Your campaign says ROAS 2.1. Fine — not great. So you nudge the budget and move on. But "2.1" is an average, and averages lie. Inside that one number is probably a 25-34 segment doing 4.0 and a 45-54 segment doing 0.6, quietly burning cash every day. The Breakdown report is where you find that out. Most founders never open it.
Where the Breakdown report lives
In Ads Manager, above your campaign table, there's a Breakdown dropdown. Click it → By Delivery. That's where Age, Gender, Placement, Platform, Time of Day, and Country live. Pick one. Your table re-splits every row by that dimension.
That's the whole feature. The skill isn't finding it — it's knowing which split to run and what to do with what you see.
Age — find the band that actually buys
Run the Age breakdown first. It's the highest-signal split. You'll usually see one or two age bands doing most of the work.
Common pattern: you targeted 18-65, the campaign reads ROAS 2.0, but the breakdown shows 25-34 at 3.5 and 18-24 at 0.7. The young band is cheap to reach (low CPM), so Meta keeps spending there, and it drags your average down.
What that tells you: your product resonates with 25-44. But don't rush to exclude — see the trap below.
Gender — usually a quieter signal
Gender splits matter most when your product skews one way and you targeted both anyway. A skincare brand running "All genders" might find women at ROAS 3.0 and men at 0.8. If the gap is that wide and both have 20+ conversions, narrowing gender is a fair call.
If the split is 2.4 versus 2.0, leave it. That's not a real difference. It's noise wearing a costume.
Placement — where the budget quietly leaks
This is the breakdown that finds wasted money fastest. Switch to Placement. Now you see Facebook Feed, Instagram Feed, Stories, Reels, and Audience Network as separate rows.
Two things to look for:
- Audience Network often shows a very low CPM and a very low conversion rate. Cheap impressions, weak results. It can soak up 15-20% of spend for almost nothing.
- Reels vs Feed — Reels CPMs are usually cheaper, but a static square image jammed into Reels converts badly. If your creative is a square image, Feed is doing the real work.
Placement also tells a creative story. If Stories is weak, it's often because your creative isn't built vertical (9:16). That's a creative fix, not a placement cut.
The trap nobody talks about
The campaign that looked fixable gets more expensive instead. The better move: use the breakdown to inform creative, not just to cut. Bad Stories number → make a vertical version. Bad 45-54 → your hook speaks to 30-year-olds, so write one that speaks to 50-year-olds. Cut only when a segment has 20+ conversions AND a clearly broken ROAS AND you've already tried fixing the creative.
Quick reference
| What the breakdown shows | What to do |
|---|---|
| One age band 2-3x better, 20+ conv each | Test a campaign focused on that band — don't just exclude the rest |
| Audience Network: low CPM, near-zero conv | Safe to exclude — it rarely earns its spend |
| Stories/Reels weak, Feed strong | Creative problem — build a 9:16 version before cutting |
| Gender gap under ~20% | Ignore it — noise |
| Any segment under 15 conversions | Wait — not enough data to act |
What to do next
The breakdown report tells you where the leak is. It doesn't tell you what to pour in. When you generate a blueprint in AdBlueprint, the audience and placement recommendations already account for which segments tend to carry Thai SMB campaigns. You start closer to the winning split, instead of finding it three weeks and ฿15,000 later.
