What do missed dental calls really cost? Build a model you can defend

1. Build a call cohort you can explain

Choose a period that represents the problem you are evaluating. That might be the last four ordinary working weeks, not a holiday week or an unusually successful marketing campaign. State the dates, locations, call windows, and definition of an unanswered call. Keep after-hours calls separate if the practice handles them differently.

Deduplicate repeat attempts using the practice’s approved reporting process. Do not assume that three rings from one caller represent three people. Then classify the request where the available information supports it: new-patient booking, existing-patient scheduling, clinical or billing question, supplier or spam, or unknown.

Keep the unknown category. If there is no reliable reason-for-call information, do not assign all unknown calls to new-patient demand. Investigate a sample through an approved process or model a range. Your internal analysis should not require copying identifiable patient data into a public calculator or an informal worksheet.

Finally, identify what happened after the missed call. A person may already have booked through a callback or online. That booking belongs in the baseline process, not in the number a new tool can claim to recover. Existing-patient requests still matter, but their value should be modeled separately from new-patient acquisition.

2. Estimate incremental bookings, not all bookings

Suppose the current callback process books 15 of those 50 new-patient inquiries. That is a 30% baseline booking rate for this particular cohort. If a pilot books 25 comparable inquiries, the pilot rate is 50%. The difference is 10 additional bookings, not 25 recovered patients.

A before-and-after comparison is only useful when the definitions and cohorts are comparable. A change in marketing, appointment availability, hours, staffing, or patient mix can affect bookings. Record these differences. Where practical, your team can compare similar call windows or use an agreed staged rollout, while recognizing that this still may not isolate causation.

Count a booking only after it is confirmed in the scheduling system. Do not treat a captured name, a requested appointment, or an unanswered transfer as a completed booking. Carry unresolved outcomes separately until they are reconciled.

3. Follow the booking through attendance and contribution

A booking does not necessarily become an attended visit or a collected payment. For the illustration, apply an 80% attendance rate to 10 additional bookings. The result is eight expected additional attended visits. Fractional expected values are acceptable in a model; actual operational reports should still show the underlying whole counts.

Now choose a financial measure and label it. Collected revenue is different from scheduled production, billed charges, or contribution after the variable costs of providing care. In this example, assume $150 of contribution per additional attended visit within the chosen measurement period. Eight visits then produce $1,200 of modeled contribution before the incremental administrative costs listed below.

Use your practice’s accounting definitions for the real calculation. If a staff-time cost is already included in contribution, do not subtract the same cost again. If the practice must add sessions, cover overtime, or buy supplies to serve additional visits, include the relevant incremental cost once and show where it enters the model.

Do not value visits you cannot accommodate. A schedule with no appropriate capacity may convert a faster response into a longer waiting list rather than near-term contribution. Test a capacity ceiling and report how many additional bookings can actually be served in the period.

4. Compare like-for-like costs

A software subscription is not the entire cost of an AI-supported process. Include usage charges, integration work, setup allocation, quality review, exception handling, staff training, and any additional phone-service cost. Use a written quote and the expected scope; do not assume every AI service has unlimited usage or free onboarding.

The same principle applies to staffing. Compare the tasks and hours that would actually change, not the annual cost of a complete front-desk role against a narrow phone workflow. Time redirected to in-person patients can be valuable without being a cash saving. Call it released capacity unless it demonstrably reduces a cost.

The Bureau of Labor Statistics Occupational Outlook Handbook provides occupation-level wage context for receptionists, not a fully loaded quote for a specific dental practice. Local pay, working hours, benefits, and coverage requirements still need to come from your own hiring or payroll information. A national median should not be presented as a precise dental staffing budget.

Source: U.S. Bureau of Labor Statistics

5. Show the downside, not just a headline return

The example becomes much more useful when you vary the uncertain input. Keep 50 eligible inquiries, a 30% baseline booking rate, 80% attendance, $150 contribution per visit, and $900 monthly cost fixed. Then test several pilot booking rates rather than presenting the most optimistic one as inevitable.

At a 40% pilot rate, the improvement is five bookings and four expected attended visits. Their $600 contribution does not cover the $900 cost. At 50%, the modeled net is $300. At 60%, it is $900. That spread tells the team which part of the process needs validation before making a financial claim.

The break-even condition in this simplified example is 7.5 additional bookings, or 15 percentage points of improvement across 50 inquiries, before rounding operational targets to whole bookings. This assumes attendance, contribution, costs, and capacity remain as stated. Change any of those assumptions and the break-even point changes too.

6. Keep lifetime value out of the annual cash-flow headline

Lifetime value can be useful for a separate acquisition decision, but its time horizon must stay visible. If a patient’s estimated value includes years of recall and treatment, multiplying that value by a year’s new-patient cohort does not mean the practice collects all that money this year.

Build a near-term case from the visits and contribution expected within the measurement period. If you also model lifetime value, label the retention assumptions, time horizon, future costs, and uncertainty. Keep that result in a separate section rather than adding it to near-term contribution and counting the same relationship twice.

Use the practice’s actual operating days. A weekday-only clinic should not multiply daily figures by 360 without explaining why the cohort includes those days. Similarly, do not mix all-day missed-call counts with a service that covers only one call window.

7. Turn the model into a decision record

Use the downloadable worksheet to name the owner and source for every input. Mark each one as observed, estimated, or not yet known. Keep the original baseline so the team can revisit the case after the pilot instead of reconstructing what it once believed.

Review the financial result alongside booking accuracy, unresolved handoffs, caller experience, and staff correction time. A positive modeled contribution is not sufficient if the process creates unsafe or unreliable outcomes. Equally, a pilot may produce operational value that has not yet converted into collected revenue; report that honestly.

When discussing an Ethon pilot, bring a description of the workflow and aggregate counts—not patient details. Agree what the implementation will handle, what it will cost, and how its results will be measured. No calculator can replace those decisions, but a transparent model makes them easier to test.