Geo-Targeting in Google and Meta Ads: What Each Setting Actually Does

One Google setting decides whether you pay for people in your city or people thinking about it.

Every platform claim below was read off Google Ads Help and Meta's Marketing API reference on 23 September 2026 and linked in place. Labels move. Check them the day you read this.

This page is part of our performance marketing guide and it stays inside one lane: the paid-ads location settings in Google Ads and Meta. Locality-level marketing strategy, the decision about which areas are worth serving at all, belongs to our hyperlocal marketing guide. Organic "near me" search is near-me SEO, and the Maps and citations groundwork is local SEO for small business. None of those settings live in an ad account, so none of them are repeated here.

The running example is a hypothetical single-studio pilates brand in Powai, Mumbai, on ₹60,000 a month. Invented, not a client.

What geo-targeting controls in a paid account

Geo-targeting controls eligibility: whether an auction you could have entered is one you are allowed to enter. It says nothing about whether the person can reach you. Two people inside the same 3 km circle can be a 6-minute walk and a 40-minute drive from your door.

• It does control which locations can trigger delivery, which locations are blocked outright, and what your reports let you compare afterwards.
• It does not control whether your Maps listing ranks, whether your creative speaks to that area, or how far someone will realistically travel.

Google Ads: the two location options that decide who sees your ad

Google Ads has exactly two target options, and picking the wrong one costs more than any bid change on a local campaign. Google's wording, from its About advanced location options page (read 23 Sep 2026): the default and recommended option is "Reach people in, regularly in, or who've shown interest in your targeted locations (Presence or Interest)", and the alternative is "Reach people in or regularly in your targeted locations (Presence)".

The exclude side has no equivalent choice. The same page calls it "Location exclusion (Presence)": it removes people likely to be situated in the area you named.

Interest is not a mystery box. Google says a person shows it "through terms used in their searches, if they were recently in a location, or through content they view related to the location". A Delhi user typing "pilates studio Powai" counts as interested, and on the default setting your studio pays for that click.

Our view: if the business has a physical catchment, switch to Presence on day one. For travel, hotels, education and real estate, leave Presence or Interest on, because the customer is buying a place they are not standing in. Google's case for the broader setting is internal data from 21 May to 1 June 2022: advertisers who switched in Travel, Real Estate and Education saw 5% more conversions. Read the verticals in that sentence. Yours may not be one of them.

Radius targeting: 1 km is the floor, and city geometry ignores it

Google requires "a radius of at least 1 km around any given location" (Target ads to geographic locations, read 23 Sep 2026). Meta's pin runs wider: "from 0.63 to 50 miles, or 1 to 80 kilometers" (Marketing API basic targeting, v25.0, read 23 Sep 2026).

A circle is a blunt instrument in an Indian city. Draw 5 km around Powai and you pull in areas cut off by the Mithi river, a railway line and a tolled expressway. Draw 5 km around a South Mumbai address and a third of it is sea.

Think in travel time, then draw the shape that approximates it. The catchment logic for gyms and fitness studios lands at roughly 3 to 5 km, because nobody commutes far for a daily session. Salons and spas stretch further, because the visit is monthly. Same map, different circle.

For a single location, start with named localities rather than one radius. Powai, Chandivali, Vikhroli West and Kanjurmarg beats a 5 km circle, and it reports back in names your client recognises.

Location exclusions do more work than inclusions

Exclusions win every conflict. Google states it plainly: "If an area is both included and excluded in your location targeting settings, exclusion takes priority" (Exclude ads from geographic locations, read 23 Sep 2026). It is the only location setting that behaves absolutely, so it should carry more of the work.

Two limits worth knowing from that page: you can bulk-add up to 1,000 locations at a time, and since 31 March 2024 campaigns targeting all countries and territories are capped at 122 country-level exclusions.

Meta's exclusion side has a rule people miss. In the Marketing API reference the recent location type, drawn "from mobile device data", is marked "Not available to exclude locations". Meta exclusions run on stated home location, so you are excluding residents, not visitors.

Location bid adjustments: check the bid strategy before you reach for them

The classic local lever was a location bid adjustment: +20% on the suburb that converts, -30% on the one that does not. On most modern campaigns that lever is no longer connected to anything.

Google's About bid adjustments page (read 23 Sep 2026) lists location adjustments as unsupported with Target CPA, Target ROAS, Maximize conversions and Maximize conversion value. They apply with Maximize clicks. Google's reason: those strategies "automatically set bids to optimize for the conversion goal".

So on a tCPA campaign, the slider is not the tool. Split that geography into its own campaign with its own budget and target, a structural call covered in our account structure guide and the bidding strategy guide. Splitting thins the conversion data in each campaign, so only do it where volume supports it.

Meta's location options in 2026: home, recent, and what Advantage+ will not touch

Meta's Marketing API reference (v25.0) lists two location types. home is "People whose stated location in their Facebook profile 'current city' is in an area. Facebook validates this by IP and information from their friends' profile locations." recent is "determined from mobile device data." Pass nothing and it defaults to both, which the reference describes as the option available.

Read those definitions twice. A profile "current city" someone set in 2016 and never updated is still a targeting signal today. That is a different kind of data from Google's live signals, which is why the same 3 km pin builds different audiences on the two platforms.

Honest gap: Meta's Business Help Centre blocks automated fetching, so we could not verify this run how those two API values are labelled inside your Ads Manager dropdown, or whether a separate travel option still appears there. Open the dropdown and read it before you promise a client a specific option.

One thing you can rely on: Advantage+ audience does not widen your map. Meta's Advantage+ audience documentation (read 23 Sep 2026) says "Non-negotiable business constraints are NOT expanded, these include location constraints, minimum age, language and custom audience exclusions". Interests get relaxed. Geography does not. More in our Meta Ads guide.

Google vs Meta on the same five jobs

The same job on both platforms, in each platform's own words. This is the table we open when a client asks why a suburb performs on one channel and vanishes on the other.

Sources: Google Ads Help pages linked above and Meta's Marketing API reference, all read 23 September 2026. Meta's Ads Manager labels were not verifiable this run; the API field names are used instead.

Why the location data is never exact

Google reports geography two ways, and they answer different questions. Targeted locations "shows your performance by the locations you targeted". Matched locations "shows your performance based on the locations where your ads appeared", which Google says "could be users' physical locations or locations of interest" (About measuring geographic performance, read 23 Sep 2026).

The same page explains the row nobody likes. Performance lands in "Unspecified" when "We can't determine the location from the IP address, and the search query didn't indicate interest in a recognized location."

Google calls the whole system "Google's best effort to serve ads to users who meet your location settings", built on "a variety of signals, including users' settings, devices, and behavior". A 1 km radius is a strong preference, not a fence.

Three India traps to check before you raise a budget

These three account for most of the location questions we get from Indian local accounts, and two of them are not documented anywhere obvious.

• PIN codes. Google's Location target types by country page (read 23 Sep 2026) says target types differ by country, and its India examples are state and union territory. Our observation, not a documented limit: in the Indian accounts we build, the smallest dependable unit below a city is a radius. Search your own account's location box before promising a client pin-code targeting.
• Tier-2 and tier-3 spill. A city target covers the city as the platform draws it, which rarely matches how the business defines its service area. Pune and its townships, Ahmedabad and Gandhinagar, Lucknow and its outer ring: one target, leads from three places. Matched locations is where you catch it.
• The searcher in the wrong city. Presence or Interest is why someone in Gurugram researching "pilates studio Powai" is eligible for your ad. Waste for the studio. Possibly the best lead of the week for a hotel or a coaching institute. Decide which you are first.

The five settings we change first on a new local account

The order matters. The first two stop the leak, the next two make the reporting trustworthy, the last one keeps it that way.

1. Switch the target option to Presence on any business with a physical catchment. Leave Presence or Interest on for travel, hospitality, education and real estate, where interest in a place is the purchase.
2. Add exclusions before you touch inclusions. Exclusion overrides inclusion, so it is the only absolute control you have. Start with other countries, the metros you cannot serve, and any locality the sales team already knows is dead weight.
3. Replace one big city target with the named localities. Four or five suburb-level targets report back in language your client uses, and they let you kill one area without rebuilding the campaign.
4. Check the bid strategy before you plan any location bid adjustment. On tCPA, tROAS, Maximize conversions and Maximize conversion value, location adjustments are not supported. If a geography needs different economics, it needs its own campaign.
5. Put the Matched locations view on a weekly check. Not a quarterly audit. A weekly look at where ads actually served, against where you thought they served, is the whole discipline.

The one geo mistake we see most on Indian local accounts

Our view, from the accounts we audit: the costliest geo mistake is a single-location service business left on the default Presence or Interest setting, with an empty exclusion list and nobody reading Matched locations. The campaign looks healthy. Cost per lead looks fine. The leads will never walk in.

Run the arithmetic on your own numbers, not ours. Open Matched locations for last month, tag every row you cannot physically serve, total the cost. On the studio's ₹60,000 budget, a fifth of spend in unservable rows is ₹12,000 a month funding enquiries the front desk has to decline. The fraction is whatever your report says. The point is that almost nobody has looked. Our performance marketing team runs that pass on every account takeover, and the wider version is in our PPC audit guide.

Choosing between the two platforms for a local budget is a separate question, and our Google Ads vs Meta Ads comparison takes it on directly.