Lead forms, landing pages, or DMs: where should my clinic ads send people?

Why this is the biggest open question on my calls

Six of the 67 sales calls in my backlog asked some version of this, the most-asked question that had no page. It sounds like "should we turn Messenger off and just do the lead form?", "lead forms are trash because Meta controls the data, true?", and, three times in two weeks, "what is CAPI?" Underneath all of them is one decision: where does the click land, and who gets to see what happens next.

Every destination trades one thing for another. A lead form trades data for friction: fewer taps, less back to you. A landing page trades friction for data: more taps, but every booking can be fed back to the ad platform. A DM trades both for a conversation. None of them is trash. Each is wrong for a specific clinic at a specific moment.

What the Benchmark says about the data gap

The Campfire Benchmark covers 15 cash PT and chiro Meta ad accounts from October 2025 to September 2026, nine of which spent money. Pooled cost per lead was $39.85, where a lead is Meta's own count: pixel or CAPI Lead events plus instant-form submissions. Strip out the lead-magnet downloads and cost per patient lead was $61.34. Only three of the nine spending accounts fired a Schedule event at all; across those three, a booked appointment cost $131.96 pooled.

Six of nine clinics running ads could not see a booking from inside their ad account. That is the real problem here. It is not that forms produce bad leads. It is that most clinics have no way to tell the platform which leads became patients, so the campaign keeps optimizing for whoever fills a form fastest. That is how you end up with the cheap lead who wanted a PDF.

What I saw in the flagged account

Last week I audited an account that Meta had just flagged as healthcare. Registrations had collapsed and the owner was asking whether a lead form or a brand-new ad account was the better fix. Three things were true in that account at once: no Conversions API, browser-only pixel tracking, and an event match quality of 6.1 out of 10. The previous ads manager's numbers looked bad partly because the bookings that did happen were never being matched back to the ads.

The flag itself, as far as I could see, attached to the domain, so a new ad account pointing at the same website would inherit the problem. Sending the click to an instant form or a DM instead kept the existing CRM follow-up working the same afternoon, because the lead still lands in the same place. The tracking fix was separate and came first: two-factor authentication on every account, then CAPI installed properly.

What CAPI actually is

The Conversions API is a server-side line from your CRM back to Meta. The pixel watches a browser; a phone with tracking blocked or a patient who booked by text never shows up. CAPI sends the event from your system instead: this phone number booked, this email paid for an eval. Meta matches it to the person who clicked, and the higher the match quality, the better the campaign learns who to show the ad to next.

So, "do I need CAPI for the opt-in?" No. An instant form is already on Meta's servers. You need CAPI for everything after the opt-in: the booking, the show, the paid plan of care. Without it, a landing page is just a slower lead form.

How the three compare

• Instant lead form. Opens inside Meta with name and phone pre-filled. Lowest friction, so the most leads, and it keeps running when a website is flagged. You get name, phone, email and your qualifying answers; Meta keeps the behavioral data. Good for: no tracked website, a flagged domain, or a solo owner who needs leads to self-book from the last question on the form.
• Landing page. Your domain, your pixel, your form. More taps, so fewer leads, but with CAPI every booking becomes a training signal. Good for: a clinic that already tracks a Schedule event, or one ready to install CAPI before spending more.
• DM ad. The click opens a conversation. Highest intent per message and the hardest to measure: the Benchmark's largest account counted conversations started, not leads, and tracked no booking at all. Good for: a clinician who will personally answer within minutes. Bad for: anyone planning to let a bot run it.

What to do, in order

If you are starting, or Meta just flagged you, run the instant form. Add two qualifying questions and end it with a booking link or a question that lets the lead pick a time. Text from your own number before you call, because in the accounts I've watched, only about one in twenty leads clicks a booking link on their own even when they're warm.

Then, before you raise the budget, install CAPI from your CRM and get the Schedule event firing. When you can see cost per booked patient in the ad account, and not just cost per lead, you have earned the landing page. Move traffic over one campaign at a time. If you want two landing pages, run two offers, not two colors.

The decision isn't form or page. It is whether the platform can see what a patient looks like. Fix that, and either one works.