Meta's fragmentation recommendation tells you to combine ad sets that share variables like objective, audience or creative, and consolidating is often the right call. But the recommendation is built to improve your account's setup score, not to protect your best result. Before you merge anything, open the cost per result column, find out which ad set is winning, and make sure the winner is the one that keeps running.
What the recommendation is built to do
Meta's Marketing API documentation describes opportunity score as a tool for understanding how well-optimized an ad account is, and says implementing recommendations will "improve setup and increase opportunity score." Meta also says its recommendations are tested to improve performance, and often they do.
The fragmentation recommendation fires when two or more ad sets have similar variables, such as objective, audience or creative, while some variables differ. So two ad sets can share an objective and still be aimed at completely different people. The documentation doesn't say the recommendation weighs which ad set is producing purchases at the lowest cost. That check is yours.
The call where the cheapest ad set got switched off
On a mentoring call, we went through the Meta account of a business selling a cleaning product for sports gear, at about $15 a unit. About $2,600 had gone into Meta in roughly two months, without a profit yet, and 15 campaigns had been launched in about 50 days.
One paused ad set stood out: a boxing influencer explaining his smelly gloves and how the product fixed them, at under $1.50 per purchase. The campaign still running had a golf ad set at $52 per purchase and a sports parents ad set at $23.
It had been paused because the recommendations panel flagged fragmentation and said combining would lower costs. So the boxing ad was moved in next to the golf and sports parents ads, and the original ad set was turned off. Boxers, golfers and parents of young athletes are not the same audience.
Our advice on the call was blunt. Don't act on a Meta or Google recommendation because the platform suggests it. Look at cost per result first. Relaunch the boxing ad set on its own and see if the results are consistent. And run two or three campaigns long enough to mature, instead of launching new ones every few days.
When to merge, and how to keep the winner
We usually argue for fewer ad sets on Meta. Ad sets chasing the same people compete in the same auctions and split the conversions each one needs to learn, which is why one campaign beats five. That still holds. The fragmentation flag is a prompt to check, not a verdict. Before you act on it:
- Compare cost per result for every ad set the recommendation names, over the same date range.
- Ask whether they reach the same people. Same audience, different creative: merge, and put the creatives in one ad set. Different audiences with very different costs: that's a test result, and splitting them can be a deliberate choice, like splitting by the states your orders come from.
- Merge into the winner. Keep the best ad set running and bring the other ads into it, not the other way round.
- Test it if you're unsure. Use Meta's A/B test tool, because two ad sets in one campaign is not an A/B test.
Google's recommendations can rewrite an account too. On either platform, a recommendation is an input. Your cost per result is the decision.
The takeaways
- Meta's fragmentation recommendation fires when ad sets share an objective, audience or creative, even if they reach different people.
- Meta says applying recommendations improves setup and raises your opportunity score. Check what it does to cost per result yourself.
- Merge ad sets that reach the same people, and merge into the winner, never out of it.
- If a paused ad set was your cheapest, relaunch it on its own and see if the result holds.
Common questions
What does Meta's fragmentation recommendation mean?
Meta's Marketing API documentation describes the fragmentation recommendation as recommending that you consolidate ad sets to improve liquidity when two or more of them have similar variables, such as objective, audience or creative, while some variables differ. It appears in Ads Manager as one of the recommendations behind your opportunity score.
Does Meta's opportunity score measure performance?
Not directly. Meta describes opportunity score as a tool for understanding how well-optimized an ad account is, says a higher score indicates a greater likelihood of improved performance over time, and says implementing recommendations will improve setup and increase the score. It is a measure of setup against Meta's best practices, not a cost per result.
Should I always merge ad sets when Meta recommends it?
No. Merging ad sets that target the same people is often the right move, because they compete for the same auctions and split their conversions. But the recommendation can fire when ad sets only share an objective. Check cost per result for every ad set it names first, and don't merge away an ad set that is clearly outperforming the rest.
How do I consolidate Meta ad sets without losing the one that works?
Compare cost per result for every ad set over the same date range, keep the best one running, and move the other ads into it rather than moving the winner into a weaker ad set and pausing it. If you want proof that merging helps, test it with Meta's A/B test tool instead of comparing two ad sets in one campaign.
When is it right to keep Meta ad sets separate?
When the ad sets reach genuinely different people, their costs are very different, and each has enough budget to get conversions. Separate audiences with separate results are a test result, not clutter. If two ad sets reach the same people with different creative, put the creatives in one ad set instead.