Obvi split a 50-ad CBO into 5 ad sets: NC-ROAS +55.65% (1.79)
Commentary byROASArchiveOur analysis, not theirs@AshvinMelwani reports splitting a scaling CBO campaign's single 50-ad ad set into 5 ad sets, and new-customer ROAS (NC-ROAS) rose 55.65% to 1.79 on day 2 of the new setup.
The post doesn't name the account; our records list it as Obvi.
It is an early before/after, not a split test.
- NC-ROAS: 1.79 after the split, up 55.65%. NC-ROAS counts only revenue from new customers, so this is a stricter measure than plain ROAS.
- Structure: 50 ads across 5 ad sets after the change.
- Timing: posted on day 2 of the new setup, so this is a very early read. Spend, purchases and the comparison period behind the +55.65% are not shared, so the size and reliability of the gain are unclear.
The change
Before
A Meta scaling CBO campaign with all 50 winning ads in one ad set, an inflated budget, a cost cap bid set to breakeven and all customers excluded. The earlier NC-ROAS figure is not shared beyond the +55.65% change.
After
The same 50 ads were split across 5 ad sets inside the CBO campaign. Budget, the breakeven cost cap and customer exclusions stayed the same, and a separate ABO testing campaign was left as is.
Our reading: why it likely worked
Our interpretation, not the advertiser's claim.
- With 50 ads in one place, Meta tends to push most spend to a few ads. Splitting into 5 ad sets forces more ads to get delivery.
- CBO can then move budget between the 5 groups, so strong themes get more money without starving the rest.
- Grouping by theme or angle may also help each ad set find a different pocket of new customers, which would show up in NC-ROAS.
Caveats: Self-reported before/after on one supplements account (Obvi), not a split test. It was posted on day 2 after the change, and spend, conversions, the comparison period and the attribution setup are not shared, so the gain could partly reflect timing or other changes.
How to run this in your account
- Hypothesis
- If we split a large single-group CBO into a few themed ad sets, then new-customer ROAS will rise because more ads get spend and budget can follow the best themes.
- Primary metric
- NC-ROAS
- Guardrails
- Spend share per ad set
- New-customer CPA
- Frequency
- Who should try it
- Accounts running dozens of ads in one ad set where a few ads take almost all the spend. It probably won't help small accounts that can't feed 5 ad sets with enough budget.
Questions
Should I put all my ads in one ad set in a CBO campaign?
In Obvi's case, splitting 50 ads into 5 ad sets came before a 55.65% rise in NC-ROAS, to 1.79. It is a single before/after, so test it on your own account.
What is NC-ROAS?
New-customer ROAS: revenue from first-time buyers divided by ad spend.
It strips out repeat buyers, so it shows if ads are actually finding new customers.