You have $1.3M sitting in patients who stopped coming.
A 3-op practice carries around 1,600 patients who have not been in for a year or more. Your front desk will never call all of them. So nobody calls any of them. Recall Engine scores every one, ranks them by what they are actually worth, and runs the outreach until they book.
The demo is the real product running on a synthetic 260-patient cohort. No signup.
| # | Patient | Lapsed | Why they score | Return prob. | Expected $ | Tier |
|---|---|---|---|---|---|---|
| 1 | Jennifer O. 8 prior visits, spouse still active |
13 mo | $5,200 diagnosed, never scheduled | 81% | $2,751 | Priority |
| 2 | Donna P. 12 prior visits, $1,100 benefits expiring |
13 mo | $5,200 treatment plan open | 72% | $2,444 | Priority |
| 3 | Patricia Z. 11 prior visits, no reason on file |
13 mo | $5,200 treatment plan open | 70% | $2,375 | Priority |
| 260 | Barbara H. Moved 162 mi away, no reply in 3 years |
44 mo | Do not spend front desk minutes here | 0.1% | $2 | Do not chase |
Reactivation campaigns fail because they treat every lapsed patient the same.
Your recall software can send 1,600 texts on Tuesday. That is not the hard part. The hard part is that a patient who moved to Arizona and a patient whose husband still comes in every six months are not the same patient, and blasting them the same message gets you a 2-4% response and a pile of opt-outs that damages the list you actually needed.
That last number is the whole product. The value is not evenly spread. It is concentrated in a few hundred patients who have unused benefits, a treatment plan they never scheduled, a family member still on the schedule, and no reason on file for leaving. Find those first, and the same front desk hours produce four times the revenue.
Score. Rank. Sequence. Stop when the chair is full.
It runs on your PMS export. No new software for the team to learn, and no rip-and-replace of the recall tool you already pay for.
Score every lapsed patient
A logistic model over 15 chart features: months lapsed, visit history, no-shows, unused insurance benefits, outstanding balance, diagnosed-but-unscheduled treatment, whether anyone else in the household is still active, distance, and the reason they left. Every score shows its own math - no black box for you or your team to argue with.
Rank by dollars, not by date
Probability alone is the wrong sort. A 40% chance at a $3,400 treatment plan beats a 70% chance at a cleaning. The list is ordered by probability times expected first-year production, then cut at whatever capacity your front desk actually has this month.
Run the sequence
Five touches over nineteen days: SMS, email, and a call task that only lands on your front desk when the patient is worth the four minutes. Copy is drafted per patient from their own chart facts. Everything below the cut line gets suppressed, which protects your deliverability.
| Your recall software | Recall Engine | |
|---|---|---|
| Who gets contacted | Everyone overdue | Ranked by expected recovered revenue |
| Message | One template, merge the first name | Drafted from that patient's chart facts |
| Front desk call list | Alphabetical, or none | Only patients above the $260 value bar |
| Patients who will never return | Messaged anyway, opt out, hurt the list | Suppressed before send |
| Reporting | Messages delivered | Production booked, by patient, by touch |
| Gets better over time | No | Refit monthly on your booked / not-booked outcomes |
One price. It either pays for itself in the first quarter or you stop paying.
- PMS export mapped and scored - Dentrix, Eaglesoft, Open Dental, Curve
- Model fit to your practice, then refit monthly on your own booked outcomes
- The 5-touch sequence built and running - SMS, email, and front desk call tasks
- Per-patient copy drafted from chart facts, reviewed by you before anything sends
- Weekly work list to the front desk, sized to the capacity you actually have
- Monthly production report - booked patients, by touch, in dollars
- Month to month after the setup. Thirty days notice, no term.
The math on a 3-op practice
Break-even sits at 6.6% reactivation. Below that this loses you money, and I will tell you before we start if I do not think your list clears it. The $863 figure is deliberately below the median patient value in our own demo cohort. Payback: 146 days.
The questions every dentist asks on the first call.
We already pay for Weave / RevenueWell / Solutionreach. Why add this?
Keep them. Those are delivery tools and they are good at delivery. This is the targeting layer that sits in front of them: it decides who is worth contacting, in what order, and what the message should say. Most practices are paying for a megaphone and using it to shout at a list nobody has ever ranked.
If your current tool is producing more than 8% reactivation on your full lapsed list, do not buy this. That is a real answer, and I will check your numbers before quoting.
My front desk already does recall calls.
They do, and they call the same forty people, because that is what fits between patients. The question is not whether calls happen. It is whether the forty they call are the forty worth the most.
In the demo cohort, the top 31% of the list holds 70% of the recoverable value. If your team's forty calls come off the top of a ranked list instead of the top of an alphabetical one, the same labor produces several times the production. The engine also strips out the patients who will never come back, which is usually a quarter of the list.
These patients left for a reason. Is chasing them going to annoy people?
Some did leave for a reason, and those are exactly the ones this suppresses. Patients who moved out of the area, who have a service complaint on file, or who have never once responded to the practice get scored down and never contacted. That is a feature: every message to somebody who will never return costs you an opt-out on a channel you need.
The majority of your lapsed list did not leave over anything. They moved, changed jobs, had a scheduling conflict, and then eighteen months went by. Those people are not annoyed to hear from their dentist. They are the ones with unused benefits sitting on the table.
Is this HIPAA-safe? Where does patient data go?
The scoring runs on models hosted on my own hardware, not a third-party API, so chart data does not leave infrastructure covered by our BAA. A signed BAA is part of setup, before any export moves. The demo on this site runs on a synthetic cohort with no PHI in it at all.
Outbound copy is reviewed and approved by your team before the first send. Nothing goes to a patient that a human at your practice has not signed off on.
What do you need from us, and how long does setup take?
One export from your PMS and about ninety minutes of a practice manager's time. Ten business days from export to the first work list on the front desk. You approve the message templates before send, and you set the weekly capacity, so the team is never handed a list longer than the hours they have.
What if it does not work?
The setup fee covers real work and is not refundable. The monthly is month to month with thirty days notice and no term, so the most you can be wrong by is one month. I report production booked in dollars, by patient, every month, which means you can tell whether it works without taking my word for it.
Send me one export. I will tell you what your dormant list is worth.
A 20-minute teardown on your real numbers: how many patients are lapsed, how the value is distributed, and what the top 200 are worth. You get the ranked list either way. If the math does not clear break-even, I will say so on the call.
Mike Rodgers - I build and run these myself. You are not talking to an account manager.