Google Ads Location Targeting 2026: Presence or Interest Explained
The campaign targets Chicago. The geographic report lists Toronto, Manila, Lagos and Warsaw — together 18% of spend. Nobody edited the targeting. This is Google’s default setting doing exactly what it says: presence or interest targeting serves ads to people who are in your area and to people who merely show interest in it from anywhere on earth.
Below: what both options actually mean, how location is inferred, which report exposes the leak, when the wider setting genuinely makes money, and how to fix it in five minutes.
Presence or interest targeting: the two options
Advanced location settings live under Campaigns → Settings → Locations → Location options. Two groups: one for included locations, one for exclusions.
| Option | What it does | Status |
|---|---|---|
| Presence or interest | Reach people in, regularly in, or who have shown interest in your targeted locations | Default for targeting |
| Presence | Reach only people in or regularly in your targeted locations | Must be selected manually |
| Presence (for exclusions) | Exclude people who are located in the excluded area | The only exclusion option available |
An important detail as of 2026: the standalone “search interest” targeting option — reaching people who search about a place without being there — is no longer available in Search, Shopping and Display campaigns, and neither is the “presence or interest” exclusion. Google narrowed the choice to two working modes, so the whole decision reduces to one question: do you serve customers who are not physically in your market?
How Google decides where someone is
“Presence” is a probabilistic estimate, not a GPS reading. The system combines several signals:
- device IP address — the primary signal on desktop;
- GPS, Wi-Fi and cell tower data on mobile;
- Google account location history, where the user has enabled it;
- how often the person appears in the area — the source of the “regularly in” wording.
“Interest” is inferred differently: from a place name inside the query (“hotels in Denver”), from the content of pages viewed, from previous searches about the region. Someone sitting in Berlin searching “flower delivery Chicago” is a relevant impression under presence-or-interest.
Both estimates carry error. Corporate VPNs, carrier-grade NAT on mobile networks and roaming all misplace users. Airtight geography does not exist in paid search — the goal is to push leakage from double digits down to low single digits.
Why the wider mode is the default
Presence-or-interest expands reach. For Google that means more auctions; for some advertisers it genuinely means more revenue. The trap is assuming “default” equals “right for you”.
The rule of thumb: if your product is bought remotely and your service area matches the targeted region, interest is a valid signal. If you cut hair, serve food, repair things or deliver physically, interest is noise you pay for.
The report that shows what is really happening
Open Campaigns → Insights and reports → Locations → Geographic report. There are two fundamentally different views:
- User locations — where people physically were. This is the one that exposes leakage.
- Locations of interest — which region the user cared about. This view looks tidy even when your traffic has scattered across three continents.
One-step diagnosis: open the user locations view, add Cost, Conversions and Cost/conv. columns, sort by cost. Anything outside your market that consumes more than a couple of percent of budget is a direct exclusion candidate.
The report is far more usable once you build the right column set in advance — the method is covered in custom columns and the report editor.
When presence or interest earns its keep
- Travel and hospitality. A Denver hotel is searched for from Chicago and Dallas. Presence-only cuts off nearly all demand.
- Relocation and real estate. People shop for housing in a new city before they move, not after.
- B2B with national service delivery. The buyer approving a Phoenix installation sits in a head office two states away.
- Events and ticketing. The audience travels in from everywhere.
- Medical and education “tourism”. Patients and students research clinics and universities from home.
When you need presence only
- Local services requiring a physical visit. Salons, dentists, auto repair, food delivery inside a three-mile radius.
- Retail with a storefront. Your shop cannot help someone in another country, but the click is billed anyway.
- Delivery with hard geographic limits. Single-city courier operations.
- Licence- or jurisdiction-bound offerings. Where serving out-of-area customers is not legally possible.
- Small-budget tests and pilots. Clean geography beats reach when you need a readable result rather than scale.
The hidden cost: you are training the algorithm on junk
Wasted impressions are only half the damage. Automated bidding learns from conversions, so if part of your leads come from areas you do not serve, the algorithm treats that traffic as good and chases more of it.
The chain: wide geography → a lead from an unserviceable area → sales rejects it → Google Ads still counts it as a conversion → Smart Bidding bids up on similar users → the share of junk grows. A month later the campaign looks excellent on CPA and terrible on revenue.
The fix is two-part: clean geography, plus splitting conversion actions so only qualified leads feed optimisation. Mechanics in primary vs secondary conversions and lead quality. When the sale closes offline, offline conversion import is sharper still — the algorithm sees deals rather than raw form fills.
Five common geo mistakes
1. Excluding regions instead of narrowing the target
“Target the whole country and exclude the rest” feels flexible, but exclusion lists grow until nobody maintains them. Including only the areas you sell in is more robust.
2. Setting a radius without checking the map
A six-mile radius in a dense metro can cover two neighbourhoods separated by a river and an hour of traffic. Verify coverage visually, not numerically.
3. Never re-checking settings after copying campaigns
Campaigns imported from another account or restored from drafts routinely arrive with defaults restored. Audit them in bulk — Google Ads Editor bulk operations exposes location options as a column.
4. Mixing regions with different economics in one campaign
If metro CPA runs double the suburban figure, a single campaign averages everything and removes your levers. Splitting by geography — and measuring the effect properly — is the subject of incrementality and geo experiments.
5. Confusing location with language
Geography and language are separate axes. Someone in your city browsing in another interface language may never see the ad. Worth testing explicitly when opening new markets — see the guide to international campaign launches.
Choosing granularity: country, region, city, radius
The second decision after presence-or-interest is how finely to slice geography. One rule covers most cases: pick the largest unit that does not produce impressions outside your market.
| Level | Fits | Risk |
|---|---|---|
| Country | Digital products, nationwide shipping, SaaS | Averages economics across very different regions |
| State / region | Distribution with regional warehouses | Administrative borders rarely match delivery zones |
| City | Local services, retail with several stores | Suburbs drop out even though customers drive in |
| Radius around an address | Single location, “we deliver within N miles” | A radius is a circle; service areas almost never are |
| Postal codes | Exact delivery zones, neighbourhood tests | Long lists that nobody remembers to update |
Practical advice: do not mix levels inside one campaign. When the list holds a country, three cities and two radii at once, reporting turns to mush — you cannot tell which included area produced the impression.
Location bid adjustments in the automated bidding era
Under manual bidding, location adjustments were the main lever: −40% on a weak region, +20% on a strong one. Automated strategies changed that.
With Target CPA or Target ROAS, location bid adjustments are largely ignored — the algorithm evaluates regional conversion likelihood itself, alongside a dozen other signals. The exception is −100%, which functions as an exclusion.
What to use instead:
- Split campaigns when regions need different CPA targets or separate budgets.
- Assign different conversion values by region when a deal in one city carries more margin. Mechanics in conversion value rules and value-based bidding.
- Exclude outright when the economics never work at any bid.
Connecting ad geography to actual sales
Google Ads knows the impression location. Your CRM knows the customer’s location. Until those two are joined, you manage geography blind — you see leads but not which ones closed.
A minimal working setup:
- Capture the customer’s city as an explicit form field. Do not infer it from IP.
- Pass the click identifier into the CRM alongside the lead.
- Export closed deals weekly and compare their city distribution against the user locations report.
- Regions producing leads but no closed deals for three consecutive months are exclusion candidates, no matter how attractive their in-platform CPA looks.
When deals close offline, the next step is feeding them back through offline conversion import — after which the algorithm stops treating a zero-revenue region as promising.
VPN and corporate network traffic
Some out-of-area impressions come from location inference rather than settings. Three usual sources:
- Corporate networks. The employee sits in your city while their traffic exits through a data centre abroad. Under presence-only they may never see your ad — the flip side of a strict setting.
- Carrier-grade NAT on mobile. The IP belongs to the carrier’s registration region, not the subscriber’s position. Other mobile signals partly compensate.
- VPNs and privacy browsers. A small but growing share of the audience that cannot be filtered out entirely.
The pragmatic stance: do not chase zero leakage. Treat 2–5% of spend outside your market as acceptable background — a guide figure, not a standard. Anything above that is nearly always settings rather than technology. If the “unknown” geography bucket in your report is large, check tracking as well: telling an inference problem apart from a measurement problem is covered in conversion tracking diagnosis.
One planning note: a geography cleanup almost always frees budget. To stop that money from accelerating spend elsewhere beyond plan, re-model your monthly ceiling — the mechanics are in Google Ads budget pacing.
A 20-minute geography audit
- Export every campaign with its location options (Editor or a settings report). Flag the ones set to presence or interest.
- For each flagged campaign, answer in writing: do we serve a customer who is physically outside the region? If not, switch to presence.
- Open the user locations report for the last 90 days. Sort by cost, view by country and region.
- List everything outside your market costing more than 1% of budget. Check whether it produced conversions and what happened to them in the CRM.
- Exclude irrelevant countries and regions at campaign level.
- Note the date of the change. After 14 days, compare out-of-market spend share and cost per conversion — the effect of a geography cleanup usually shows inside that window.
- If volume drops, do not immediately revert. First check whether you have hit a ceiling: impression share and auction insights will tell you whether the constraint is budget or rank.
One campaign for all regions, or several?
The splitting question arrives right after a geography cleanup. There is no universal answer, but three criteria make the decision almost mechanical.
| Criterion | Single campaign | Separate campaigns |
|---|---|---|
| Regional economics | CPA varies by less than 30% | CPA varies by 1.5–2× or more |
| Budget | Shared pool, no regional priorities | You must guarantee spend per market |
| Conversion volume | Under 30–50 conversions per region per month | Each region reaches learning volume on its own |
| Creative and landing pages | Identical everywhere | Different addresses, phone numbers, languages, delivery terms |
The strongest argument against splitting is data. A strategy fed ten conversions a month learns slowly and badly; five such campaigns are worse than one with fifty. The 30–50 conversions per campaign per month figure is a working guide, not a rule — in high-ticket verticals you will run on less, but then you should lean on micro-conversions. Choosing them without misleading the algorithm is covered in micro-conversions and funnel optimisation.
The usual compromise
Most accounts land on: major markets as dedicated campaigns, and the long tail of small regions grouped into one. Your volume gets granular control while the tail stops fragmenting the data. Revisit the split quarterly — a region that was tail last quarter may have earned its own campaign.
Account-level exclusions
If there are countries you will never sell to, exclude them once at account level rather than campaign by campaign. That way a new campaign built from a template by a junior does not quietly open you up to half the planet. Verifying that exclusions survived a campaign copy is fastest through a bulk settings export.
Why this scales with your budget
Geography is one of those settings where the cost of the error grows with spend and stays invisible at small volumes. At $2,000/month, 15% leakage is $300 — easy to write off as noise. At $200,000 it is $30,000 and a completely distorted read on channel performance.
Location checks belong at the top of any account audit, next to conversion setup and negative keywords. The full sequence is in the Google Ads account audit checklist. If you want a second pair of eyes on structure and settings, Google Ads training built around your own account is the practical route; for teams running several markets in parallel, Google Ads agency accounts come with billing already configured per geography.
FAQ
What is the difference between presence and presence or interest?
Presence serves ads only to people physically in, or regularly in, your targeted area. Presence or interest adds everyone showing interest in that area from anywhere.
Which setting is the default?
Presence or interest. If you never changed it manually, you are running the wider mode.
Where do I change it?
Campaigns → Settings → Locations → expand Location options → select Presence under targeting. The setting is per campaign.
Can I target interest only, without presence?
No. The standalone search-interest option is no longer available for Search, Shopping and Display campaigns. Two modes remain: presence, and presence or interest.
How do exclusions work now?
Exclusions operate on presence: ads are withheld from people located in the excluded area. Excluding by interest is not available.
How do I see where impressions actually came from?
The geographic report, viewed by user locations. The locations-of-interest view is not fit for this purpose — it shows intent, not reality.
Why do I still see impressions from countries I excluded?
Location inference is probabilistic. VPNs, corporate networks, roaming and carrier NAT all cause misplacement. A small percentage is unavoidable; a double-digit percentage points to a settings problem, not an inference problem.
Does geo targeting affect automated bidding?
Directly. Conversions from areas you do not serve still enter the training data and push the algorithm toward similar traffic. It is one of the most common causes of the “great CPA, poor revenue” gap.
Should I split campaigns by region?
Yes, when regional economics differ meaningfully or you need separate budget control. The trade-off is data fragmentation — very small campaigns learn slower and optimise worse.
Does this setting exist in Performance Max?
Location options are set at PMax campaign level just like in Search. Geographic reporting there is thinner, which makes periodic checks more important, not less.
How fast does a geography cleanup show results?
Typically within 7–14 days: out-of-market spend share drops immediately, while cost per conversion improves as bidding re-learns. Treat that window as a guide — long sales cycles stretch it.
Volume fell after switching to presence. Now what?
Find the ceiling first: lost impression share to budget means money, lost share to rank means relevance or bids. Only revert to the wider mode if you genuinely serve customers outside the region.
One campaign-level setting, two minutes to change, and somewhere between 5% and 20% of budget stops flowing to places where you sell nothing. Start with the user locations report — it tells you whether you have this problem before you change anything at all.
