Sunday night, a little after eleven, and a molar that was fine at dinner has turned into a throb nobody in the house is sleeping through. The guy holding frozen peas to his jaw types “emergency dentist near me” into his phone, and the first thing on screen is an ad. Dental PPC is the business of being that ad, and it lives or dies on one comparison: what the click costs against what the patient behind it is worth. Below is what clicks run in 2026, how to set the account up, and how to hold cost per patient down. What is really working is different from a well-run account, and we get there at the end.
One note on numbers: search volumes, difficulty scores, and cost-per-click figures in this article come from our own keyword research, a pull of 1,341 healthcare keywords, July 2026. Outside facts are linked to their primary source. Hours, channel rankings, and timelines are LabRanked planning estimates, not measured benchmarks.
Why this matters right now
The demand under a dental account is large, and some of it has been sitting untreated for years. CDC counts 1 in 5 adults aged 20 to 64 with at least one untreated cavity. CDC says nothing about when those people search, so read the next bit as ours: a cavity does not fix itself, so that pool keeps turning into somebody’s 11pm emergency.
They are also spending their own money, and that changes how the click behaves. The ADA’s Health Policy Institute reports that 45% of the US population had a dental visit in the past 12 months in 2022, with out-of-pocket the largest share of the $189 billion spent on dental care in 2024. Most medical advertising reaches somebody spending an insurer’s money, while a lot of dental advertising reaches somebody working out whether they can afford you this month. Price shows up in the search itself, and a page that dodges it loses the click it just paid for.
Now the blunt part. We are not going to print a national cost per click for “dentist near me” or “dental implants,” because every public benchmark we went looking for traced back to companies that sell ad management, we could not check their method, and repeating them would make us part of the problem. What our own keyword dataset does cover is the other side of the auction, the terms aimed at you rather than patients, and there “dental ppc” carries a cost per click of $30.82, “dental local seo” $32.00, and “dental office answering service” $135.74. That is what the auction prices a click at for the chance to pitch one dentist.
Hold every benchmark you are shown up against that number.
What used to work, and what it’s still good for
Most dental accounts we inherit are still built on a playbook from about 2018. None of it was foolish then. It is aimed at the wrong job now, and in a paid channel the wrong job sends you an invoice every week.
Broad match on “dentist.” It was how you discovered searches you had not thought of, and in a mature account with clean conversion data it still works. Run as your main targeting, the search-terms report fills with dental assistant jobs, hygiene schools, denture repair kits, and teeth cleaning for dogs. All at your rate.
The $59 cleaning special as the whole offer. It produces appointments, which is not the same as producing patients. Coupon traffic fills the schedule with low-value visits and no implant or ortho cases. Keep it as one offer among several, never the account’s only door.
Sending every click to the homepage. Fair enough when nobody had the resources for fifteen pages. Now somebody who searched for one specific problem lands on a general page and has to hunt for what they came for, at eleven at night, in pain. We would not bet a paid click on them bothering.
Boosted Facebook posts. Boosts buy engagement, and engagement is what they report back. They reach people who are not in pain yet, and the ones who are have already opened Google.
Handing an automated bid strategy the budget on day one. Google’s guidance is that conversion tracking has to be running before you start, and that Smart Bidding should be judged over a period holding at least 30 conversions, or 50 for Target ROAS. Google will let you switch it on with no history at all, and with tracking that is not right yet it chases whatever cheap conversion it can find, which in the accounts we take over is often a form from somebody who wanted a job.
Counting form fills as new patients. A form fill is a hand raised, and a booked and attended appointment is revenue. If your account reports the first while your practice software knows the second, the two numbers have never met and your budget is set by the wrong one.
They were built to buy attention broadly, back when attention was cheap. What pays now is buying one specific intent and answering it the second it arrives.
What’s working now, ranked
| What you fix | Effort | Speed | Verdict for 2026 |
|---|---|---|---|
| Conversion tracking with calls counted | About 3 hours | Immediate | Do this before you touch a bid |
| Exact and phrase match, service plus city | About 4 hours | Same week | Where the core of the budget goes |
| Negative keyword list | About 3 hours | Immediate | The cheapest work in the account |
| One landing page per advertised service | About 5 hours per page | Weeks | Moves cost and bookings together |
| Call assets scheduled to staffed hours | About 1 hour | Immediate | Dental books by phone, so not optional |
| Ad and landing page quality work | About 4 hours | Weeks | Lowers the click price without raising the bid |
| Emergency and same-day terms | About 2 hours | Same week | The highest intent in the vertical |
| Automated bidding | About 1 hour | Judge it over 30+ conversions | Good later, expensive early |
| Ranking pages for the same services | Low cash, months | Slow | What eventually shrinks the invoice |
That ranking and those hours are our framework, not measured data, so price them against your own account.
Fix the measurement before you fix the bids. Dental books by phone, and in the accounts we inherit calls outnumber form fills, so an account counting only forms cannot see where its patients came from. Google’s call assets put a tap-to-call button on the ad, and clicks on that button cost the same as a click on your headline, with call reporting, once you switch it on, routing through a Google forwarding number so calls can be counted as conversions. Set the asset to show only when somebody answers. A paid click that rings an empty office at 6pm is worse than no click, because now the patient knows you did not pick up.
Buy the service, not the profession. A search with a service and a city in it is somebody planning to spend money. A search with just the profession might be a browser, a comparison shopper, or a student writing an assignment. Build the account on the first, add the second once your data can tell them apart, and put emergency terms at the front of the queue.
The negative keyword list is the cheapest work in the account. Jobs, salary, hygienist, assistant, school, courses, free, cheap, DIY, denture repair, and dog. Fifty terms in an afternoon, then fifteen minutes a week reading the search-terms report and adding what turned up. Nobody enjoys it, which is why it stays undone and why we start there.
One page per service, and answer the money question on it. The headline repeats the search, the page covers what the treatment involves, what it costs or what the range depends on, and what insurance and financing you take, with one form and one tracked number visible without scrolling. Since out-of-pocket is the largest share of dental spending, a page that hides price asks a cash payer to phone and ask, and plenty will phone somebody else.
Quality is the lever that lowers your click price without raising your bid. You control the bid too, but moving that up just pays more for the same patient. Google explains that Ad Rank decides whether and where your ad shows, calculated at auction time from six factors including your bid and the quality of your ads and landing page, and it states plainly that higher quality ads can often lead to lower CPCs. Be precise here, because the market is loose about it. Quality Score is built from expected clickthrough rate, ad relevance, and landing page experience, and Google’s documentation says Quality Score is not an input in the ad auction and is a diagnostic tool. Use it as a list of which keywords have a weak ad or page behind them, and ignore anyone selling Quality Score as a product.
Set the budget so it cannot ambush you. The practices we talk to underfund ads because they are picturing a runaway bill. Google publishes the guardrail: you will not pay more than twice your average daily budget on any single day, or more than 30.4 times it in a month, for most campaigns. Pick the monthly number you can lose without flinching, divide by 30.4, and let it run long enough to learn something instead of throttling it every Tuesday.
Do the arithmetic in your own numbers. Take what the service is worth to you and multiply by the share of enquiries that become booked and attended appointments, which your front desk can pull from last year. That is what a lead is worth. Divide it into what the account pays per lead. If the ratio is comfortable, spend more, and if not, work the leaks above before touching a click price. We are not printing an example table, because a plausible set of made-up figures is how practices talk themselves into a budget that never made sense.
The 30-day action plan
- Days 1 to 3, install the measurement. Conversion tracking live, calls counted through call reporting, and your booked-and-attended rate pulled from last year’s schedule. About 3 hours.
- Days 1 to 3, write the negative list. Fifty terms from the obvious categories plus your own search-terms report, then a standing fifteen minutes a week. About 3 hours.
- Week 1, pick one service and one city. Highest value, clearest intent, and real capacity to see the patients. One service done properly beats five done thinly. About 1 hour.
- Week 2, build the landing page. Headline matching the search, what it costs or what the range depends on, insurance and financing, one form and one tracked number above the fold. About 5 hours.
- Week 2, rebuild the campaign on exact and phrase match. Service plus city, ad copy that repeats the search wording, call asset scheduled to staffed hours. About 4 hours.
- Week 3, work the quality diagnostics. Pull the keyword-level scores, find the components graded below average, and rewrite the ad or the page behind each one. About 4 hours.
- Week 4, read what happened and cut. Which searches produced booked appointments, not enquiries. Kill the rest and move that budget onto what worked. About 3 hours.
- Week 4, start the page that makes the ad optional. The organic version of your best service page, written to rank rather than to be paid for. About 4 hours.
Those hours are LabRanked planning estimates rather than measured benchmarks, and they come to about 27 across the month, most of them in the first fortnight. Steps 1 and 2 are the ones we would run before touching anything else, and the two we most often find skipped.
So what’s really working in 2026?
Not a cheaper click. A smaller share of your patients arriving through the meter at all.
Every practice in your area is bidding into the same auction on the same margins, which makes paid search the one place where somebody else sets your costs. You can tune it hard, and you should, and there is still a ceiling. Dentistry has an advantage most advertised categories do not, and almost nobody builds an account around it: a patient you win once comes back for recall visits for years. Your real ceiling in the auction is set by your recall system and your front desk, not your bid strategy. A practice that keeps patients can pay more for the next one and comes out ahead.
So the account worth having is the one quietly making itself unnecessary. The service page that converts your ad clicks is the same page that ranks, and the cost answer on it is what an AI assistant quotes when somebody asks what an implant runs in your city. Every ranking you gain moves a slice of the budget from rented to owned.
Cost per patient comes down when fewer of them have to be bought.