Facebook Ads for Real Estate Agents (2026): What You Can and Cannot Target

Almost every guide to Facebook ads for agents, including the version of this page I published in May, tells you to build a Special Ad Audience. It is described as the compliant replacement for a lookalike, the workaround that keeps you legal.

Meta deleted Special Ad Audiences. Creation was removed from the API on 15 September 2022, as a term of a settlement with the Department of Justice. The advice is four years stale and cannot be followed. I had it on this page and I have taken it off.

Housing ads are the most heavily restricted category on the platform. You cannot target by age, gender, ZIP code, income or homeowner status, you cannot use a radius under 15 miles, you cannot exclude locations, and you cannot build a lookalike. This page is what is actually left, what it costs, and which of the numbers you have read survive a check. I sell houses and run real estate coaching, so read the next box first.

The advice everyone still gives you was deleted in 2022

Search this topic and you will be told, repeatedly and confidently, that housing advertisers cannot use lookalike audiences but can use a Special Ad Audience instead. Build one at 1 to 3 percent from your database, the advice runs, and you have the compliant version of a lookalike.

There is no such thing any more. Meta removed Special Ad Audience creation from its Marketing API on 15 September 2022, with the release of version 15. It was retired as a term of the company's settlement with the Department of Justice, which is covered further down this page. So the current position on both tools is simple:

I am starting here because this page carried that advice until today. The May version of this article called the Special Ad Audience its secret weapon and told you to build one inside a 15 mile radius. It was wrong, it had been wrong since before I wrote it, and I did not check.

That is worth a moment, because it explains something about this whole category of content. The advice is not wrong because writers are careless. It is wrong because a rule changed once, quietly, inside a legal settlement, and the entire ecosystem of marketing blogs copies from itself. Nothing forces a correction. Four years later the dead product is still being recommended as the compliant option.

So check the date on anything you read about this, including this page, and check it against Meta's own developer documentation rather than against another blog.

Declaring it is your job, not Meta's

This is the error underneath most of the others, and it is a misunderstanding about who does what.

Meta's advertising standards put the obligation on the advertiser:

"Any United States advertiser or advertiser targeting the United States, Canada or certain parts of Europe that is running financial products and services, housing or employment ads, must self identify as a Special Ad Category."
Meta advertising standards, read 28 August 2026

Self identify. There is a toggle in the campaign setup and you are the one who has to switch it. Meta's documentation says that if you do not, there is a risk your ads will be paused until the campaign is adjusted.

Here is why that matters more than it sounds. An agent sets up a campaign, does not declare it, builds a three mile radius around a neighbourhood, adds an age band of 30 to 55, and the ad runs. Nothing is blocked, because nothing was declared. The agent concludes the housing restrictions must not apply to their kind of ad.

The restrictions did apply. The undeclared campaign is the violation, and the fact that it ran is not evidence that it was allowed. If you have been running property ads with tight ZIP targeting and an age band and wondering why other people complain about the rules, that is the likely explanation.

Which of your ads count

Three of Meta's categories touch what agents do. Housing is the obvious one. Employment is the one agents miss: if you are running ads to recruit agents to your team or brokerage, that is an employment ad and it carries its own declaration and its own restrictions. Financial products and services covers mortgage and insurance advertising.

Everything you cannot target on a housing ad

This is the complete current list, read from Meta's own Marketing API documentation on 28 August 2026 rather than from a marketing article.

Meta Marketing API documentation on special ad categories, read 28 August 2026. Meta changes these without announcement, so check the date on this table before relying on it.

Read the first three rows together and you can see what has actually been taken away. You cannot advertise a home to the neighbourhood it is in. The smallest audience you are permitted is a 15 mile circle, which in most metropolitan areas is several hundred thousand people and crosses school districts, price bands and municipalities.

That is the point of the rule rather than a side effect of it. Fine geographic targeting is how housing discrimination was historically done, and the restriction exists because ZIP code is a close proxy for race in most American cities. Whether you find that reasonable or not, it is the reason, and it is not going to be relaxed for you.

What survives is worth naming, because it is not nothing:

• Your own customer list, uploaded as a custom audience, if it passes the eligibility check. Your database is the single most valuable targeting asset you have left.
• Retargeting from your pixel, meaning people who already visited your site.
• Engagement audiences, meaning people who interacted with your page, your video or your form.
• An approved interest list, narrow and shrinking, plus the 15 mile geography.
• The creative itself, which is now doing most of the targeting work. More on that further down.

Check your own campaign

Two questions. What the ad is about, and what you were planning to target. It tells you whether you have to declare, and which of your plan survives. It runs in your browser and nothing is sent anywhere.

Two settings are worth trying. Switch it to recruiting and watch the category change to employment, which catches almost everyone. Then switch to the branding option and read the answer carefully, because it is the one case where I am telling you what I do not know rather than what I do.

The compliance tax nobody prices

Take targeting away from an ad platform and the ads get more expensive, because the system is spending your money on people who were easier to reach rather than people who were more likely to convert. Everyone in this business assumes that. Almost nobody has measured it.

One team has. A peer reviewed paper presented at the ACM Conference on Fairness, Accountability and Transparency in 2025, by researchers at Princeton and the University of Southern California, ran paired experiments on Meta's Variance Reduction System, the fairness mechanism the company built under its Department of Justice settlement. Their finding:

Imana, Shen, Heidemann and Korolova, ACM FAccT 2025, preprint arXiv:2506.16560, read 28 August 2026.

Now the limits, because a good study deserves accurate reporting. It is 36 paired experiments at $20 an ad, run over 24 hours, using a North Carolina voter file. It isolates the fairness delivery system specifically, not the full removal of targeting options that the Special Ad Category also imposes. So the honest reading is that roughly 12 percent is the measured cost of one component, and the total cost of housing compliance is larger than that by an unknown amount.

It is still the only number in this entire subject that comes from an independent, methodologically transparent source. Meta publishes nothing about what the Special Ad Category costs the advertisers it imposes it on.

The practical version for you: budget as though your housing ads will cost more per result than a comparable ad in an unrestricted category, and do not benchmark yourself against a general Facebook advertising figure, because you are not running a general Facebook ad.

There is no real cost per lead benchmark, and that is the finding

I set out to replace the cost table that used to sit on this page. I could not, because no credible one exists, and after a full pass through the sources I think saying that plainly is more useful than another invented range.

Here is what is actually out there.

All read 28 August 2026. Nobody anywhere separates real estate costs by buyer lead, seller lead, listing or brand campaign, which is the breakdown agents most want.

Every figure in circulation is published by a company that sells advertising, software or leads to real estate agents. That does not make any of them false. It does mean not one of them is research, and the most quoted number in the category may rest on two campaigns.

So what do you do instead? Benchmark against yourself. Your own cost per lead last quarter, in your own market, at your own budget, is the only number that describes your situation. If you want to compare Meta spend against other channels on a like for like basis, I built out the arithmetic in what real estate lead generation actually costs.

One figure I can give you with a sample attached. The National Association of Realtors 2025 Technology Survey found 75 percent of members use social media and 39 percent say it generates the highest number of quality leads, from 1,241 usable responses out of 49,233 invitations, a margin of about 2.8 points. Worth knowing, though, that in the same survey digital ad campaigns ranked fifth at 12 percent. Agents themselves report that organic social produces far more for them than paid does.

Three numbers going around that do not survive a check

All three of these were on this page until today. I traced each one.

"$1.36 average cost per click for real estate"

Traceable, and stale. It is WordStream's 2024 edition, where the real estate cost per click was $1.36 and cost per lead $13.87, from 2,946 campaigns between February 2023 and April 2024. WordStream's 2025 edition supersedes it at $1.57.

It keeps circulating because a real estate software vendor republished it, correctly citing WordStream but from the wrong edition, on a page still updated in August 2026. Everyone downstream copied that. If you see $1.36 presented as a current figure, you are reading a copy of a copy of a two year old number.

"Reels ads get a 70 percent engagement lift over static images"

Untraceable, and contradicted twice over.

Nobody publishes 70 percent. Socialinsider, analysing 35 million posts across 447,613 profiles for the full year 2025, found Reels at 0.52 percent engagement against static at 0.37, which is about 40 percent, and found carousels beating Reels. Buffer, across more than 52 million posts through December 2025, found Reels came last, losing to carousels by 109 percent and to single images by 34 percent on engagement, winning only on reach.

There is also a category error inside the claim. All of that data is organic. The claim is about ads. No paid Reels versus static comparison exists anywhere I could find. The number should not be adjusted, it should be deleted, which is what I have done.

Any tidy table of real estate cost per lead by campaign type

Covered in the section above. Nobody segments real estate Meta costs by buyer lead, seller lead, listing or brand campaign. If a page shows you that table, the numbers were assembled rather than measured. This page used to show you one.

The settlement that set these rules, stated accurately

The rules on this page exist because of one legal action, and it is worth stating precisely, because the distinction between what was alleged and what was agreed matters.

On 21 June 2022 the Department of Justice announced a settlement in United States v. Meta Platforms, filed in the Southern District of New York. The government alleged that Meta's advertising targeting and delivery systems violated the Fair Housing Act. That allegation was never adjudicated, and the settlement agreement records that Meta denies liability and any and all wrongdoing.

What Meta agreed to do:

• Stop using the Special Ad Audience tool for housing ads by 31 December 2022. This is why the product no longer exists.
• Build the Variance Reduction System, subject to Department of Justice approval and court oversight.
• Withdraw targeting options semantically related to classes protected under the Fair Housing Act.
• Pay a civil penalty of $115,054, the statutory maximum available under the Act.

Department of Justice press release and settlement agreement, and the January 2023 metrics agreement, all read 28 August 2026.

In January 2023 the parties agreed the compliance metrics: variance of no more than 10 percent on 91.7 percent of housing ads for sex, and on 81.0 percent for estimated race and ethnicity, verified by an independent third party. Court oversight was stated to run to 27 June 2026.

That date has passed. I could find no primary source stating what happened at the end of the term, and the Department of Justice case page shows no filing after October 2024. So I am not going to tell you the oversight lapsed and I am not going to tell you it continues, because I do not know.

What that does not change is anything practical for you. The Special Ad Category is Meta's own platform policy, published on Meta's own documentation, and it applies to your account whatever the status of a court's supervision of the company.

What shifted in 2026, and what did not

The federal fair housing ground has moved this year and most agent facing advice has not caught up. I am reporting this as changes to the record, not as a view on any of it.

• A Federal Register notice published 6 April 2026 withdrew eight Fair Housing and Equal Opportunity guidance documents, with a withdrawal date of 17 September 2025. One of them is the April 2024 guidance on applying the Fair Housing Act to advertising through digital platforms, which is the document that spoke most directly to what this article covers.
• The same notice withdrew the February 2021 memo implementing Executive Order 13988, which had directed enforcement of the Act's sex provision to cover sexual orientation and gender identity.
• A proposed rule published 14 January 2026 would rescind HUD's disparate impact regulations. Its comment period closed in February with 1,109 comments and it had not been finalised when I checked. A supplemental proposed rule followed on 10 August 2026, with comments closing 9 October.

Federal Register notices, read 28 August 2026.

There is a live disagreement you should know about rather than have resolved for you. HUD's current overview lists seven protected classes and does not include sexual orientation or gender identity. The National Association of Realtors still describes the protected class of sex as including gender identity and sexual orientation. Both statements are on the respective organisations' own pages right now. If you are a Realtor, you are bound by the Code of Ethics whatever a federal agency's guidance page says.

Your state has its own advertising rules on top

Meta's rules are not the only ones your ad has to satisfy. Every state real estate commission has advertising requirements, and a paid social post is an advertisement. Four examples, because the differences matter.

All read 28 August 2026. This is four states out of fifty one jurisdictions. Check your own.

Texas gives agents the single most useful concession I found. The broker name disclosure can be satisfied by a link rather than by cramming it into the post, provided the linked page carries the required information:

"An advertisement on social media will comply with the rule as long as the license holder has linked from the advertisement to the account profile page or a separate page, and that page contains the required information for an advertisement."
Texas Real Estate Commission, read 28 August 2026

Do not assume your state offers the same relief. Florida's rule is the opposite in spirit, requiring the brokerage name physically adjacent to the contact information in internet advertising.

Two practical habits cover most of this. Put your brokerage name in the ad creative itself rather than relying on your profile, and run your first ad past your broker before you spend anything. Their licence is on the line with yours, and in Virginia the rule explicitly puts your advertising under their supervision.

What actually works inside the constraints

All of the above is a list of things you cannot do. Here is what is left, and it is more than the restrictions make it sound.

The creative is now the targeting

This is the whole strategic shift and most agents have not made it. When you could target a 30 to 55 age band inside three ZIP codes, the audience settings did the qualifying. You cannot do that now, so the ad itself has to do it.

An ad that opens with "thinking about selling in Vienna this autumn" filters its own audience. People it does not describe scroll past, which costs you an impression and nothing else. Naming the town, the price band, the situation and the timeframe in the first line of copy does legitimately what ZIP targeting used to do illegitimately, and it is not restricted, because you are describing your service rather than selecting people by a protected characteristic.

Be careful with the line, though. Describing a property or a service is fine. Describing who should respond is where advertisers get into trouble. "Perfect for a young family" is a familial status problem, and "great for professionals" reads as a proxy for several things. Write about the home and the transaction, not about the buyer.

Your database is the asset now

Customer list custom audiences survive the Special Ad Category, subject to Meta's eligibility check. So does pixel retargeting and so do engagement audiences. Those are the only precise tools you have left, and they all depend on having a list and a website.

Which reorders the priorities. Spending on cold reach into a 15 mile circle with no interest targeting is the most expensive thing you can do on this platform. Spending to bring people to something worth their email address, then advertising to that list, is the cheap route. If you do not have anything worth an email address, that is the first problem to fix, and I went through the options in lead magnets that actually convert.

Send them somewhere built for it

A restricted, expensive click deserves better than your homepage. The single highest return change most agents can make is not in the ad account at all, it is what happens after the click, which I covered in landing pages that convert.

Consider whether this is the right channel at all

I would be doing you a disservice if I did not say this plainly. Agents in the National Association of Realtors survey rank digital ad campaigns fifth as a lead source, at 12 percent, well behind organic social. Housing ads carry a measurable compliance premium. The targeting that made this channel efficient for real estate has been removed on purpose and is not coming back.

Two adjacent channels are worth weighing before you commit a budget. Google ads for agents catch people already searching, which is a different and often better intent, and are not subject to Meta's category. And Facebook groups reach the same platform organically with no ad account and no category at all.

None of that means do not run Meta ads. It means run them knowing the channel got harder, budget for a higher cost per result than a general benchmark implies, and measure against your own numbers rather than against a table somebody assembled.

And if the honest answer is that you are reaching for paid ads because the rest of the business is not producing yet, spend the money there first. Building a database, a follow up habit and a listing presentation beats any ad account, and it is what I work on with agents through real estate coaching.

Frequently asked questions

Can real estate agents still use lookalike audiences on Facebook?

No. Meta's documentation states lookalike audiences are unavailable for housing, employment and financial products and services ads. The Special Ad Audience that replaced them between 2019 and 2022 no longer exists either, because creation was removed from the API on 15 September 2022 under Meta's settlement with the Department of Justice. There is no substitute product, and any guide recommending one is out of date.

What is the minimum radius for a real estate Facebook ad?

15 miles, or 25 kilometres, in the United States and Canada, measured from any city, address or dropped pin. It is 15 kilometres in Europe. ZIP code, postcode, neighbourhood, subcity and metro area targeting are all blocked for housing ads, and excluding a location is not supported at all.

Do I have to declare my ad as a Special Ad Category?

Yes, and it is your responsibility rather than Meta's. Meta's advertising standards require the advertiser to self identify, and its documentation warns that ads risk being paused until the campaign is adjusted if you do not. An undeclared housing ad that runs successfully has not been approved, it has simply not been caught.

Does a personal branding ad count as a housing ad?

This is genuinely unsettled. Meta's own definition is not readable on any page I could access, and the only authoritative definition available, from the Department of Justice settlement, is transaction specific: an ad about a specific opportunity to rent, lease, sell or buy a residential dwelling, or a related transaction such as a mortgage or appraisal. A listing ad is clearly inside it. A pure branding ad with no property mentioned is arguable. Declaring costs you targeting precision. Not declaring risks a fair housing problem, so I would declare.

Do ads recruiting agents to my team fall under these rules?

Yes, but under employment rather than housing. Employment is one of Meta's three special ad categories alongside housing and financial products and services, and it carries its own declaration and its own targeting restrictions. This is the one most agents miss.

What is the average cost per lead for real estate Facebook ads?

There is no credible answer. Every figure in circulation is published by a company that sells advertising, software or leads to agents, and none discloses a real estate specific sample size. The most quoted benchmark comes from a report whose methodology describes search campaigns rather than Facebook campaigns and requires only two active campaigns per subcategory. Benchmark against your own past results instead.

Do housing ads cost more than normal ads?

On the only independent evidence available, yes. A peer reviewed study presented at ACM FAccT in 2025 found that Meta's Variance Reduction System, applied to housing ads under the Department of Justice settlement, raised cost per person reached by 12.02 percent and cut reach by 9.82 percent. That measures one component rather than the full effect of the Special Ad Category, so the real premium is likely larger.

Educational content only. Not legal advice. Fair housing law and platform advertising policy are described here as they were published on the dates named, and neither the author nor Jamil Academy is a law firm. Consult your principal broker and your own counsel before running paid advertising. Jamil Academy is not affiliated with, endorsed by, or sponsored by Meta Platforms, Inc., the National Association of Realtors, or any company named in this article. Company and product names are trademarks of their respective owners and are used here only to identify what is being discussed. This page carries no affiliate links, no commissions and no sponsored placements, and the author earns nothing whichever way you decide. Platform rules, regulations, prices and published benchmarks were read from the sources named on 28 August 2026 and change without notice, frequently and without announcement. Claims described as alleged in litigation are allegations, not findings of fact, and unresolved matters are described as unresolved. Views expressed are the author's own opinion. If you believe anything here is inaccurate, tell us and we will correct it.