Plot your inquiries by hour and you will find a second peak nobody staffs
The first useful exercise is not a page redesign, it is a chart of when people try to reach you.
Export your calls, forms and messages for the last ninety days, bucket them by hour and day, and look. Practices in this valley are often surprised. Alongside the expected mid morning peak there is usually a second cluster in the evening, another early in the morning, and a scatter overnight.
That second peak is the round the clock economy showing up in your data. Hospitality, healthcare and casino operations run continuously, so a real share of your market is free at hours your phone is not.
You do not have to staff every hour. You do have to decide what happens in each one: a live answer, an answering service with a script you wrote, or an immediate text back that sets a time.
Do the same chart for missed calls specifically. The overlap between when people call and when nobody picks up is the cheapest problem you will find all year.
The phone menu you added to save time is where new patients quit
An automated attendant costs a practice more new patients than any button color ever will.
Existing patients tolerate a phone tree because they already know you. A new patient with three tabs open does not. Every extra step is a chance to hang up and call the practice on the next corner.
Route new patient calls to a live person as the first option, ideally with a dedicated tracking number so you can hear those calls separately. If your greeting runs longer than a few seconds before offering a human, it is too long.
Then measure hold time and abandon rate, which most phone systems report and almost no practice reads. An abandoned call is a paid click that reached a dead end.
Lunch is the classic gap. If your phones roll to voicemail while the team eats, you are closed during one of the few free hours a shift worker actually has.
Ask what hour you may call back, and then call at that hour
For patients working shifts, a callback at the wrong time is not a second attempt, it is the end of it.
Someone working the floor at a resort cannot take a call at two in the afternoon. They cannot step away, and they will not try you again from the parking lot. The result looks like a bad lead in your report and it was not one.
Add a single field to your form asking the best hour to reach them, and add the same question to your call script for messages taken. Then treat the answer as an appointment your practice has made.
Offer the reply in the channel the patient used. If they texted, text. Chasing a texter with three voice calls is a way of losing them politely.
Track the honored callback rate for a month. It is usually lower than the office believes, and it is fixable by assigning one named person and one time slot per day rather than by buying software.
A nine o'clock appointment for a graveyard worker is a no show you scheduled
Broken appointments in this market are frequently a scheduling mismatch rather than a patient failure.
When your only offer is mid morning, a patient who works nights will take it, mean it, and sleep through it. The practice records a no show and draws the wrong conclusion.
Ask what shift the patient works, or simply what time of day suits them, before offering a slot. Then look at your own broken appointment data broken out by appointment hour and by which submarket the patient drives from, whether that is North Las Vegas, Green Valley or the far side of the 215.
If the pattern is real in your numbers, act on it: hold a block of early and late slots for shift workers rather than filling them with whoever calls first.
Confirmations matter more for this group. A reminder timed to reach someone who is asleep during the day is a reminder that did not happen.
An in house membership plan answers the uninsured caller your script loses
A large service economy means a steady stream of callers with no dental coverage, and most front desk scripts have nothing to offer them.
Listen to ten recorded calls that ended without booking. A predictable share will be someone asking what it costs without insurance, followed by a number, followed by a polite goodbye.
A membership plan gives the front desk a third answer between full fee and nothing. Whether it fits your practice is a business decision involving pricing, terms and regulatory questions worth confirming with your own advisor, and it is not something to launch from a marketing conversation alone.
If you already have one, check whether it appears on the website in plain text, whether the front desk mentions it unprompted, and whether it is offered before the caller asks the price. In most practices the answer to at least one of those is no.
Measure it as a conversion step, not a product. The number that matters is what share of uninsured callers book.
Compare like weeks, because one big event weekend will wreck a week over week read
A single practice in a single metro does not generate enough volume for small tests, so change the method rather than pretending otherwise.
Split testing a headline or a button needs traffic no single dental practice in this valley has. Waiting for statistical significance on a small effect means waiting past the point where the answer is useful.
Test big things instead. New hours. A different first offer on the phone. A membership plan mentioned before price. Live booking replacing a form. Changes large enough to show up in booked appointments without a calculator.
Run each for a defined period and compare like with like. The valley's event and convention calendar makes some weeks unrepresentative, so compare against the same week pattern rather than the week before, and note anything unusual in a log as you go.
Judge on kept appointments and production, not on form fills. And give each change a fair run before replacing it, because a practice that changes three things a month never learns which one worked.
Questions we actually get
- Where do most dental inquiries actually leak?
- In the order we usually find them: calls that ring out or hit a menu, callbacks that never happen at a time the patient can take, insurance questions the front desk cannot answer on the spot, and appointment times offered that do not suit the patient's shift. The website is normally the fourth or fifth problem, not the first.
- What should we measure first?
- Four numbers, kept separate: answered inquiries, booked appointments, kept appointments and treatment accepted. Most practices track one blended figure and cannot tell which stage is failing. Add a split by the hour of the inquiry and by the area the patient comes from, and the picture usually becomes obvious within a month.
- Is A/B testing worth it for one practice?
- For small changes, no. A single location does not produce the traffic needed to read a small difference in a useful timeframe, and anyone promising otherwise is selling reports. Test changes big enough to move booked appointments, run them for a defined period, and compare like weeks so the valley's event calendar does not read as a result.
- Should we answer the phone overnight?
- Not necessarily with a person. The realistic options are an answering service working from your script, an automatic text back that offers times, or online booking that writes into the real schedule. Look at your own overnight inquiry volume before deciding. If the data shows a genuine cluster, one of those three is worth funding.
- How long before intake changes show up?
- Faster than advertising changes, though we will not name a date. Fixing an unanswered lunch hour or an unstaffed evening tends to show in booked appointments quickly because the demand already exists. Changes involving scheduling policy or a membership plan take longer, since they depend on the whole team using them consistently.
