Facebook Ads for Loan Officers Under Meta's Housing Restrictions

What Facebook Ads for Loan Officers Lost, and Why It Was Never Optional

Facebook ads for loan officers run inside Meta's special ad category, which means the account is stripped of nearly every audience control a normal advertiser uses before the first dollar is spent. The category has to be declared at campaign creation. A broker who quietly skips the declaration and promotes a loan product anyway is not being clever; ads get disabled on detection, and repeated failures put the whole ad account at risk of restriction, which is a far more expensive problem than a rejected creative.

The specifics are worth stating plainly, because brokers arriving from an agency background routinely assume one or two of these survived. Age targeting is gone and every ad serves to the full adult range. Gender targeting is gone. ZIP-code targeting is gone, and the location radius cannot be drawn tighter than fifteen miles, which for a broker licensed in one metro means the campaign will pay to reach a large amount of geography that is technically in-market but practically irrelevant. Detailed targeting is reduced to a scrubbed list with anything correlated to a protected class removed, so the homeowner-status, likely-to-move, and income-proxy segments that once made this channel trivial are simply not there. Special Ad Audiences, the constrained lookalike substitute Meta offered for years, were retired outright. There is no replacement waiting behind a support ticket.

None of this is a platform quirk. Housing and credit are the two categories where targeted advertising has an actual civil-rights history, and the restrictions exist because ad delivery in these verticals has been shown to reproduce exclusion patterns the Fair Housing Act was written to end. Treating the rules as an obstacle to route around is how a marketing decision becomes a legal one.

Creative Is the Only Targeting Layer Left

With the audience controls removed, the only mechanism left for reaching a specific kind of borrower is the ad itself, and the practical shift is that the ad set is built broad on purpose while the creative does the qualifying. A campaign aimed at self-employed applicants who have already been declined once by a bank does not need an audience segment; it needs a first line that names that experience so precisely that nobody else stops scrolling. That is not a workaround for lost targeting so much as an admission that the targeting was always a lazier substitute for saying something specific.

The second half of the problem is what Meta's delivery system is being optimized toward. The algorithm needs roughly fifty optimization events per ad set per week to exit the learning phase and deliver stably. A broker who funds eight loans in a month cannot feed the system on funded loans; there simply is not enough signal. So the campaign optimizes to an upper-funnel event that clears the volume threshold, usually a lead submission, and then quality is controlled elsewhere rather than by the bidding algorithm. Running Meta ads for mortgage brokers means accepting that the platform will optimize toward the event you name, will find the cheapest possible version of that event, and will do so with real enthusiasm. Naming the wrong event is how an account produces two hundred inquiries a month and no closings. Creative is the only targeting left after the housing category restrictions, which reorders the best mortgage marketing strategies on Meta.

Instant Forms Versus a Landing Page, and the Arithmetic Behind the Choice

Meta's instant forms are the default recommendation for a reason: they are cheap, they pre-fill from the profile, and they will produce inquiry volume immediately. They also produce the weakest intent available anywhere in paid media, because a borrower who submitted a pre-filled form in two taps has invested nothing and frequently does not remember doing it by the time the loan officer calls. Mortgage lead ads built this way convert at a rate that flatters the dashboard and disappoints the pipeline.

The economics decide the argument. A broker fee of 1 to 2 percent of the loan amount means a $250,000 file returns somewhere between $2,500 and $5,000. Cold social traffic converts to a funded loan at a far worse ratio than search does, because nobody on Facebook woke up intending to get a mortgage. If sixty inquiries produce one closing, a $30 instant-form lead has cost $1,800 of media to reach a file worth $2,500 at the low end of the fee range, and the campaign is underwater before anyone has paid for a credit report. The two levers that change this are friction and follow-up. Switching the form to the higher-intent setting, adding qualifying questions about loan purpose, timeline, and property type, and reviewing the answers before the call all cut volume and lift the ratio. The questions themselves have to stay inside fair-lending limits — asking about family status, national origin, or anything adjacent is out of bounds even when the intent is innocent. The alternative is sending the click to a real page and letting the page do the qualifying, which is why the design of that page carries so much of the load in landing pages that carry the rate disclosure.

Refinance Is the Campaign This Channel Exists For

When rates fall, search demand takes several days to build and arrives inside an auction where every lender in the country is bidding on the same handful of refi queries. Meta requires no such queue. A refinance campaign can be built and live in an afternoon and it reaches homeowners who have not yet had the thought, which is the entire proposition: this channel creates the demand rather than capturing it, and in a refi window measured in weeks, being first matters more than being cheapest.

The restriction makes this harder than it sounds. Homeowner status cannot be targeted, so the campaign cannot ask Meta to find people with mortgages. What it can use is first-party data — a customer list of past borrowers uploaded as a custom audience, site visitors, and people who engaged with previous video — plus broad geography and creative that self-selects. A line addressed to anyone who closed a loan in a high-rate stretch will be ignored by everyone it does not apply to and will stop the person it does. What that creative cannot do is name a rate without carrying the annual percentage rate at equal prominence, which is precisely why so much refi copy stops at the break-even question rather than the number. Speed also has a compliance cost: a campaign assembled in an afternoon still has to go through review, and creative built to be refreshed weekly needs a disclosure template that does not have to be re-argued each time.

The Purchase Side, Pre-Approval, and the RESPA Problem With Co-Marketing

On the purchase side the ask is not a loan. It is a pre-approval, ninety days ahead of anything funding, from a first-time buyer whose main question is what they actually qualify for. Ads that lead with a program rather than a rate work best here, and the copy stays deliberately unspecific about numbers because naming a down payment percentage or a monthly payment figure turns an ordinary ad into a triggering-term problem that drags the full disclosure block onto a square image with no room for it.

The purchase side also carries a trap that has nothing to do with Meta and everything to do with mortgage. Loan officers routinely co-market with real estate agents, splitting the cost of an ad that features both. Under RESPA Section 8, each party has to pay a share proportional to the value they actually receive from the ad. A loan officer who covers eighty percent of the cost of a campaign that is mostly an agent's listing has not bought advertising; they have paid for referrals, and that is a kickback whatever the invoice says. Co-marketed mortgage lead ads need the split documented against the actual prominence of each brand, and the loan officer needs to be able to defend that split later. Brokers who build their purchase pipeline through agent relationships should read this alongside the way property managers advertise for owner accounts rather than tenants, where the same referral-economics question shows up in a different regulatory frame.

Measuring a Channel That Cannot See the Closing

Meta's default attribution window is seven days after a click and one day after a view. A mortgage decision takes considerably longer than seven days, and the borrower who saw an Instagram video in March, searched the broker's name in April, and funded in June will be reported by Meta as nothing at all. Every honest assessment of Meta ads for mortgage brokers has to start by accepting that the platform will systematically under-report its own contribution while simultaneously over-reporting cheap form fills.

Two fixes do most of the work. The Conversions API sends server-side events with the lead identifier attached, which restores signal that browser-side tracking loses to consent prompts and iOS restrictions. Offline conversion uploads then push the downstream stages — application started, application submitted, loan funded — back into Meta weeks later, so the bidding algorithm eventually learns which inquiries were real. The reporting discipline is to judge the channel by cohort: take the leads generated in a given month, follow them for ninety days, and count the funded loans against the media spent to acquire them. A borrower who arrives cold from a video and needs six weeks of contact before booking a call is not a failed lead, and keeping that borrower engaged for those six weeks is the job of the longer-form video that explains a loan program properly once the fifteen-second social asset has done its work.