Geo Targeting: Where Your Ads Run And Where They Should

The short version

Geo targeting decides where your ads run. The part that costs people money is not the country they chose — it is the option underneath it.

Most platforms distinguish between people in a location and people interested in a location, and the default frequently includes both. If you sell locally, the second group is spend you did not intend.

Check this setting on every campaign. It is the single most common geo mistake.

Presence versus interest

Presence means the person is physically there, by device signal or IP.

Interest means they searched about it, browsed content about it, or otherwise showed interest in the place.

For a local service business, interest targeting means advertising to people planning a visit, doing research, or reading about your city from another country. They will never become customers, and they will click.

For travel, relocation, and education, interest is exactly what you want. The setting is not wrong — the default is just wrong for most advertisers.

Choosing the granularity

Granularity should follow how the business actually works.

Country. For anything that ships or sells nationally.
Region or state. Where regulation, language, or logistics differ meaningfully.
City or metro. For a service area business. The common default and usually right.
Radius. For genuine catchment areas. More honest than city boundaries for a physical location, since customers do not respect administrative lines.

A frequent error is targeting a city name when the real catchment is a radius that crosses city boundaries. You exclude nearby customers and include distant ones inside the same city.

Excluding is as important as including

Exclusions solve problems that inclusions cannot.

Areas you cannot serve or ship to.
Regions with historically poor conversion where you have enough data to be sure.
Places generating clicks and no business, which sometimes indicates click quality problems.

Exclusions apply before bidding, so they are cheaper than bid adjustments for anything you genuinely cannot serve.

Bid adjustments by location

Where supported, adjusting bids by location is usually better than excluding, because it keeps the option open.

The caution is sample size. Location performance data fragments quickly — split spend across fifty regions and most have too few conversions to support any conclusion. Adjust on regions with real volume; leave the long tail alone.

This is the same trap as dayparting: geography feels like a solid dimension, and the data underneath it is usually thinner than it looks.

Things that quietly go wrong

Language and location confused. Targeting a country does not target a language. Both need setting.

Time zones. Schedules usually run on the account time zone, not the viewer's. A campaign scheduled for business hours may not be running during business hours where the customer is.

Currency and pricing. Ads shown in a market whose landing page prices in another currency convert badly for a reason that is not the ad's fault.

Automated expansion. Some campaign types widen geography on their own. Verify where spend actually went rather than where you set it.

Practical rule

Set presence rather than interest unless you specifically want travellers and researchers, and check where spend actually landed rather than where you targeted. The gap between those two is where geo waste lives.