9,233 patient enquiries at $12.18 each.

Smile Loft Dental runs a group of practices across Maryland. We built their website, their brand and their print, then ran patient acquisition across Meta and Google — and automated the part almost nobody automates, so that an enquiry arrives at the front desk as a name, a number, a service and a location rather than as a notification someone has to go and find. Twenty-nine months of Meta activity has produced 9,233 enquiries at $12.18, against a published dental benchmark of $28 to $55.

This was not a media engagement with a website attached. The order matters, because most of the cost advantage below comes from things that are not advertising.

Most dental advertising fails after the click, not before it. The ad works, the form fills, and the enquiry lands in a platform inbox that a practice manager checks on Thursday. By then the patient has booked with somebody else.

So the capture and the routing were built before the spend scaled. On Meta, the conversation itself does the collecting: an automated thread that will not complete until the patient has given four things — their name, their phone number, the treatment they want and the location they want it at. On Google, the same four fields. Both write straight into one sheet, by location, in the order they arrived.

That design decides lead quality before a single dollar is spent. A prefilled form asks a patient to tap once; a conversation asks them to answer. Fewer people finish it, and the ones who do have told the practice what they want before anybody picks up the phone.

Dentistry is one of the most expensive categories in paid search. Published 2026 benchmarks put the average cost per click for dental at around $7.85 to $8.00, and cost per lead between $50 and $85 depending on the source and the treatment.

On Meta the account runs at roughly a quarter of the lower end of the published range. And there is a reason to think that comparison understates rather than flatters us.

The $28 to $55 benchmark is for Meta lead ads — the low-friction format the platform built for volume, where the patient’s name and contact details are prefilled and the action is a single tap. The enquiries counted here required four answers typed into a conversation. That is a harder bar for a patient to clear, and it is being cleared at a quarter of the cost of the easier one.

If you would rather judge us only on classic lead-form submissions and set the conversational captures aside entirely, that figure is $18.62 — still a third below the bottom of the band, on the strictest like-for-like available.

There is no Google cost-per-lead row in that table. Conversion tracking on the website was not correctly implemented for most of the Google period, so the platform’s count understates the leads those campaigns produced, and we would rather publish nothing than publish a figure we know to be wrong. What can be stated is the click cost, which the platform measures itself: $3.03 against a category average near $8.00, across 36,672 clicks.

The binding constraint on this account was never the advertising. It was how fast the practice could call people back.

When enquiries arrived faster than the front desk could work them, the right thing to do was stop buying more — so campaigns were paused, then restarted once the backlog cleared. That happened repeatedly over two and a half years, and it is written all over the account: 242 lead campaigns, 70% of them named with an explicit date range like “16th–20th June”, at a median spend of $414 each.

Every restart puts a campaign back into Meta’s learning phase, which it leaves at roughly fifty optimisation events. It costs real money:

Interrupted campaigns cost 73% more per enquiry. They took 47% of the lead budget to produce 34% of the enquiries. The $12.18 on this page is a weighted average of a channel running properly and a channel being switched on and off — with continuous delivery it would sit closer to $9.77.

Two more, briefly. Nine locations in two languages produced a fragmented account structure that current Meta guidance would consolidate, and some of that fragmentation is testing that should have been folded back in sooner.

And the measurement stops at the enquiry. We built the routing so that leads reach the practice; closing the loop the other way — from booked patient back to the campaign that produced them — is a separate build, which we proposed and which has not been commissioned. That is why this page can tell you what an enquiry cost and cannot tell you what a patient cost. We would still argue for it, here and anywhere else, because it is the only measurement that actually settles whether the advertising is working.