Six front doors into the practice, and nobody owns the web chat
Count the ways a patient can reach you, then name the person responsible for each one.
A typical practice has more inbound channels than it realizes: the main phone, the web form, the Google Business Profile message thread, a chat widget, text messages to a tracking number, referrals from other offices, and people who walk in from the block. Most practices staff two of those properly.
Do the inventory on paper. For each channel: who watches it, how fast, what happens after hours, and where the record ends up. Any channel without a name next to it is a leak, and the unread ones are usually chat and profile messages.
Turn off what you cannot staff. A chat widget that collects inquiries nobody answers until Monday is worse than no widget, because the patient believes they have made contact. Same with a contact form that routes to an inbox one person checks on Fridays.
Then set a response clock. Inquiries decay quickly in dentistry, especially anything urgent, and the practice that replies first usually gets the patient regardless of who is better.
The referral slip from the Clayton specialist sits in a pile until Thursday
Referrals are the highest converting inquiries a practice gets and the least likely to have a process.
Referrals arrive by fax, by portal, by phone from another office's front desk, or in a patient's hand. They convert at rates no advertising channel will match, because a professional already vouched for you. And they routinely wait days because nobody owns the pile.
Give referrals the same clock as a paid inquiry. Someone opens them daily, calls the patient the same day, and logs the referring office so you can see which relationships actually produce.
The reciprocal loop matters too. Send a note back when the case is complete. Specialists in Clayton, Chesterfield or Kirkwood remember which general practices closed the loop, and referral relationships in this metro are personal and durable.
Track it. Referral source belongs on the chart at intake, entered by a person, not guessed at by software afterward.
Answer the money question on the first call or the Webster Groves caller dials the next practice
Coverage and cost end more new patient calls than anything else, and most practices handle them by promising to call back.
The most common intake failure in dentistry sounds polite. We will verify your benefits and get back to you. The caller says fine, hangs up, and books with whoever answered the question live.
Fix it with preparation rather than heroics. Build a one page reference of the plans you are in, the plans you are not, and what a new patient exam and cleaning generally runs for a self pay patient. Put it at the desk. Coverage details are always worth confirming with the payer afterward, but the caller needs an answer now.
Where you can verify live during the call, do. Where you cannot, book the appointment first and verify after, with a clear note that the estimate is subject to the plan's confirmation.
Self pay callers need a path too, not a shrug. A stated fee for the first visit and a payment option turn a price shopper into a patient more often than a vague invitation to come in and discuss it.
Sort last quarter's new patients by municipality and the map redraws itself
A metro average hides the fact that the city, the county and the Illinois side convert differently.
Ask every new patient which town they are coming from and write it on the chart. Not the ZIP on the insurance card, the place they would say out loud. After a quarter you will have something no analytics tool can give you: a real map of where your patients come from and where they do not.
Read booking rate, show rate and production against that map. Patterns show up quickly. A municipality that books well and shows poorly is a drive time problem, and the fix is which appointment times you offer, not the website. A municipality that calls and never books may be a plan participation problem.
This is also how you find out whether crossing the county line is realistic for routine hygiene. Often it is not, and the honest answer changes where you advertise and what you write.
The Metro East deserves its own row in the table. Different state, different plan landscape, different drive. Averaging it into a metro number tells you nothing about either side of the river.
The recall column and the new patient column leak in different places
Two funnels, two sets of failures, and treating them as one hides both.
New patient intake leaks at first contact: unanswered calls, unanswered messages, unresolved money questions. Recall leaks later, in confirmation and rescheduling. A patient who cancels a hygiene appointment in February and never rebooks is a different problem from a caller who never got through.
Measure them apart. For new patients: contacts, booked, kept, and production from the first ninety days. For recall: percentage of the hygiene schedule prebooked, reappointment rate, and how many lapsed patients were contacted this month by a person rather than an automated message.
Confirmation sequences do most of the work in the recall column. Text confirmations, an easy reschedule path, and a call for the appointments you cannot afford to lose. Winter weather here empties a schedule without warning, so a written plan for refilling a canceled morning is worth having before the first hard freeze.
The cheapest booked appointment in any practice is usually a patient already in the system. Work that list before you buy another click.
You will never have the volume to prove a headline, so change what needs no proof
One practice in one metro cannot power a split test, which is an argument for bigger changes, not for guessing.
Split testing needs volume that a single location dental practice does not have. Running a button color test for six weeks to learn nothing is a common way to spend an agency retainer.
So change things whose effect does not require statistics to see. Answering the phone during lunch. Replying to profile messages. Publishing your fees for a first visit. Adding a text option. Each of those either happens or does not, and the schedule board shows the difference.
For the rest, run sequential changes with a clock and a control period. Grade twenty recorded calls before and after a script change. Compare a full month against the same month with the change in place, and accept that you are looking for large effects only.
Write down what you changed and when, in one shared document. Without that log, a busy quarter gets credited to whatever was most recently installed, and a practice ends up paying to keep something that never worked.
Questions we actually get
- Where should we start if we can only fix one thing?
- Answer every channel. Count how many ways a patient can reach you, assign a name and a response time to each, and turn off the ones you cannot staff. Unanswered chat threads and profile messages are the most common leak we find, and closing them costs nothing but attention.
- Is it worth running A/B tests on our website?
- Generally not at a single location practice. The traffic will not support a readable result on small changes. Sequential changes with a control period, plus call grading before and after, are more honest tools. Save split testing for changes big enough that you would notice them on the schedule board.
- How do we track where patients actually come from?
- Ask at intake and write it on the chart: the town they would name, and how they heard about you, entered by a person. Software attribution guesses, and in a metro split between an independent city, a county of small municipalities and an Illinois side, the guesses are unusually wrong.
- Should the front desk quote prices over the phone?
- A prepared range for a first visit and a clear list of the plans you take will keep more callers on the line than a promise to call back. Confirm coverage with the payer afterward and say plainly that the estimate depends on that confirmation. What loses patients is silence, not the number.
- How long before intake changes show up in production?
- We do not give timelines, and any agency that does is guessing. The mechanics are worth understanding instead: answering channels shows up in booked appointments within weeks, confirmation and recall work shows up in kept appointments over a quarter, and case acceptance changes take longer because they depend on the operatory, not the phone.
