Dental Practice Marketing Attribution: Which Channels Actually Work

Most practices paying $5,100 a month for marketing are buying two things: some new patients, and the credit for patients they were going to get anyway. The second item costs more than the first, and no dashboard on the market will separate them for you.

Our view, stated plainly so you can argue with it: for a general practice under roughly $2M in collections, almost every marketing dollar above a maintained review profile and a phone that gets answered is buying attribution, not patients. That is a genuinely contestable claim. Plenty of good operators will tell you their paid search is the only reason the schedule is full. Some of them are right. Most of them have never audited the call log against the invoice, and the gap between what an agency reports and what an agency caused is where six-figure marketing budgets go to hide.

The 466 Patients Who Explained How They Actually Chose

Start with the least convenient piece of evidence in the field. In a 466-participant telephone survey published in a peer-reviewed analysis of how older patients select a dentist, recommendation and word of mouth drove 65.6% of awareness and 56.7% of final selection, while internet and modern technology accounted for 7.3% of awareness and just 2.6% of selection. That is not a rounding difference. That is one channel doing twenty times the closing work of another.

The obvious objection: that study skews old, and older patients search less. Fair. So put it next to a broader consumer sample. In a 1,500-respondent U.S. survey, 43.3% said they picked a dentist based on online reviews, 19.4% on proximity, 18.4% on search-result rank, 14.4% on price, and 4.5% on website design. A separate 2023 survey of 1,099 Americans found 8 out of 10 use online reviews as the first step in choosing a new dentist, and 83% use reviews to evaluate one.

Read those three together and the picture is not contradictory, it is sequential. A person hears your name from a coworker, a neighbor, or a hygienist who moved. Then they check reviews to make sure the name is safe. Then they check the drive. Reviews are not a discovery channel. They are a verification layer sitting downstream of a conversation you did not pay for and cannot see.

The scale of that verification layer is settled. An analysis of 204,751 U.S. dentist listings found 154,683 (75.55%) had received at least one patient rating, and 150,571 (73.54%) held an overall rating of 3 or better out of 5. Three quarters of the profession is already rated and already acceptable. Which means your reviews are rarely why someone picks you. They are why someone does not eliminate you.

What Most Owners Get Wrong: Paying for Demand Capture and Calling It Demand Creation

The specific mistake: crediting the last click, then funding the last click as if it created the visit.

The owners who make it most often are three to eight years in, one or two locations, and just came off a soft quarter. They hire an agency. Within 90 days the schedule firms up, the report shows 20-plus attributed new patients a month, and the retainer becomes untouchable because it looks like the thing that fixed the quarter. Second most common: the associate who just bought a mature practice with thirty years of referral gravity, turns on paid search, and books the inherited goodwill as a marketing win.

Here is the mechanic. A patient hears your name at work on Tuesday. On Thursday they type your practice name into Google. Your ad sits above your own map listing, so they click the ad. The platform records a paid conversion. You are now billed for a patient the coworker delivered. Multiply that across branded search, retargeting, and a call-tracking number that only fires on the last touch, and the report is not lying so much as it is congratulating itself for opening a door someone else walked through.

The same error runs the other way on websites. Only 4.5% of respondents in that 1,500-person survey named website design as their deciding factor. We would not spend $18,000 rebuilding a site that loads, states hours, lists insurances, and has a working request form. We would spend $400 making the phone number tappable and go back to work.

$61,200 of Marketing, Three Different Answers

Take a composite general practice: 1,600 active patients, 38 new patients a month, and a marketing stack of $2,500 a month in agency fees plus $2,600 a month in ad spend.

That is $5,100 a month, or $61,200 a year.

Total new patients run 456 a year, and the dashboard claims 21 a month, or 252 a year, from paid channels.

On those numbers the reported cost per new patient is $242.86, and everyone in the room nods.

Now audit the calls. Pull 90 days of recordings on the tracking line and sort them by one question: did the caller name a person, or name your practice, before they named a search? Say 13 of the 21 did, because they were referred or they were looking for you specifically. True incremental new patients from paid drop to 8 a month, or 96 a year.

Incremental cost per new patient is then $637.50, not $243.

At $1,240 in first-year production and 62% overhead, each of those patients contributes $471.20 in year one.

So the first year on paid loses $166.30 per patient.

Across 96 patients that is $15,964.80 of year-one negative contribution, which is the number nobody puts in the report.

Run it to three years and the verdict flips. At $2,900 of cumulative production per retained patient, contribution is $1,102.

Against $637.50 of acquisition cost that clears $464.50 per patient.

Which is $44,592 across the cohort, and a perfectly good use of money.

Read what just happened. Paid acquisition in this practice is not a marketing decision at all. It is a retention bet. It is solvent only if those patients reappoint, and it is insolvent inside twelve months if they do not. If your hygiene reappointment rate is under 80%, you are not buying patients at $637.50. You are renting them at a loss and calling it growth.

Third answer. Move $18,000 of that $61,200 into a review and recall system plus two protected hours a week of team time asking for referrals by name. For that reallocation to beat paid on its own terms, it has to produce 28.2 new patients a year.

Across 1,600 active patients, 29 referrals is 1.8% of your existing base bringing one person each.

We would take that bet every time. One in fifty-five patients making one introduction is not a campaign. It is a Tuesday.

The Review Asset Half the Profession Refuses to Maintain

There is a quieter problem sitting on top of all this. ADA membership survey data found that 88% of responding dentists said they had received online reviews from patients, and 39% said they were unable to respond because of HIPAA regulations. So nearly nine in ten own the asset that 83% of prospective patients use to evaluate them, and four in ten have decided they are not allowed to touch it.

That is a training gap dressed as a compliance constraint. You cannot confirm someone is a patient, discuss their treatment, or argue about their crown in public. You can absolutely post a general response about how your practice handles concerns and how to reach the office. Our call: a practice with 60 reviews, none answered, is leaving the cheapest trust signal in the funnel on the floor, and the fix costs an hour of policy writing and a saved template.

Three Questions That Beat Any Attribution Platform

You do not need better software. You need the front desk to ask better questions, in this order, on every new-patient call.

• "Who told you about us?" Not "how did you find us," which prompts people to name the last screen they looked at. Ask for a person. Log the name. Names are the only referral data that compounds, because they tell you who to thank.
• "Had you heard of us before you searched?" This single question splits demand creation from demand capture. If the answer is yes, that conversion belongs to word of mouth no matter which ad it touched.
• "What made you call today?" Pain, insurance renewal, a move, a school form. Timing is the variable your ads cannot manufacture, and knowing it tells you whether you are buying urgency or interrupting it.

Then separate branded from unbranded spend in the ad account and look only at unbranded cost per booked appointment. Branded search is a toll booth on a road you already built. Pay it if competitors are bidding on your name, but never count it as acquisition.

Where We'd Put the Next $10,000

In order, and we would not shuffle it.

First, $0 on anything until the phone is measured. If 12% of new-patient calls go to voicemail at lunch, no channel in the world fixes that, and every dollar upstream is discounted by the miss rate. Second, review volume and recency: a standing ask at checkout, plus responses to everything, because being unrated is the only review problem that actually eliminates you when three quarters of dentists already clear 3 stars. Third, the referral loop: named asks, specific thanks, and a reason for a satisfied patient to say your name out loud. Fourth, geography, since 19.4% decided on proximity, which means signage, parking clarity, and map accuracy are marketing. Unbranded paid search comes fifth, sized to a three-year payback you have actually verified, not a twelve-month one you assumed.

We would drop the website rebuild. We would cut the retainer in half and hold the ad spend flat for one quarter while you instrument the calls. If new-patient volume does not move, you just found $15,000 a year.

Do This Before Your Next Agency Call

Pull 90 days of tracking-line recordings and sort every new patient into referred, brand-aware, or genuinely cold. Separate branded from unbranded ad spend in the account this week and recalculate cost per new patient using cold callers only. Write a four-line review-response template your team can post without touching any clinical detail, and set a standing weekly ask at checkout. Then bring the cold-only number to the agency and ask them to defend the retainer against it.

What we are tracking next: whether any practice can show a durable unbranded paid channel that clears its cost inside twelve months rather than three years. That would change our position, because it would mean paid acquisition is a marketing decision after all instead of a retention bet in disguise. Send us the numbers if you have them.

All of this math turns on one figure most owners are carrying from last year: your real overhead percentage, which decides whether a $637 patient is profitable or expensive. Pull your last twelve months of collections and expenses and run your own numbers through our free overhead calculator, then recompute your cost per new patient against the contribution margin it actually leaves you.