How detailed targeting works
Detailed targeting is the box inside a Meta ad set where you add interests, behaviours and demographic categories on top of the basics of location, age and language. Each entry describes a group of people Meta believes fits that category, based on what they do across its platforms and on inferences drawn from that activity.
The logic in that box catches people out, so it is worth stating plainly. Everything you type into the same field is joined with or: the audience becomes anyone matching any one of the entries, and adding more makes it larger. Using the option to narrow the audience creates a second field joined with and, so a person must match something in the first field and something in the second. Widening and narrowing therefore look almost identical on screen and do opposite things.
Two limits apply. Certain campaigns fall into special advertising categories — credit, employment, housing and social or political issues in the markets where those rules exist — and detailed targeting is heavily restricted for them. And the list of available categories has been trimmed over time as sensitive options were withdrawn, so an interest that worked in an old campaign may simply no longer exist.
Why detailed targeting matters
It is the part of a core audience where advertisers spend the most effort and, often, gain the least. Used sparingly it does real work: excluding an obviously wrong crowd, or reaching a professional group that ordinary demographics cannot describe. Used heavily it produces an audience that nobody can explain and that the delivery system would have found anyway.
It matters more in small markets than large ones. In Nepal the total addressable audience for many businesses is modest to begin with, so a stack of interest filters can shrink an ad set to the point where delivery becomes unstable and costs swing from day to day.
Common mistakes with detailed targeting
The first is reading an interest as an intention. A category records a pattern of behaviour, not a decision to buy, so someone grouped under an interest in property may be a broker, a student or a person who once watched a house tour. The creative, not the checkbox, is what separates a buyer from a browser.
The second is testing interests without a clean test. If expansion is switched on, delivery can serve outside your selections, so a result cannot be attributed to the interest you chose. The third is layering so many filters that the ad set never gathers enough results to leave the learning phase, at which point you are reading noise and calling it a finding.
How to act on it
Start with the fewest selections that make the audience sensible, and write down what each one is supposed to prove. Test a small number of hypotheses against a genuinely broad ad set, with the same creative in each, and give every one enough budget and enough days to produce a stable cost per result.
Use the narrowing option deliberately rather than by habit, and check the audience size estimate after every change so you can see whether you have just widened the pool when you meant to tighten it. If a broad ad set matches or beats your carefully built interest stack, keep the broad one and put the saved effort into stronger creative, which is doing most of the targeting now anyway.