Updated
Meta ads work for ABA when you respond within minutes: they reach parents before they ever search, but that borrowed intent decays fast. Text within a minute, call within five, qualify area, payer, age and diagnosis in the first exchange, give undiagnosed families an evaluation path, and judge the channel on cost per started client, not cost per lead.
Meta is the best tool in the category for one specific job: reaching the parent who has not searched for anything yet. A mother who has been worried about her two-year-old for months, who has googled "speech delay" twice at midnight and never once typed the word "ABA," is unreachable on search. She is very reachable on Instagram.
That is an enormous audience, and it is cheap relative to search. It is also the reason Meta has a reputation for junk leads in ABA. Both things are true, and they have the same cause: you are interrupting someone rather than answering them. Everything below follows from that one fact.
Meta does three things no other ABA channel does well. It creates demand instead of harvesting it, reaching families before they know a service category exists. It scales in days rather than quarters, which makes it the practical answer when a new location opens with empty schedules. And it targets geography precisely, which matters enormously when your service area is four counties and not a state.
It is also the channel where the message does the most work. On search, the parent brings the intent and your ad just has to be there. On Meta, your creative has to produce the intent from scratch.
A search lead decided to contact you. A Meta lead was persuaded to, in about eight seconds, while doing something else. That difference does not make the lead worthless, but it does mean the interest is on loan and it is already evaporating when the form submits.
Practically, this shows up in three ways. Meta leads answer the phone at a lower rate than search leads. A larger share are outside your service area or on a payer you do not take, because they were not filtering for you the way a searcher was. And a much larger share have no diagnosis yet, because you reached them earlier in the journey — which is an opportunity, not a defect, as long as you have somewhere to send them.
The failure mode is a practice that runs Meta, judges it on the same yardstick as search, concludes the leads are garbage, and turns it off. Usually the leads were fine and the follow-up was six hours late.
There is a test that separates the two, and it is worth running before touching the ad account. Call within minutes for a couple of weeks. If families answer and most of them qualify, the targeting is fine and the old follow-up was the problem. Only when both halves fail — almost nobody responds to a five-minute call, and almost none of the few who do are in your area, on your payers, or in your age band — is the channel genuinely bringing the wrong families, and then the fix is geography, creative and audience rather than anything downstream.
Health-related interest targeting is restricted on Meta, and you should not plan around reaching "parents interested in autism." Nor should you want to: audiences built on sensitive inferences are both unreliable and a bad look for a healthcare provider.
What works is unglamorous. Tight geography around each clinic or service radius, not the whole state. Broad parent-aged demographics, and let the delivery system find responders. Lookalikes built from families who actually started care, not from everyone who filled a form. And exclusions for current clients and recent converters so you stop paying to reach people you already serve.
Radius or county targeting that matches the area you can genuinely serve, including drive time. Paying to reach families you will have to turn away is the most common Meta waste in ABA.
Give the delivery system room. Over-layered targeting on a small local audience raises costs and starves learning.
Seed from admitted families, not raw leads. Optimizing toward form-fillers optimizes toward people who fill forms.
A video that opens with the specific worry — an 18-month-old who stopped babbling, a four-year-old who melts down at every transition — self-selects the right audience better than any interest setting.
Speak to the worry, not the modality. Parents at this stage do not know what ABA is, and "evidence-based applied behavior analysis" means nothing to them. "Is my child just a late talker?" means everything.
Show real people and real rooms. Stock photography of smiling children reads as an ad; a therapist in your actual clinic reads as a place. Name the geography out loud, because a parent scrolling has no idea you are local until you tell them. State the payer situation plainly if you take Medicaid or the dominant commercial plan in your market, since that single fact removes the biggest objection before it forms.
And be careful with tone. Content that implies something is wrong with a child performs badly and, more importantly, is the wrong thing to put in front of a frightened parent. The best-performing ABA creative is reassuring and concrete: here is what the first step looks like, here is how long it takes, here is who you would talk to.
Instant forms produce the most leads and the weakest ones, because friction is near zero and so is commitment. Landing pages produce fewer, better leads, because a parent who clicks through and reads has spent something. Both are defensible, and the choice depends entirely on what happens next.
The middle path is better than either: an instant form with two or three qualifying questions — ZIP or county, insurance, child’s age — followed immediately by a thank-you step that offers the real next action. Not "we will be in touch." A link that starts actual intake, or a button that calls you now. The moment of highest intent is the second after they submit, and almost every practice wastes it on a confirmation message.
Whatever you choose, do not send a cold Meta lead a fifteen-page PDF packet. It will not come back.
This is the whole channel. A Meta lead contacted an hour later behaves like a stranger; contacted in two minutes, they behave like someone who just raised their hand, because they did.
The discipline that works: a text within sixty seconds that names the child and the practice, a call within five minutes from a local number, and a fallback sequence of several attempts across text and phone over the following week rather than one voicemail and a shrug. Ads run at 10pm and on Sundays — 48% of family contacts in our published cohort arrive outside business hours — so any coverage plan that stops at 5pm is throwing away roughly half of what you paid for.
Speed alone is not enough, though. The first contact has to advance the family, not just greet them: confirm you serve their area, confirm the payer, get the intake started while you have their attention, and if there is no diagnosis yet, hand them the evaluation path instead of a dead end.
Meta looks bad on cost per lead compared with nothing, and great compared with search. It looks bad on lead-to-contact rate. It can look excellent on cost per started client — which is the only number that matters. Judge the channel on admitted children and be honest about the lag: an ABA family sourced in March may not start until May, and a 30-day reporting window will tell you to turn off the thing that is working.
Feed conversions back. Meta optimizes toward whatever you tell it success looks like, so if the only signal it receives is "form submitted," it will find you excellent form submitters. Sending back the later events — qualified, intake completed, started care — retrains delivery toward families who become clients. This is the single highest-leverage account change most ABA advertisers have never made.
You are a healthcare provider advertising to families about a child’s development, which means the ordinary rules deserve real attention. Do not upload client lists as custom audiences without a defensible legal basis. Keep protected health information out of ad platforms entirely. Make sure any form that collects a phone number carries clear consent language for calls and texts, with an opt-out honored everywhere. And if a third-party pixel sits on a page where families enter health information, know exactly what it is collecting.
Yes, for reach and for reaching families earlier in the journey than search can. They work poorly as a lead-collection exercise judged on cost per lead. The practices that succeed on Meta respond within minutes, qualify during the first conversation, and have a defined path for families who do not yet have a diagnosis.
Because you interrupted them rather than answering them. A Meta lead never searched for ABA, so more of them are outside your service area, on a payer you do not take, or without a diagnosis yet. That is inherent to the channel, and it is manageable with instant qualification and an evaluation referral path — not by blaming the leads.
Instant forms with two or three qualifying questions, followed by an immediate next step on the thank-you screen, outperform both plain instant forms and most landing pages — provided you respond within minutes. If you cannot respond fast, a landing page with more friction will produce fewer and slightly better leads.
Minutes, not hours. Text within a minute and call within five. Because ads run around the clock and roughly half of family contacts arrive outside business hours, the response plan has to cover nights and weekends or it only works for half the leads you buy.
Name the worry in a parent’s own words, name the place, and name the next step. Avoid clinical jargon and avoid implying something is wrong with the child. Stating that you take the dominant local plan or Medicaid removes the largest objection before the parent forms it.
This page comes down to one discipline: reach the parent while the borrowed intent is still there, and make that first contact move them forward. Carelu does both the moment the form submits, including at 10pm and on Sundays when your ads are still running.
The account is still yours: creative, geography, audiences and budget. Carelu makes sure the leads those choices produce are judged on how they actually convert, not on how late someone called them.
Meta Ads connects to Carelu, so a lead form is answered by text within seconds and the conversation continues on phone, text or chat, whichever the parent prefers.
Service area by ZIP or drive time, payer, age and diagnosis status confirmed in the first exchange. Families you can’t serve are turned away politely instead of left waiting.
Meta reaches parents early, and 22% of families reach out before a diagnosis. They are routed to an evaluation path and kept in touch, not dead-ended.
The family gets a link to a phone-friendly intake prefilled from the conversation, and follow-up flows by text and email keep going until they respond or opt out.
Reporting shows speed to contact, intake completion, conversion and cost per started client by source, so you can see what each campaign actually produces.
The ad account is the easy half. What decides whether Meta is profitable for an ABA practice is what happens in the twenty minutes after a form submits — including at 9pm on a Saturday.
Meta leads are not bad. They are perishable.
Carelu answers every Meta lead the second it lands — texts, calls, qualifies, starts intake, and routes undiagnosed families to an evaluation path — at 2pm and at 2am.
Get a Demo