Most dental marketing agencies will happily show you a return-on-investment number. Far fewer will show you how that number was produced, and fewer still will show you the same number twice. If you are evaluating dental marketing ROI for your practice, the most useful thing you can do is decide what you want measured before you sign anything — because measurement set up after the fact tends to measure whatever looks best.
This guide is written from the buyer’s side: what to instrument first, the three numbers worth asking a vendor to report every month, how call tracking and campaign tagging work, and how to sanity-check a report before relying on it. The same discipline applies to pay-per-lead channels such as Local Services Ads, where you are charged for a lead rather than a click.
Dental marketing ROI starts with instrumentation, not attribution
Attribution is a reporting layer, not a marketing strategy. Before a single dollar goes to attribution, four things need to be true, because each determines whether your numbers mean anything.
Your practice management system is the source of truth for “new patient.” Most systems can mark a patient record as new, or you can define new patient as one with no production history. Agree the definition in writing before anyone reports on it. A vendor who counts new patients one way and the front desk another produces a permanently confusing report.
Your website has one obvious conversion action. For most practices that is a phone call, then a form, then a booking request. Google documents how conversion tracking works for calls in Ads — configure it correctly from the start.
Your analytics can attribute a session to a source. Google Analytics 4 records traffic-source dimensions — session source, session medium, session campaign — which is how you see which channels produced traffic and which produced calls.
Your phone system can report where a call came from. If it can show call tracking data — number, source, duration — you can attribute calls without buying an attribution platform first.
If any of those four are broken, fix them before paying anyone to interpret data on top of them. Attribution software reporting into a broken foundation produces confident numbers attached to the wrong causes.
The three numbers worth asking every vendor to report
Ask for these three, by channel, every month. If a vendor cannot produce them, you are not buying measurement.
1. Cost per new patient, by channel
This is your most basic efficiency metric: channel spend divided by the new patients that channel produced. Calculate it yourself from raw data before trusting anyone’s version. If the vendor’s number cannot be reproduced from the underlying spend and the new-patient list, it is not auditable.
Two cautions. First, require the denominator to be new patients, not leads, not calls, not form fills. Leads inflate the number and cost you money. Second, be suspicious of a cost per new patient dramatically better than every other channel for no explainable reason. Most practices have one channel that looks far better when attributed loosely — usually the one that gets credit for patients who would have found you anyway.
2. Lifetime value to acquisition cost
LTV:CAC compares what a new patient is worth over time against what it cost to acquire them. If a channel costs a small amount to acquire a patient and those patients stay and generate production for years, that channel is doing real work. If acquisition cost approaches the value of the patient, the channel is expensive regardless of how many leads it produces.
Be precise about the assumptions, because this ratio is highly sensitive to them. A two-year LTV and a five-year LTV produce very different answers from the same spend. Ask what retention period was used and whether it reflects your actual retention rather than an industry assumption.
3. Phone answer rate
This is the one most practices overlook and the one that quietly determines whether your marketing spend does anything at all. If a quarter of the calls from your best channel ring into a voicemail box, a chunk of that channel’s apparent cost per new patient is money spent on calls nobody answered. Ask for the answer rate as a first-class metric alongside the patient numbers, and ask how missed calls are handled — whether they are routed to a callback, a text, or lost.
Google’s official documentation on conversion measurement and calls from ads describes how this is set up.
How call tracking actually works
Two mechanisms do most of the work, and both are simple once you know what to look for.
Dynamic number insertion means different phone numbers appear on different marketing channels. Search ads might display one number, your website a different one, and a printed mailer a third. When a patient calls, the tracking system knows which number they dialled and attributes the call to that channel. This is the standard mechanism behind most call tracking products, and it is why providers like Google Call Tracking, CallRail and Invoca exist as separate tools.
Campaign parameters are the tags you attach to a link so the analytics layer knows where a visit came from. Google’s traffic-source documentation sets out the manual tagging parameters — utm_source, utm_medium, utm_campaign, utm_term, utm_content — and how they surface as session and first-user dimensions, alongside the auto-tagging available when Analytics is linked to your ad platforms. Google’s ads and calls documentation covers the advertising side, including phone call conversions and a Google forwarding number for tracking when a call came from one of your ads.
Two questions separate a useful setup from a decorative one. Is the number insertion applied to all your channels consistently? If one channel is tracked and another is not, the comparison between them is meaningless. And is there a holdout? Some providers let you run a small portion of a channel untracked so you can compare tracked performance against a clean baseline. Without a way to test, you are trusting attribution you cannot check.
How to check a vendor’s report instead of trusting it
This is where buyer-side diligence is worth real money: every agency in this market sells ROI, and most sell it with a black box.
Ask for the raw inputs. A monthly report should be reconstructible from a spend ledger and a patient list. If a vendor cannot show both, you cannot audit them, and an unauditable number is a marketing claim, not a measurement.
Ask what they would have to be wrong about. A good report can be interrogated: if the number moved, what moved — spend, volume, close rate? A vendor with a real attribution model can explain the change. A vendor with a screenshot cannot.
Check whether your own data agrees with theirs. Your front desk knows how many new patients came in. If the agency report and your internal numbers diverge materially, one has a definition problem, and you want to know which.
Watch for the incentive. Vendors who sell a channel earn more when it looks good. Ask which channel they would recommend cutting, and whether they have ever advised a client to reduce spend. An agency that has never recommended cutting anything is selling, not measuring.
Be sceptical of multi-touch attribution models on a small budget. Sophisticated models need enough conversion volume and time to be meaningful. With a modest number of new patients a month, a complex model produces numbers that look sophisticated and are not stable. A simple, transparent channel breakdown is often more reliable at that scale.
What to do with the numbers once you have them
Measurement only pays if it changes decisions. A practical monthly loop:
- Check answer rate first. It is the cheapest thing to fix and it gates every other number.
- Compare cost per new patient across channels and reallocate toward the efficient one, subject to the caveat about assisted conversions.
- Watch LTV:CAC by channel over time, not in a single month, since retention takes longer than a quarter to establish.
- Investigate anomalies in both directions. A sudden drop is worth investigating, and so is a sudden spike — it often means a tracking change, not a marketing triumph.
If you are still working out how much to spend, our guide to what doctor marketing actually costs breaks down how agencies price their work, which is useful context when you compare a proposed ROI figure against the fee.
And if the goal is more patients through the phone rather than a spreadsheet, how to get more patients covers the acquisition mechanics that these numbers are measuring. If a campaign is running on Google Ads, our guide to Google Ads for doctors covers how to structure it so the numbers attribute back to a channel.
A note on what nobody can promise
No agency can guarantee a ranking, a traffic figure or a number of new patients, and any report promising one is making a sales claim rather than a measurement. Google treats manipulative practices aimed at rankings as a violation of its spam policies, and its documentation on AI-assisted content notes that automated tools still require human review before publishing — a reminder that no measurement system substitutes for someone checking the work.
What a credible agency can do is instrument your marketing properly, report the three numbers above consistently, and let you see the underlying data. That is worth paying for. A number you cannot reconstruct is worth paying nothing for.
If you want a second opinion on whether your current tracking setup supports the ROI figures you are being shown, talk to us about it and we will say plainly whether the underlying data supports the claim.
Ready to grow your practice?
Book a Free Strategy Call
