In an ad platform's dashboard, location targeting often looks like a single dropdown: type a city, pick it, move on. Behind that dropdown, though, sit two quite different questions, and mixing them up can send part of a budget to people who could never actually become customers. For an agency running campaigns in a market as small as North Cyprus and one as large as Türkiye at the same time, getting that distinction right matters more than most of the rest of the campaign build.
Where someone is, versus where someone is interested in
Google Ads and Meta ads both fold two separate audiences into a single location setting: people physically present in an area right now, and people searching about that area because they are interested in it. The default setting usually covers both, because the platform's logic leans towards widening reach rather than narrowing it.
An example makes the difference concrete: someone who has never set foot on the island but is planning a holiday and searches "Cyprus" is exactly the person a hotel or tour operator wants to reach. If that same click lands on a neighbourhood bakery's ad, it is simply wasted spend. The bakery needs the "interested in" layer switched off and only the "located in" option left on; for the hotel, that same layer is where the real value sits.
When a radius beats a city
Targeting by city or district follows an administrative line, but customer behaviour rarely respects that line. For a physical shop or a business that provides service on site, a radius drawn around a single point usually reflects the real audience better than an administrative boundary does. Someone living just outside Nicosia's city limits can often reach a shop in the centre more easily than someone from a different district technically inside those limits; targeting by district name alone can leave that person out entirely.
What decides the radius is the business itself. A plumber or a removals firm with a wide service area might reasonably extend it to thirty or forty kilometres; a café or beauty salon relying on daily footfall stays within a few. City-level targeting still makes more sense for a business with several branches or one serving an entire province — the point isn't that one approach is universally better, it's that each has to match the physical reality of the business.
North Cyprus's scale problem: narrow targeting can shrink an audience past the point of learning
Because the population in North Cyprus is small, narrowing targeting carries a cost that often goes unnoticed. Combining a tight radius with a narrow interest or behaviour layer at the same time can leave too small an audience for the platform to learn from. Once the algorithm can't gather enough signal, delivery becomes erratic, cost swings unpredictably, and the campaign never settles into a stable pattern.
The fix isn't to widen everything, but to decide which layer gets loosened. If location has to stay tight, opening up the interest or behaviour layer a little — or the reverse — brings the audience back to a size the system can actually learn from. Building a campaign on the island with the same tightness as one in a large city ignores what the smaller market actually is.
On the Türkiye side: by province, or country-wide
Targeting Türkiye with the same budget leaves two main routes: keeping provinces like İstanbul, Ankara and İzmir in separate campaigns, or setting one country-wide target and letting the platform handle distribution. That choice is made with accumulated data, not instinct. Splitting by province makes it visible which region actually converts, but when the budget per province is too thin, each campaign can end up too small to produce meaningful data on its own.
Starting with a single country target lets the platform decide internally where to weight spend, and on a modest budget that usually delivers more consistently. The sound approach is normally to use both in sequence: begin with a country-wide campaign to see which regions stand out, then, as budget grows, split the strongest provinces into their own campaigns. There's no cost to waiting for that decision; the real risk is splitting by province too early, on an assumption about "the big city" rather than on data that has actually accumulated.
Language targeting is not location targeting
This is another pair of settings that gets conflated. A platform can flag a user as Turkish-speaking purely from device or browser language, which has nothing to do with where that person physically is. Someone living in Germany with a Turkish-language browser can end up seeing an offer meant specifically for North Cyprus or Türkiye in a campaign where location targeting wasn't set up properly — even though they could never use that offer.
The reverse happens too: someone living in Türkiye whose device is set to a different language can be excluded entirely from a campaign built around a language filter alone. Getting this right means treating location and language as separate layers, each set on its own logic; using one as a stand-in for the other quietly loses part of the audience.
A Google Business Profile answers the same intent through a different door
On the local search side, a Google Business Profile is not an ad, yet it serves exactly the same intent as one: someone searching "near me", or a service alongside a city name, trusts a map result about as much as an ad. Keeping that profile current — the right address, opening hours, the right category — strengthens the ground that paid targeting is built on. Running the two apart is effectively greeting the same user at two different doors and leaving one of them unattended.
The risk of shifting budget without checking the location breakdown
When a campaign's overall numbers look good, the decision to increase budget tends to get made quickly — but a decision made without looking at where those numbers actually came from risks growing the wrong thing. Reallocating budget without ever opening the location breakdown can mean ignoring a weak-performing region while pouring more money into an average that only looks healthy in aggregate. That risk grows larger on a channel like remarketing, which speaks to an already warm audience and so tends to look good by default. Opening the location breakdown before deciding is the one step that actually shows which setting is doing the work.