Targeting Everyone Who Walked Into Your Competitor Sounds Great. Most Stores Are Too Small for It.

The email is short and confident. Someone you have never met says they can put your Meta ads in front of everyone who walked into your biggest competitor last month. Not people who live near it. Not people interested in what it sells. The actual people who went through the door. They will build you a sample audience for free, so you can see it for yourself.

It sounds like a superpower, and it is not a scam. The technology is real, it is sold openly, and Meta will accept the audience it produces. The questions worth asking are the ones the email skips: where the list comes from, how many people it actually finds for a business your size, what it costs once the free sample is over, and whether you are allowed to use it at all.

We have written before about geofencing as it is usually sold, bundled into streaming TV and programmatic packages, in Someone Will Pitch You Streaming TV and Geofencing Ads. This pitch is the newer version: the same location data, delivered as a Facebook and Instagram audience. We looked into all four questions, because it is the kind of thing we would want to offer if it held up. Here is what we found, and why, for most small businesses, the targeting already sitting in Ads Manager does most of the job for free.

Data Brokers Sell Your Competitor's Visitors by the Phone

Every audience like this starts in somebody else's app. A weather app, a coupon app, a game: any app that asked for your location and got a yes can record where your phone goes, and many of them pass that data to companies called data brokers. The brokers combine feeds from many apps into one large panel of phones and the places those phones have been.

Each phone in that panel carries an advertising ID: Google's advertising ID for Android phones, and Apple's identifier for advertisers for iPhones. Advertising systems have used these for years to recognize a device without a name attached.

From there, it takes four steps:

1. Draw a boundary. The vendor traces the outline of a building on a map and picks a date range.
2. Pull the phones. Every advertising ID seen inside the outline during those dates goes on a list. Better vendors filter by how long the phone stayed and how often it came back, so a driver cutting through the parking lot does not count as a customer.
3. Upload the list to Meta. Meta's customer list audiences accept a column called mobile advertiser ID, and Meta's formatting guide says it takes both the Android and the Apple versions.
4. Let Meta match it. Meta checks each ID against the phones its own apps have seen and builds an audience from the matches. You never see who is in it.

Meta asks one question on the way in. When you upload a customer list, you declare whether the data was collected directly from your customers, provided by partners, or both. A list bought from a data broker is partner provided.

Meta's Tightest Radius Is Two-Thirds of a Mile. That's a Whole Shopping Center, Not One Store.

Before paying for any of this, be clear about what Meta already gives you, because it is more than most owners use.

Meta lets you target by ZIP code, by city, or by a pin dropped on an address with a circle drawn around it. Its help pages describe the reach as people who live in, have recently spent time in, or often visit the places you choose. Meta's developer documentation puts the smallest circle around a pin at 0.63 miles, or one kilometer.

A circle that size covers about 1.25 square miles. Drop it on a suburban strip mall and it takes in the mall, the neighborhoods behind it, the high school across the road and everyone who drove past on the way to work. For a business that serves its neighbors, that is often exactly right. For a business that wants the 400 people who walked into one particular gym, it is a blunt tool.

So the honest answer to "what does a location-data audience do that radius targeting can't?" comes down to two things:

• It knows who went inside. Radius targeting knows who was nearby. A list of phones seen inside a competitor's building is a list of people who already buy what you sell. That is a stronger signal than proximity, and it is the entire pitch.
• It follows people home. Radius targeting stops reaching people once they leave the area. A list of advertising IDs does not. The 20,000 people at a trade show go home to 40 different states, and the list still reaches them there.

Everything else these vendors promise, Meta already does. We covered the settings that quietly widen your radius, and why Meta no longer separates a resident from a recent visitor, in Meta Checked a Box for You. It Sends Your Ad Out of Town. If your radius is leaking, fix that first. It is free.

One Store Doesn't Have Enough Phones to Fill a Meta Audience

Here is the part the free sample is designed to skip. Meta has to match enough of a list to real accounts before it can deliver ads well, and its guidance is plain: "We recommend that your custom audience includes at least 1,000 customers." A single location rarely gets close.

Nobody publishes the end-to-end numbers, so we built an estimate from the pieces that are published. Treat it as an illustration of scale, not a measurement:

• How much of the population a broker sees. A peer-reviewed study in PLOS ONE compared one major location panel with census counts from 2018 to 2022 and found it covered about 7.5% of the population on average, ranging from 4.5% to 14.5%. Most of the people who walk into any store are simply not in the data.
• How many of those phones carry a usable ID. Android phones generally do. iPhones share theirs only when the owner taps Allow on Apple's tracking prompt, and Adjust, a mobile measurement firm, puts that opt-in rate at about 35% of users shown the prompt. With iPhones at a little over half of US mobile web traffic, roughly two-thirds of the panel ends up usable.
• How many of those Meta can match. Meta does not publish a match rate for advertising IDs, and we could not find a credible independent figure. We assumed half. On iPhones the real number is likely lower, because Meta can only match an ID it has seen in its own apps, which takes a second yes from the same person.

Run a store with 1,500 visits a month through that:

• Visits in a month: 1,500
• In the broker's panel (7.5%): about 113
• Carrying a usable advertising ID: about 72
• Matched to a Meta account: about 36

Thirty-six people. And 1,500 visits is not 1,500 people, because regulars come back every week, so the number of different people is lower still. Even a full year of one small store lands at a few hundred matched accounts, short of the 1,000 Meta recommends.

A small audience costs more than its size suggests. Meta spreads your budget across too few people, so the same handful see your ad again and again, and you pay more for every thousand impressions to reach them. The ad wears out fast, and so do they.

Scale changes the picture. The same math on an event with 20,000 attendees gives roughly 480 matched people, and anywhere from about 300 to about 900 depending on how much of the crowd the broker's panel sees. Run it across every location of a competitor chain, or across 12 months instead of one, and you can clear the bar. That is the shape of where this works.

Location Data Is Priced for Big Contracts. The Free Sample Is the Only Cheap Part.

Here is what the data costs after the sample, from the prices vendors have actually published. Most publish nothing.

• DRAKO. Advertising IDs seen at an address or a drawn boundary, with three to five years of history (its listing says both), exported for you to upload. Listed on the Datarade data marketplace as "starts at $2 / CPM." The listing does not say what the thousand refers to, so ask.
• Irys. Raw location records with advertising IDs, which you filter to your boundary yourself. $5,000 per one-time purchase on Datarade.
• OnSpot Data. Location audiences with a direct push into Meta, so you never handle the file. Quote only.
• Intuizi. An ongoing US feed of raw location data with device IDs. $10,000 a month on Datarade, the lowest published recurring price we found for US data.
• Veraset. One of the largest panels, with a self-serve tool that advertises exportable device IDs. No public price. Vendr, which tracks software contracts, reports a median of $165,000 a year; SafeGraph, a competitor, says pricing starts at about $40,000 a year.

Two things stand out. First, nothing here is self-serve the way a Meta campaign is. You are buying one-off pulls, buying through a managed service, or signing a contract measured in tens of thousands of dollars a year.

Second, it explains how a tool can offer this for a small monthly fee. A company signs one large data license, then sells slices of it to many small businesses, often with a percentage of ad spend on top. That can be a legitimate business. It also means your audience is only as good as the panel they licensed, and you will rarely be told which one it is.

Five States Ban Selling Location Data. Meta Still Makes You Swear Your List Is Legal.

This is where most pitches go quiet, and it matters more than the price.

Meta puts the legal promise on you. Before you upload any customer list, you accept Meta's Customer List Custom Audiences Terms. The first clause has you promise that you have "all necessary rights and permissions and a lawful basis" to use the data. The second says your audience criteria will not be "based on" health information, financial information, or "other categories of sensitive information (including any information defined as sensitive under applicable laws...)." If an agency uploads for you, it binds you to the same terms. Meta does not verify the broker's data when you upload it. You are telling Meta it was legal, and Meta reserves the right to audit that promise later.

The FTC keeps closing down the supply. Since 2024 the Federal Trade Commission has finalized orders against five location data brokers. One of them, InMarket, may no longer sell precise location data at all. The most recent, against Kochava and the subsidiary that now runs its data business, was announced in May 2026 and entered by the court in June. It bars selling sensitive location data unless the person gives express consent and the data is used for a service they asked for. Some orders also reach backward: a broker can be required to tell past customers that data they already bought should be deleted, de-identified or made non-sensitive.

Five states now ban the sale outright. Oregon, Maryland, Virginia, Connecticut and New Jersey prohibit selling precise geolocation data, and in none of them does the person's consent make the sale legal. California treats location within 1,850 feet as sensitive personal information that people can tell a business to limit how it uses, and Texas requires consent before it is processed at all.

Health care is off the table. A list built from phones that visited a chiropractor, a dentist or a physical therapist is an audience based on health information, which Meta's terms rule out on their own. Several states add their own bans:

• New York bans geofencing any health care facility other than your own to deliver any digital ad at all, within a boundary of 1,850 feet or less.
• Washington bans geofences within 2,000 feet of any in-person health care provider when they are used to track people seeking care, collect health data, or send ads related to their health or health care, with no consent exception.
• Nevada does the same at 1,750 feet.
• Connecticut bans geofences within 1,750 feet of mental, reproductive or sexual health facilities used to track people or send them messages about their health data.

New York's version reaches businesses that are not in health care at all. Its ban covers any digital ad, so a tight boundary drawn around a competitor's store in a building that also holds an urgent care could put you on the wrong side of it.

A common assumption is that HIPAA is what stops health businesses from doing this. It is not. HIPAA covers health care providers, insurers and the companies they hire, not a data broker reselling app data. The barriers are Meta's terms, the FTC and state law, and they are enough.

None of this is legal advice, and these laws are changing quickly. If you are seriously considering location data for your business, a short conversation with a privacy attorney in your state is cheaper than finding out the hard way.

Events and Competitor Chains Are Where Geofencing Pays

None of this makes location data useless. It works for a narrow set of businesses, and they share two traits: a big enough pool of phones, and a customer worth enough to justify the cost.

Where it can earn its keep:

• Trade shows, conferences and large events. Thousands of qualified people in one place for a few days, who scatter afterward. This is the strongest case on Meta, because radius targeting cannot follow them home and a list can.
• Competitor chains, not single stores. Every location of a regional furniture chain or a gym brand over six to twelve months can produce an audience Meta can actually deliver to.
• High-value purchases. Car dealers, furniture and mattress stores, home improvement showrooms, pool and spa dealers. When one customer is worth thousands, a small, expensive audience can still pay.
• Your own locations. If you run several stores, an audience of your own past visitors, checked against your sales, can tell you something your pixel cannot.

Where to skip it:

• Local service businesses. A plumber or a roofer wins by being there when the need appears. Radius targeting plus a sharp ad reaches the neighborhood, and Meta's broad delivery finds the people who are ready.
• A single small location. The math above is the reason.
• Health care, and anything in a special ad category. Health is ruled out by Meta's terms and the state laws above. Housing, employment and financial ads already run under Meta's special ad categories, which strip out most targeting, so building a precise audience for them works against the rules you are already under.

A Broker Who Won't Give You a Phone Count Doesn't Have the Phones

If your business fits the first list, test it properly before you commit money. The test is cheap. Skipping it is not.

1. Get the count before you pay. Ask how many unique phones the vendor's data shows inside your boundary over your date range, before anything is purchased. A serious vendor can run that count in minutes. If they will only show a finished audience size after you sign up, assume the raw number is small.
2. Set a floor. With half or fewer of those phones likely to match on Meta, a count under about 2,000 is unlikely to produce an audience Meta can run well. That threshold is our rule of thumb, not a Meta figure, but it follows from Meta's 1,000-person guidance.
3. Ask where the data comes from. Which apps, collected with what consent, and whether the vendor removes sensitive places such as clinics, schools and places of worship. Recent FTC orders require several brokers to verify consent from their own suppliers, so a vendor that cannot explain theirs is a risk you are about to take on.
4. Ask what you actually receive. A raw file of advertising IDs you upload yourself, or an audience pushed straight into your ad account. A direct push is simpler; a file lets you see what you bought.
5. Run it against a control. Put the location audience in one ad set and a normal radius audience in another, with the same ad and budget. If the location audience does not beat the radius on cost per customer, it is not earning its price.
6. Judge on customers, not clicks. Meta will happily report cheap clicks from a small, curious audience. Count the sales in your own records.

One more comparison is worth making before you buy anyone's list. Meta matches a list on whatever identifiers it carries, and your own customer list, with emails and phone numbers, gives it far more to work with than a column of bare device IDs from strangers' phones. If you have never uploaded your customers, start there.

Radius Targeting and a Sharp Ad Still Do Most of the Work

The pitch is clever because it sells precision, and precision feels like control. But on Meta in 2026, most of the targeting happens inside the ad itself. Meta reads your image, your words and who responds, then goes looking for more people like them. We made that case in full in You Think You Narrowed Your Audience. Meta Calls It a Suggestion. A clear offer shown to a well-set radius will find your competitor's customers on its own, and it will find the ones who are ready to switch.

So when that email arrives, you do not need to argue with it. Ask for the phone count, ask where the data comes from, and ask which states they will not sell in. The answers will tell you whether you are looking at a real audience or a free sample built to impress.

This is the kind of question we work through every day for the businesses whose Facebook and Instagram ads we run, and we are happy to help if you would rather hand it off. If you would rather keep running your own account, the six questions above work on any vendor, and they will save you a few thousand dollars on the ones that cannot answer them.