Ask every caller where they are driving from, and write the answer down
Origin is the single most predictive field you are probably not capturing, and it costs one sentence to collect.
Front desks ask for a name, a phone number and an insurance carrier. Almost none ask where the caller is coming from, which means nobody can later explain why show rates differ between two campaigns that looked identical.
Add the question to the script early and naturally. It helps the caller too, because it lets your team offer the office that is actually closer and warn them about the drive honestly rather than discovering it later.
Store it as a field, not as a note. A ZIP code in a structured field can be reported on. The same information typed into a free-text comment might as well not exist.
Within a couple of months you will have something genuinely useful: booking and show rates by origin. Practices are routinely surprised by which areas produce patients who come back and which produce a single visit.
The cancellation two days later is usually a drive, not a scheduling mistake
Late cancellations cluster by geography in a metro where crossings are limited, and treating them as random noise hides the pattern.
Someone in Dunedin books enthusiastically on a Sunday evening. On Tuesday they map the route to your Brandon office at four in the afternoon, and the appointment quietly disappears. The front desk logs it as a cancellation with no reason given.
Start recording a reason code, even a rough one. Distance, cost, plan, timing, changed mind. Three months of coded cancellations tells you more than any heat map of your website ever will.
When distance is the driver, the fix is upstream. Stop advertising into areas your book shows do not stick, and be straight with patients about the drive before they commit. A declined booking costs less than a held slot that empties.
For the crossings you do want to serve, schedule around traffic. Offering a Pinellas patient a mid-morning slot rather than a late-afternoon one is a conversion improvement disguised as a scheduling preference.
A household on a transfer clock books with whoever answers first
Relocating families around MacDill decide on a compressed timeline, and speed of response beats almost everything else you could improve.
Someone sorting out a household after a move is working through a long list. They will call two or three practices in one sitting and book with the first one that gives them a real appointment time.
Measure time to first human contact, not time to first response. An automated acknowledgment is not contact. Track the minutes between an inquiry arriving and a person speaking to the patient, and look at it by hour of day.
Cover the gaps you already know about. Lunch, the hour after closing, and weekends are when inquiries arrive and when nobody is watching. Whether that is answered by a service, a rotation or a callback queue matters less than someone owning it.
Have the record request ready as a form your team sends, not as homework for the patient. Removing that task from a person who is already overwhelmed converts appointments that would otherwise never be made.
Every missed call should text back before the caller opens the next tab
The recovery window on an unanswered dental call is minutes, and an automatic text is the cheapest intervention available.
Missed calls are usually the largest single leak in a practice, and the least visible, because nothing is logged when a phone rings out. Pull the call report before assuming your number is small.
An immediate automated text, sent from the same number, does most of the recovery work. Keep it brief, name the practice, offer two concrete times or a link, and make it repliable. Confirm your messaging setup meets consent requirements before switching it on.
Rank your callback queue by inquiry type rather than by arrival order. A caller in pain and a caller asking about whitening are not the same urgency, and treating them identically loses the one that would have booked today.
Grade a sample of recorded calls each month against a short checklist: were two appointment times offered, was the plan question answered, was the call ended with something booked. Coaching against real recordings changes behavior in a way memos do not.
Two offices on opposite sides of the bay are not a control and a variant
A single practice in a single metro rarely has the traffic for small-effect testing, and comparing two locations across the water is not a substitute.
Splitting a test between a Hillsborough office and a Pinellas one compares two different populations with different competitors, different plan mixes and different traffic patterns. Any difference you observe is confounded before you begin.
Spend your effort on changes large enough to see without statistics. A phone answered at seven instead of five. New patient slots held open each week. A first-visit price published where it was previously absent. Effects of that size show up in the booking numbers without a testing tool.
If you do test, test the script and the offer rather than page furniture. Button colors need volume that one practice will never have. Whether the front desk offers two specific times or asks what works for you is a change you can hear in a week of call recordings.
Use before-and-after periods carefully. Compare like months, note what else changed, and be honest about seasonality in the patient book. A change made in a slow month can look brilliant for reasons that have nothing to do with it.
The cheapest booking this month is a patient you already treated
Reactivation converts at a rate no acquisition channel matches, and most practices have a list sitting unworked in their software.
Run the report for patients with no visit in the past eighteen months and no scheduled appointment. In most practices the number is larger than the owner expects, and a share of them simply drifted rather than left.
Contact them like people rather than like a campaign. A short message from the practice, naming the last visit and offering two times, outperforms a discount blast. Some will have moved away, which in this metro is a real and frequent outcome, and that is useful information too.
Fix the leak that created the list. Most lapses start with an appointment that was never rebooked before the patient walked out. Rebooking at the chair, before checkout, is a conversion step that costs nothing.
Track reactivations as a separate line in your reporting. Rolling them into new-patient numbers hides both how cheap they are and how much acquisition spend they are quietly offsetting.
Questions we actually get
- Where do we start if we can only fix one thing?
- Missed calls. Pull the report, find out how many go unanswered and when, and close the largest gap. It is usually the biggest leak in the practice and the cheapest to address.
- How many calls should we listen to before changing anything?
- Enough to hear patterns rather than anecdotes, and from a spread of days and times rather than one busy afternoon. Grade them against the same short checklist each time so the results are comparable month to month.
- Is online booking worth adding?
- It helps when it writes into the real schedule with genuinely available slots. It hurts when it collects requests nobody works quickly. Decide who owns the queue before switching it on.
- Should we discount to win price shoppers?
- Generally the wrong lever. Discounts attract patients who leave for the next offer and they compress the value of the treatment you actually want. Availability, clear pricing and a fast human reply convert better and cost less.
- Can we A/B test our website?
- With one practice in one metro, only for changes with large effects, and even then carefully. Script changes, hours and offers are testable in weeks. Small page tweaks would need traffic a single practice will not have, so measure them differently or leave them alone.
