maybe it slides through without disrupting anything.

It feels like Meta might be treating some changes as more minor than before. And that could be a very welcome shift for advertisers.

What should you take away from this?

This doesn’t mean you should start dumping new creatives into every live ad set tomorrow. But it does mean you might not need to be as cautious as you once were.

If you’ve been holding back on refreshing creatives just to avoid the learning reset, it could be worth testing again. Add one new ad to a healthy ad set and keep an eye on whether the learning phase gets triggered or not. You may find that nothing changes and that opens up some breathing room for creative testing and iteration.

Still, treat this carefully. This might not apply to all accounts, campaign types, or optimization goals. Until Meta makes it official, consider this a working insight, not a permanent rule.

Final thought

Meta’s ads platform evolves quietly sometimes. Rules that felt carved in stone can shift in the background, without any formal announcement. That’s why small tests like this matter. And credit goes to Jon Loomer for not just spotting the change, but running the kind of test that helps everyone else pay attention too.

Want to try it yourself? Pick a live ad set and run your own version of the test. Just one new ad. Watch what happens. You might be surprised.