What a breakdown does
A breakdown takes the results you are already looking at and divides them along one dimension. In Meta Ads Manager the choices fall into a few groups: who saw the ad (age, gender, country, region), where it appeared (platform, placement, device, publisher), when it ran (day, week, time of day) and what happened (action type, destination, video view type).
Nothing is recalculated. The campaign total stays the same; you are looking at it through a different cut. That is worth remembering, because a breakdown never gives you new data, only a more specific view of data you already had.
Why breakdowns matter
A campaign average conceals the thing you need to fix. An acceptable overall cost per lead can hide one placement quietly consuming budget at several times the price of the others, and a healthy click-through rate can be one age band performing well while the rest do nothing. Until you split the numbers, you are optimising against a figure that describes nobody.
Placement is usually the most valuable split, because the same creative behaves very differently in a full-screen vertical Reel and in a right-column desktop unit. Region matters in markets where demand is concentrated in a few cities, which is the normal pattern in Nepal. Time of day matters for anything that depends on someone answering a phone.
Common mistakes with breakdowns
The biggest is acting on a segment with almost no data. Slice a campaign finely enough and every cut contains a handful of results, at which point the differences you are reading are noise. Look at the volume in each row before you believe the cost column.
The second is excluding segments too eagerly. Turning off a placement or an age band narrows the pool the delivery system has to work with, and a narrower pool often costs more overall even when the excluded slice looked expensive. Meta’s optimisation is already moving budget within the ad set; removing options is not always an improvement.
The third is expecting every metric to be available in every cut. Some combinations of breakdown and conversion metric cannot be reported, because the underlying measurement does not support that split, and the interface will simply not show the number. That is a limitation to work around, not a bug.
How to act on it
Use breakdowns to form a question, not to reach a verdict. If Reels shows a much lower cost per result than feed, the honest conclusion is that vertical creative deserves a proper test, not that feed should be switched off tonight. Build a separate ad set or a separate creative and let a real comparison answer it.
Keep the split shallow. One dimension at a time, over a period long enough to have volume, and always with the count column visible next to the cost. Then read it beside unique metrics so you can see whether a segment is genuinely reaching more people or simply serving the same people more often — the pattern that frequency exposes. Turning those splits into a report someone will actually read is part of ongoing Meta ads management.