A ring drawn from the Central West End buys clicks in two states
Radius targeting is a circle, and nothing about this metro is circular.
Draw a ring around a Central West End address and it will cover the city, a slice of St. Louis County, and a stretch of the Illinois side across the river. Draw one around a Chesterfield address and a good part of it lands in St. Charles County or in farmland. Neither circle matches the map of people who will actually drive to you.
Build geography from a list instead. Municipalities, ZIP codes, and where they exist, drive time polygons. The county is a patchwork of small separate cities, so a targeting list here is genuinely longer than in most metros. Fifteen named places beats one radius every time.
Then treat the city, the county and the Metro East as separate lines in the account rather than one blended blob. They convert differently, they cost differently, and blending them hides both facts.
The test for whether the geography is right: could your office manager read the list out loud and agree that a patient from each of those places would realistically show up on a Tuesday at four?
Presence, not interest, or you fund people who merely looked up St. Louis
The default location setting in most ad platforms includes people who are not here.
Ad platforms will, by default, show your ads to people who show interest in a location as well as people in it. For a dentist that is close to pure waste. Someone in another state reading about St. Louis is not booking a crown.
Set targeting to people in or regularly in your locations, and set exclusions to people in or regularly in the places you refuse. Check it again after any account change, because these settings have a habit of resetting when campaigns are copied or upgraded.
Automated campaign types deserve the same suspicion. They will spend outside every line you drew if the signals point that way, so give them tight geography, a real negative list and a conversion definition worth optimizing toward, or keep them out of the account until the fundamentals are earning.
Pull the geographic report monthly. Any city on it that surprises you is a targeting bug, not a discovery.
Free clinic, denture kits and dental assistant jobs: three ways to spend Tuesday
The exclusion list is written before launch, not after the first invoice.
Dental keywords attract traffic that will never book. People looking for free or reduced cost care. Students searching for hygiene and assisting programs. Job seekers. Shoppers looking for mail order aligner kits or denture kits. Patients hunting for the school clinic rather than a private practice.
Every one of those clicks costs the same as a genuine implant consult. So the list goes in on day one: jobs, salary, hiring, school, program, course, free, cheap, coupon, kit, DIY, insurance definitions, and the names of any procedures you do not perform.
Add the competitor names you do not want to pay for and the neighborhoods you will not serve. If you do not take a plan, the plan name goes in the negatives, not just on your website.
Then keep reading the search terms report weekly for the first month and monthly after. An account left alone drifts, and it drifts toward cheaper clicks from worse patients.
A Chesterfield click and a Florissant click are not worth the same money
Value varies by place, so bids and budgets should vary by place too.
The mix of procedures, the plan participation and the likelihood of a household transferring in all differ across this metro. A click from one municipality may lead to a family of four with coverage, and a click from another may lead to a single emergency visit that never comes back.
You will not know which is which until you can attribute booked appointments to the campaigns that produced them. That means unique tracking numbers per campaign, a booked or not booked field the front desk actually fills in, and a monthly reconciliation against the schedule.
Once you have two or three months of that, segment. Fund the areas that produce kept appointments, cut the ones that produce phone calls about prices, and stop letting a metro average tell you the account is fine.
The cost of skipping this step is subtle. The account looks acceptable at the top line while half the money quietly buys the wrong half of the map.
Send kept appointments back to the auction, because a no-show teaches it the wrong lesson
Bidding algorithms optimize toward whatever you feed them, so feed them the thing you actually want.
Most dental accounts report form fills and phone calls over thirty seconds. Both are proxies. A practice can double its lead count and fill fewer chairs, and the ad platform will call that an improvement because it was told to.
Feed it better data. Import the appointment that was booked, and where you can, the appointment that was kept, with a value attached. Even a coarse value band, one number for hygiene and new patient exams and another for surgical and restorative cases, changes bidding behavior for the better.
Cost per booked case is the number that belongs on the report. Cost per lead belongs in the appendix. A practice can accept a high cost per click if the chair hour behind it is worth it, and cannot accept a low one that fills the schedule with people who never arrive.
Kept appointments take weeks to resolve, so read the account on a rolling window rather than a daily dashboard. Daily numbers in a single location practice are noise wearing a suit.
The Saturday toothache campaign needs somebody answering on Saturday
Urgent search is worth buying only when the phone gets picked up.
Emergency and toothache searches convert faster than anything else in dentistry and they are unforgiving. The caller is in pain, they will call three practices, and the first one to answer wins. If your Saturday plan is voicemail, do not buy Saturday clicks.
Give urgency its own campaign so it can have its own budget, its own hours, its own ad copy and its own landing page. Say what you can actually do same day, and say who is on call. Vague reassurance loses to a specific promise of a time.
Run the schedule against the front desk roster, not against the office hours on your door. The hour before opening and the hour after closing are often the best inquiry windows a practice has, and both are usually unstaffed.
Everything else in the account can be paced against the schedule board. When the hygiene column is full and the doctor column is not, the money should be moving toward the searches that fill the doctor column.
Questions we actually get
- What should we spend on dental ads in St. Louis?
- There is no honest benchmark we can hand you, and any agency quoting a market cost per click is repeating a number from somewhere else. The useful way to set a budget is backward from chair time: how many open new patient slots per week, what a kept appointment is worth to you, and what you can afford to pay for one. Then fund enough campaigns to be meaningful and no more.
- Should we run one campaign for the whole metro?
- No. City, county and the Metro East belong in separate campaigns with separate budgets that do not borrow from each other. Blending them hides which half of the map is working, and it makes it impossible to bid differently for a Chesterfield patient and a Florissant patient.
- Do Performance Max or automated campaigns work for a dental practice?
- They can, but only with tight geography, a real negative list and a conversion definition tied to booked appointments rather than form fills. Started cold, they generally find the cheapest clicks available, which in dentistry means job seekers and free care searches. We would earn the fundamentals first.
- How do we know whether a click became a patient?
- Unique tracking numbers per campaign, a booked or not booked field the front desk fills in, and a monthly reconciliation against the schedule. Then import booked and kept appointments back into the ad platform with a value attached. Without that loop you are optimizing toward phone calls, not dentistry.
- Is it worth bidding on our own practice name?
- Usually yes, and it is usually cheap. Competitors and directories bid on brand terms in this market as they do everywhere. The counterargument is that you may already rank first organically, so test it by pausing brand for a period and watching total booked appointments rather than clicks.
