Instagram ads for real estate agents

Start with the consequence nobody draws. A housing ad set has its age range fixed to 18 through 65+ and its gender set to all — those controls are simply not yours any more. But the two apps do not have the same people on them, so choosing which one to buy is choosing an age distribution. Pew's November 2025 survey puts Instagram at 80% of adults aged 18 to 29. Facebook reaches 74% of 50- to 64-year-olds and 57% of the over-65s, where Instagram reaches 40% and 19%. That is a targeting decision wearing a media-planning costume, and after the special ad category has finished with your ad set it is close to the only one left.

Which tells you what an Instagram budget is for in this industry: buyers and renters. First-time buyers who think they need twenty percent down, renters who have never run the numbers, relocating professionals, accidental landlords in their thirties who inherited a house and a phone that will not stop ringing. The seller side — downsizers, estate sales, the owner of a house that outlasted the family in it, the expired listing whose owner is sixty-two — skews to exactly the cohort where this app loses more than half its reach. Run those on the Facebook twin. An agent who buys listing-acquisition ads on Instagram because that is where they personally spend their evenings is paying a reach premium to talk to people's children.

The shape argument is the second thing, and in this category it fights you. Meta's ads guide recommends 4:5 for feed and full-screen 9:16 for Stories and Reels, and those are the only two shapes worth exporting here: unlike the Facebook cell of this cluster the 1.91:1 export has nowhere useful to go, because our own publisher routes that ratio to Facebook's right-hand column and search results and to no Instagram position. (Instagram Feed's technical requirements do accept up to 1.91:1 — corrected 7 September 2026; the shape is allowed, the surfaces built for it are Facebook's.) That matters more than it sounds. Every agent has a folder of listing photography, all of it landscape, shot on a short lens backed into the corner of a room precisely to make the room read as a room. Crop it to portrait and you throw away the width that was the entire point of the lens. Instagram rules out the one asset you already own, and it is doing you a favour, because that asset is also the one every competitor in your market is posting.

The third difference is what you are allowed to do with the ad set's remaining scraps, and it is not nothing. Detailed targeting is not switched off — behaviour and demographic targeting are prohibited and interest exclusions are prohibited, but a previously approved list of interests remains selectable. Guides say either "you can still target interests" or "all targeting is gone", and both are wrong in a way that costs you: the honest instruction is to open the picker, see what your account offers under this category, and expect a short list.

Ads Manager offers six — Awareness, Traffic, Engagement, Leads, App promotion and Sales — and for a real estate agent under the housing category, one of them does almost all the work and one of them is a trap that the dentist guide in this cluster actively recommends. That contrast is the fastest way to understand what the category has done to you.

Leads is the default and should be nearly all of the budget. Eight of the ten ads on the examples page were written for it, two for Traffic, and none for Sales or Awareness — which is not a stylistic preference but the category speaking. Nobody buys a house from an Instagram ad. The most an ad here can cause is a conversation, and Leads is the objective that optimises toward one.

Awareness is the trap. On the dentist page it is a rational buy, because a practice can draw a three-mile circle and a household either drives past the building or does not. Here the smallest circle you may draw is fifteen miles, you may not exclude anywhere inside it, and you may not narrow it by age. That is a few hundred thousand people, the overwhelming majority of whom are not moving this year, and Awareness is the objective that asks you to pay to be seen by all of them. Same button, opposite verdict, and the reason is entirely in the ad set.

The offer list is where a real estate campaign is actually decided, and it is the part that has to be written for this app rather than for the industry. Everything below skews to the people Instagram reaches — buyers, renters, younger landlords — with the seller-side offers marked as belonging to the Facebook twin so you are not tempted to run all of them everywhere.

Each row pins four decisions: who the offer is really for, which angle family argues it best (each is a page in the angle library, with the sub-angles spelled out), which objective buys it, and, where the examples page has one, a finished ad that shows the whole thing assembled. Two of the eleven belong to adjacent businesses — property management and short-term-rental management — because the same person very often runs them, and because they are the two offers on this list that do not depend on somebody deciding to move.

One structural note. Every one of these ads has to open on a person rather than on a property, and the reason is in the targeting section: the ad set cannot qualify anyone, so the first line has to. That is why the angle recommendations here lean on families that name a situation — Problem & Pain, Objection & Risk Reversal, Comparison & Positioning — rather than on the ones that describe a product.

Any ad promoting the sale or rental of housing must run in the Housing special ad category — one of exactly four Meta accepts, alongside employment, financial products and services, and social issues, elections and politics. Declaring it is not a choice you get to weigh: Meta's classifiers will apply the category from the creative whether or not you ticked the box, and finding out that way means finding out after your ad set has been rebuilt underneath you.

What it removes is most of an ad set. Age is fixed to 18 through 65+ and gender to all. Lookalike audiences are unavailable. Saved audiences are unavailable. Postal codes, neighbourhoods, subcities, subneighbourhoods, metro areas, small geographic areas and electoral districts are all unsupported as location types, and any city, address or dropped pin carries a minimum radius of fifteen miles or twenty-five kilometres in the US and Canada, fifteen kilometres in Europe. Two restrictions here are almost never mentioned and both matter: you may not exclude a location either, so there is no carving the far side of the highway out of your circle; and behaviour and demographic targeting are prohibited along with interest exclusions, leaving only a previously approved list of interests selectable.

Customer-list audiences are the fiddly case and the one to check rather than assume. They are not simply gone — a list carries an eligibility flag, and a business admin has to review and accept Meta's non-discrimination policy before the account can run these ads at all. Website and engagement custom audiences are the interesting silence: Meta documents the lookalike and saved-audience removals explicitly and says much less about those, so the honest instruction is to build one and see whether the ad set accepts it, rather than trusting a guide. Including this one.

All of which means the ad set can no longer tell you who saw the ad, so the ad has to. Every ad on the examples page opens by naming a person — sellers whose listing expired, renters who think they cannot buy yet, the owner of a house that outlasted the family in it — and on any other page in this cluster that would be a note about craft. Here it is the last targeting instrument you own, and it is inside the picture.

The answer splits by side of the transaction, which is unusual — most industries have one right answer — and getting it backwards is the most expensive mistake available on this page.

Seller-side offers are a form. A valuation, an instant-offer comparison, a market report: the reader wants a number and you want a conversation, and there is no page that improves on asking. Use the Instant Form's higher-intent type, which adds a review step letting the person confirm what they entered, and add two or three qualifying questions — timeframe, address, whether they have already had an offer. You will get fewer leads and more of them will pick up the phone, which is the trade you want in a business where you personally follow up every one.

Buyer-side offers are a page, and specifically a page they can search. A buyer who taps an ad about first-time purchasing and lands on a form has been asked to hand over their phone number in exchange for nothing; the same buyer landing on live inventory with a filter has something to do for the next twenty minutes, and they register when they want to save a search. That is a better lead and it costs you nothing to collect. The exception is relocation, where the offer genuinely is a conversation rather than a listing feed.

For both, do not send anyone to a brokerage homepage. The ad made one specific promise — the twenty percent figure was wrong, your listing did not fail on price — and a homepage answers a different question. If there is no page for the offer, a form that matches the ad beats a beautiful site that does not.

The rule that changed most recently is the one worth leading on, because almost every competing guide is still wrong about it. On 24 April 2026 HUD issued a Dear Colleague letter (HUD No. 26-028) stating that sharing neighbourhood crime rates and school-quality information is permissible when that information is shared consistently and without discriminatory intent, and the National Association of REALTORS® updated its guidance to match. The blanket “never mention schools” advice that has circulated for a decade is no longer the rule. What has not changed is the thing underneath it: the requirement is consistency and objectivity — the same kind of neighbourhood information offered to every client regardless of their background, factual rather than editorial — and courts have found that racially-coded commentary about crime and schools can be evidence of discriminatory intent even where the data itself is accurate. So it is sayable. It is not sayable selectively, and it is not sayable with an opinion attached.

The second thing is the opening line, and it carries more weight here than on any other page in this cluster. Because the ad set cannot qualify anyone, the first sentence is your targeting: it has to make the right person recognise themselves within about a second and let everybody else scroll past without costing you a click. Open on the situation, not the house. “It wasn't the price. It was the photos.” finds a specific person. A wide-angle living room finds nobody.

Third, shoot for the shape rather than cropping into it. Instagram gives you 4:5 and 9:16 and nothing else, and your existing listing photography is landscape by design. Reshoot the two or three frames the ad actually needs vertically, or build the ad out of something that is not a room — a document, a calendar, a comparison table, a phone that is not ringing. Six of the ten ads on the examples page are not photographs of houses at all, which is not a stylistic accident.

Three stages, and the middle one is doing more work here than in any other cluster. Cold is a blunt instrument by regulation — a fifteen-mile circle you cannot narrow, cannot exclude from and cannot age-band — so its job is not efficiency, it is finding the people who react. Warm is where the ad set finally gets to be specific again, and it is where a housing campaign actually converts.

Which means the usual budget instinct is wrong here. In most industries you spend most of the money cold and retarget on the remainder; in this one, plan for the warm stage to earn its keep and resist the temptation to pour everything into a circle that cannot be aimed.

The arithmetic every advertiser should do is the learning-phase one, and in this category it comes out worse than anywhere else on this site. Meta's delivery system needs roughly 50 optimisation events per ad set per week to leave the learning phase; below that an ad set sits in “learning limited”, which is not a penalty but is an admission that the system cannot optimise with what it has been given. Multiply your own cost per lead by 50 and that is the weekly spend at which one ad set is working with full information. Housing leads are expensive precisely because the audience cannot be qualified before delivery, so that number is high — and it is the argument for one cold ad set with several ads in it rather than one ad set per offer.

Do not respond to that by widening. Widening is what the category already did to you, and the only lever left that improves cost per lead is the creative — specifically the opening line, which is where the qualifying happens. An agent who doubles the radius has bought more of the same problem; an agent who rewrites the first sentence has bought targeting.

Then leave it alone. Pausing an ad set, or changing its optimisation event, audience or creative, counts as a significant edit and restarts the learning phase, and budget or bid changes can too depending on their size. Daily adjustment feels like management and is the largest self-inflicted cost in a small account.

Seasonality is sharper in this industry than in most and it is regional, so plan against your own market's calendar rather than a national one. The listing season and the relocation season are the two moments where the same budget buys materially more, and the year-end lull is the right time to run the slow offers — the market report, the agent page — that compound rather than convert.

The first four are all the same mistake wearing different hats: planning a campaign around an ad set that the housing category will not let you build.