Why geo-targeting is not just setting a radius in kilometers
When someone sets up a campaign with geographic targeting for the first time, they tend to think that drawing a circle around their store is enough. In practice, geo-targeting in ads is a decision layer that combines location signals (IP, GPS, Wi-Fi, self-reported profile data) with account structure, business type, and the funnel. And that is where the nuances begin that only appear once you have real campaigns under your belt.
According to the platforms’ official documentation (Google Ads, Meta Ads, Microsoft Advertising), user location is inferred, not always verified with precision. That difference between physical, interest-based, and self-reported location is the root of most of the errors we will see in the following cases.
Case 1: the dental clinic paying for clicks from another province
A healthcare client had a Search campaign with keywords such as “dental implant price.” The report showed a high volume of clicks, but the calls were not coming in. When reviewing the settings, the campaign was targeted to “People in, or who show interest in, your location.” The result: users from other provinces looking for comparative information entered the ad and consumed budget with no intention of traveling.
The fix was not just switching to “People in your location.” We added a ZIP code radius layer to the highest-value ad groups (invisible orthodontics, implants) and left the broad radius only for the brand campaign. CTR dropped slightly, but the call conversion rate rose because the traffic was genuinely local. Practical lesson: for proximity businesses, the “interest” option almost never pays off; for tourism or ecommerce with national shipping, it can make sense.
Case 2: ecommerce with national shipping and the myth of excluding major cities
A store selling bulky products (furniture) assumed that large cities generated more returns due to logistics and excluded Madrid and Barcelona from its Shopping campaigns. The reasoning seemed logical, but the attribution data showed that those provinces concentrated the highest average order value and that many returns came from rural areas with delivery problems, not from the major cities.
Geo-targeting should not exclude entire provinces, but rather adjust bids for regions where shipping costs were disproportionate. We applied negative bid modifiers in areas with logistics surcharges instead of total exclusions. This is an important nuance: excluding a location eliminates the possibility of capturing demand; modifying the bid only adjusts profitability. For ecommerce, total exclusion is rarely the first option.
Case 3: B2B service with metropolitan region targeting
An industrial maintenance company wanted to attract clients in specific industrial parks. Radius targeting from the headquarters did not work because the industrial parks are on the outskirts. The solution was to use ZIP code targeting and “custom location” targeting with polygons drawn on the map. Here a real limitation appears: not all platforms allow drawing polygons with the same precision, and on Meta the minimum radius and the way location is inferred make precision lower than on Google Ads.
In practice, for industrial B2B, ZIP code targeting in Search was more reliable than radius targeting. And it is worth checking geographic location reports periodically, because ZIP code boundaries and metropolitan areas do not always match what one expects.
Case 4: lead generation campaign in several cities with limited budget
A client with franchises in five cities wanted a single campaign. The temptation is to target all five cities in one campaign and let the algorithm distribute. The problem: with a limited budget, the system concentrates spending where there is the most search volume, leaving smaller cities without exposure. The solution was to separate campaigns by city (or at least by clusters of similar size) to control the budget per market.
This is not a universal rule. If the budget is ample and the cities are homogeneous, a single campaign with geographic targeting can work. But when there is an imbalance in volume, a separate structure prevents some cities from eating up the budget of others.
Case 5: the interest location error in brand campaigns
In a brand campaign for a gym chain, the team activated “interest” to capture people who had searched for the brand before traveling. The result was clicks from users who were browsing from another city and then could not go to the center. In local brand campaigns, what works is targeting by physical location and, if anything, using interest location only in awareness campaigns, not in conversion campaigns.
Here it is worth being transparent: there is no configuration that guarantees results. Location inferred by IP can fail on mobile networks and VPNs, and platforms do not always warn about this clearly in the interface. That is why reviewing the geographic location report (country, region, city) is an ongoing task, not a one-time adjustment.
How to decide the granularity: radius, ZIP code, or region
The choice depends on the business model more than on the platform:
- Radius: useful for proximity services with a clear catchment area (restaurants, auto shops, clinics). Be careful with overly broad radii that include areas with no real intent.
- ZIP code: more precise for B2B and for services with logistics surcharges. It requires checking that the platform supports uploading lists and that the codes are up to date.
- Region or country: suitable for ecommerce with national shipping or awareness campaigns. Here geo-targeting is more useful for excluding unserved areas than for fine-tuning bids.
A detail that is often overlooked: in Search campaigns, location targeting does not prevent impressions from users searching from another area but typing the city name. If the ad includes the city in the text, the geographic filter and the copy must be aligned; otherwise, it creates a confusing experience and a click with a low probability of conversion.
Limitations worth acknowledging before scaling
Geo-targeting does not solve offer or pricing problems. If the product does not fit the area, better targeting will not fix it. Nor does it correct a landing page that does not mention the location or does not show local hours. I have seen well-targeted campaigns fail because the landing page was generic and the user could not find the nearest store.
Another limitation: measurement. Conversion attribution by location has delays and does not always reflect the user’s real location at the time of purchase. In businesses with a long decision cycle, the geographic report can mix first-click and last-click locations. That is why cross-referencing platform data with the CRM or call records helps validate whether the targeting is capturing the right traffic.
What to review in practice before launching
- Define whether the business is proximity-based, national, or mixed, and whether conversion depends on travel.
- Choose the location option (physical vs. interest) according to the funnel, not by default.
- Check whether the platform allows polygons, ZIP codes, or only radii, and adjust granularity to that capability.
- Separate campaigns when volume between markets is very uneven.
- Review the geographic location report periodically and cross-reference it with your own conversion data.
- Avoid total exclusions when bid modifiers are sufficient.
Geo-targeting in ads is a control layer, not a guarantee of performance. The cases above show that the value lies in deciding with judgment which location is used at each stage of the funnel and in accepting that precision has technical limits. Those who understand it this way tend to adjust their budget better and avoid the two extremes: targeting too much and ending up with no volume, or targeting so little that the ad is shown to people who will never be able to buy.