How the lookalike percentage works
When you build a lookalike audience, Meta asks you to pick a similarity band alongside the country. That band is the lookalike percentage. It sets how large a slice of that country’s adult population Meta is allowed to include, and therefore how closely each person has to resemble your source list.
The narrowest band gives you the people Meta considers the closest match, and there are relatively few of them. Each step wider adds people who resemble the source less well, so the audience grows and the resemblance loosens. The bands are nested: a wider band contains everyone in the narrower ones, plus more. That single fact explains most of the trouble people run into with it.
Why the lookalike percentage matters
It is the dial between precision and reach. A narrow band tends to buy fewer, better-matched impressions, so it can hold up on a specific offer but will run out of room quickly. A wide band behaves much more like broad targeting: cheaper to reach, easier to spend against, and far more dependent on the creative to do the qualifying.
Country size changes the answer completely. In a market the size of Nepal, even the narrowest band is a modest number of people, and a wide one starts to approach ordinary interest-free targeting. In a large market the same settings behave very differently, which is why advice copied from an American case study often disappoints here.
Common mistakes with the lookalike percentage
The most expensive mistake is running several bands side by side without excluding the narrower from the wider. Because the bands are nested, those ad sets are competing for the same people, and you pay for the privilege.
The second is believing narrow is always better. A tight band on a small seed list can be so restrictive that delivery stalls, frequency climbs and costs rise — the opposite of what was intended.
The third is fiddling with the band when the real problem is the source. Widening or tightening a lookalike built from a weak or stale list only changes how many vaguely described people you reach.
How to act on it
Start with one band rather than a ladder of them. Choose narrow if your seed is strong and your offer is specific; choose wider if the country is small, the seed is thin, or delivery has stalled at a tighter setting. Judge it on cost per result over a full week, never on how the audience size looked when you created it.
If you do want to compare bands, exclude the narrower audience from the wider one so each ad set is buying different people. And before touching the dial at all, look at the seed audience: refreshing it with people who actually paid will move results further than any change to the band.