When six practices share your block, hold music is a referral
Density means the alternative to you is thirty seconds away in a search result, so answer speed is a competitive position rather than a courtesy.
A caller who reaches voicemail at 12:40 does not usually call back. They tap the next result. In a city where several practices sit within a few blocks of each other, the cost of a missed call is a patient who booked somewhere else before you noticed.
Count the misses first. Most practice phone systems can report unanswered calls by hour. Nearly every office finds the same two holes: the lunch hour and the twenty minutes around opening and closing.
Then close them with coverage rather than a script. Someone covering lunch, a rollover, or an answering service that can actually see the schedule. A message taken by someone who cannot book anything is only marginally better than voicemail.
Missed calls should trigger a text back within a minute or two, in the same thread, from the practice number. People here text. Many will never call again but will answer a message.
Out of network needs a two-sentence answer, not a callback
The insurance question ends more New York dental calls than price does, and it usually ends them because nobody at the desk could answer it cleanly.
The caller asks whether you take their plan. If the answer is a pause, a maybe, or a promise to check and call back, the call is effectively over even when it ends politely.
Write the answer down. Two sentences the whole team says the same way: what you are in network with, how out-of-network benefits generally work at your practice, and what happens next. Consistency matters more than elegance.
Verify while the caller is still on the line wherever you can. A benefits check that happens after the call ends is a booking that never happens, because the caller has kept shopping in the meantime. Specific coverage varies by plan and employer and is worth confirming with the carrier before quoting anything firm.
Track how many calls die on this question for two weeks. It is usually the single largest category of lost inquiries in a practice with any out-of-network exposure, and it is fixable with training rather than spend.
Late arrivals here are a wayfinding failure, not a patient failure
Patients who get lost in a lobby, a freight corridor or an unmarked stairwell arrive late, lose their slot and do not rebook.
A first-time patient walks into an office building, meets a security desk that wants photo ID, works out which elevator bank goes to your floor, and looks for a door that may only have a suite number on it. Ten minutes disappear. In a walk-up or a brownstone with a basement entrance it can be worse.
Put arrival directions in the confirmation and in the reminder, not just on the website. Which entrance, what the desk needs, which elevator, which floor, which door, and a photo of the front of the building.
Look at your own late arrival log by patient visit number. If first visits run late far more often than returning ones, the building is the cause and the fix costs you nothing but writing.
Small courtesies convert. Telling a first-time patient to arrive a few minutes early because of the desk, and telling them who to text if a train stalls, changes the outcome of the appointment before it starts.
A late-cancellation policy written for drivers will cost you reviews
Transit failures are outside a patient's control, and a rigid fee schedule turns an unavoidable delay into a public complaint.
A patient stuck on a stalled train did not choose to be late. A policy that treats them the same as a no-show generates exactly the review you would expect, and in this market reviews carry a lot of weight for people who have no one to ask for a referral.
Keep the policy, and give the front desk explicit discretion. Written discretion, so it is applied consistently rather than by mood.
The better protection is a standby list. When a slot opens at short notice, a text to a short list of patients who wanted something sooner refills it far more often than waiting does. Build the list from people who asked for an earlier appointment and from treatment already diagnosed but not scheduled.
Track refill rate as its own number. A practice that refills most short-notice cancellations has a very different economic profile from one that does not, and nothing on the website affects it.
One person cannot check out a patient and answer the phone at the same time
In a small practice the biggest conversion problem is usually workload at the desk, not wording on a page.
Watch the front desk for an hour at midday. One person is collecting a copay, answering a question about a claim, scheduling a hygiene visit and watching the phone light up. The call that goes unanswered was not a training failure.
Before rewriting any scripts, look at what can be moved off that desk: online forms completed before arrival, automated reminders, a payment link sent by text, and someone else covering the phone during checkout peaks.
Reserving a couple of new-patient slots each day is a related fix. If your first available is three weeks out, the answer given on the phone kills the booking regardless of how well the call is handled.
Then read the whole path as separate numbers: inquiries received, appointments booked, appointments kept, treatment accepted. They fail for different reasons and averaging them hides which one is broken.
Grade the phones by shift, because your variance is a person, not a page
One practice in one city will never generate enough traffic for small-effect split tests, so the readable experiments are in the intake.
Split testing a button color needs volume no single dental practice has. Trying anyway produces numbers that look like results and are not, and decisions made on them are coin flips with a report attached.
Grade calls instead. Pull twenty recorded calls, score them on the same short rubric, and you will learn more in an afternoon than a quarter of page tests would tell you. Do it again in a month with the same rubric.
Test big, structural things rather than small ones: adding evening hours, changing who covers lunch, verifying benefits live, moving from a callback promise to a booked time. Changes of that size are readable in the schedule without statistics.
Give each change a fair run and change one at a time. If a practice alters hours, scripts and the website in the same month, nobody will ever know which one worked, and the next decision gets made on a guess.
Questions we actually get
- Where do most New York dental practices lose inquiries?
- In our experience of the trade the recurring leaks are unanswered calls at midday and around opening and closing, insurance questions the desk cannot answer on the spot, a first available appointment that is too far out, and first-time patients getting lost between the street and the suite. None of those are website problems, and all of them are cheaper to fix than buying more traffic.
- Is it worth split testing our website?
- Rarely at the scale of one practice. Small-effect tests need traffic volumes a single office does not produce, and running them anyway yields numbers that feel like evidence but are not. Test large structural changes instead, such as adding evening hours, live benefits verification or lunch coverage on the phones, and read the result in booked and kept appointments over a few months.
- What should we measure first?
- Four numbers kept separately: inquiries received, appointments booked, appointments kept, and treatment accepted. Then break booked and kept down by hour of day and by where the patient is coming from. Averaging everything into one conversion rate hides which stage is failing, and the stages fail for completely different reasons.
- Should we charge for late cancellations?
- Having a policy is reasonable. Applying it rigidly in a city where a stalled train can make a punctual person twenty minutes late tends to cost more in reviews than it recovers in fees. Give the desk written discretion so it is applied consistently, and put your effort into a standby list that refills short-notice openings, which is worth more than the fee ever will be.
- How long before intake changes show up in the numbers?
- We will not put a date on it. Answer rate and booking rate respond quickly because they are measured at the moment of the call. Kept appointments and accepted treatment lag by weeks because of how far out the schedule runs, so give any change at least a full scheduling cycle before judging it, and avoid changing several things at once.
