Guides

What a missed patient call actually costs your clinic

Last reviewed September 2026

A missed call leaves no trace. A cancellation leaves a visible hole in the schedule; a call that rings out leaves nothing at all, which is why most clinics have no idea what the phone is costing them. This guide covers where the calls actually go, why so few people leave a voicemail, how to count your own misses in an afternoon, and how to do the money arithmetic with your assumptions written down rather than borrowed from someone selling you something.

The call nobody remembers

Someone rings your clinic at twenty to three on a Tuesday. Your front-desk person is checking a patient in, the second line is already going, and the phone rings out. That is the entire event. No alert, no flag, no note in a chart. By Friday nobody at the clinic knows it happened.

That is what makes missed calls hard to manage rather than merely expensive. A cancellation leaves a visible hole in the schedule. A missed call leaves nothing behind at all. You cannot go looking for a patient you never knew had rung.

Where the calls actually go

"Missed call" covers four different failures, and they want four different fixes, so it is worth separating them before you do anything else.

The phone rings out and nobody picks up. The call diverts to voicemail. The caller waits in a hold queue and hangs up before anyone answers. Or somebody does answer, parks the caller on hold to deal with the person standing at the desk, and the caller gives up.

The fourth one is the cruel one, because your phone system records it as an answered call. If you judge the desk on answer rate alone you will never see it, and it is the failure your staff are most likely to be pushed into on a busy morning.

Then there is a fifth destination, which is not a failure of the desk at all: the call arrives when the clinic is shut. A clinic that closes at five on Friday and opens at nine on Monday has its front door locked for sixty-four hours straight — more than a full working week — across exactly the window in which a person with a job can comfortably make a phone call.

Why so few people leave a voicemail

You will find a confident figure for this everywhere: eighty per cent of callers sent to voicemail leave no message. It is worth knowing where that number comes from before you plan around it. The trail leads to a 2014 CRM Magazine article, which attributes the eighty per cent to Forbes without naming any underlying study — and which, a few paragraphs later, quotes the chief executive of a phone-support company putting the rate at somewhere between fifty and seventy-five per cent depending on the type of business. A number that moves by thirty points inside a single article is not a measurement. It is a talking point, and it is usually pointed at your wallet.

What you can reason about without a study is the behaviour, and the behaviour is not mysterious. A voicemail is a one-way message with an unknown reply time, left with a clinic the caller may never have spoken to. They have no idea whether the box is checked hourly or on Monday morning. If they are in pain and they have three clinics open in three browser tabs, leaving a message means choosing to wait while the other two options are one tap away. Most people do not choose that. They dial the next number.

That is also why "did they leave a message?" matters less than it looks. A voicemail returned within the hour is a save. A voicemail returned the next afternoon is often a patient who has already been seen elsewhere and is too polite to say so.

How many of your own calls go to voicemail

Stop hunting for the industry number and get yours. It takes an afternoon and costs nothing.

Export a full month of call detail records. Every VoIP provider and most telcos give you a downloadable log with timestamps, duration and disposition. Count the calls with zero talk time, the ones that ended in voicemail, and the ones abandoned in a queue. That is your floor, and it is a real number about your clinic instead of somebody else's.

Then split it by hour and by weekday and look for clusters. Check the first hour of the day, the lunch hour, the half hour after your busiest arrival block, and everything outside opening hours. If your misses pile up in one predictable window, the fix may be a shift pattern rather than a purchase.

Finally, listen to the voicemails you did get and tag them: new patient, existing patient rebooking, clinical question, admin or billing. Now you know what share of your phone traffic is actually revenue. It is usually a smaller share than people assume, and that cuts both ways — it means the frightening annual figures overstate your loss, and it means the small slice that does matter deserves far more protection than it is currently getting.

Two things are worth watching beyond answer rate, because speed alone is not the goal. When researchers looked at telephone access across US Veterans Health Administration primary care, the system mandated an average speed of answer of thirty seconds or less and an abandonment rate under five per cent — demanding targets — but self-reported performance against those metrics was not related to how patients rated their access. What did track with better-rated access was how well the people answering the phone communicated with the clinical teams behind them. And a Journal of the Royal Society of Medicine paper on primary care access argues for measuring abandoned calls, time from first contact to clinical resolution, and repeat contacts for the same problem, rather than counting appointments. The practical version for your clinic: a fast answer that resolves nothing produces a second call, and your volume goes up rather than down.

Why a missed new patient costs far more than a missed existing one

The cost of missed calls is not spread evenly across them. Most of them cost you very little. A few cost you a great deal, and the dividing line is almost always new versus existing.

An existing patient trying to move Thursday's appointment has a relationship with you, your number in their phone, and usually no urgency. They try again. A new patient with a sore back has no relationship with you at all, is ringing three clinics, and is not waiting for a call back from the first one.

So "we miss about a fifth of our calls" is not a useful sentence. A fifth of which calls?

A worked example, with every assumption on the table

Here is the money reasoning done transparently. Every input below is an assumption, not a fact about clinics generally — swap in your own and redo the arithmetic. The method is the point.

Suppose your average visit fee is $90, and suppose a typical course of care at your clinic runs six visits. Then one new patient is worth 6 x $90 = $540 over that course of care. If you want to count a second year, a maintenance cadence or referrals, add them — but write the assumption down, because unwritten assumptions are where overstated estimates come from.

Now suppose six in ten new callers who hit voicemail book somewhere else and never come back. The expected cost of one missed new-patient call is then 0.6 x $540 = $324.

Now the existing patient. Suppose one in four existing patients who cannot get through drifts — books elsewhere, forgets, lets it go — and suppose you count only the single visit they were ringing about. The expected cost of one missed existing-patient call is 0.25 x $90 = $22.50.

Under those assumptions a missed new-patient call is worth about fourteen times a missed existing-patient call. Change the assumptions and the multiple moves; the direction does not.

Scale it. Suppose two new-patient calls a week go unanswered: 2 x 52 x $324 = about $33,700 a year. If your real permanent-loss rate is three in ten rather than six in ten, the same sum gives about $16,800. If your course of care runs ten visits rather than six, it gives about $56,200. The spread across perfectly plausible assumptions is enormous, which is exactly why no stranger's web page can tell you your number.

In one line: missed new-patient calls per week x 52 x the share who never come back x the value of a course of care.

Two qualifications that most pages on this subject leave out. First, this is expected value across a population of callers, not cash sitting somewhere waiting to be collected. Treat it as a forecast, not an invoice. Second, if you are already booked solid six weeks out, an extra new patient displaces an existing one — fixing the phone then changes your patient mix and your waiting list rather than your revenue. Still worth doing, because new patients renew a practice, but it is not found money.

The after-hours window

There is decent Canadian evidence that a lot of demand is arriving when clinics are closed. In the Commonwealth Fund's 2023 international survey, reported by CIHI, only 26% of Canadian adults were able to get a same- or next-day appointment, against a 42% average across peer countries. Less than a quarter — 23% — found it easy to get care in the evenings, at weekends or on holidays without going to an emergency department, against a 32% peer average.

Whatever that says about the system, the operational reading is simple. People are trying to reach care outside business hours and mostly failing. Those sixty-four dead hours between Friday and Monday are not quiet because nobody wants anything.

What to do about it

Tighten how the desk answers. Move the phone out of the treatment room. Put a second person on the desk for the first hour and the lunch hour. Stop using hold as a parking space — if you cannot deal with the caller now, take a number and say when you will ring back.

Set a call-back rule and hold yourself to it. Voicemails returned within the hour save appointments; voicemails returned tomorrow mostly do not. Give one named person the box, and give them a deadline rather than a vague instruction.

Take booking off the phone entirely. The cheapest missed call is the one that never gets dialled. Real online booking — where the patient picks a slot and it lands in your schedule — removes call volume rather than answering it. A request form that generates a callback task for your front desk does not; it moves the work, and a patient who submits a form and hears nothing for a day behaves much like a patient who got voicemail.

Use a live answering service. A person takes a message and, in some configurations, books into your calendar — ask specifically, because that is the difference that matters. Published pricing varies enormously and is usually quoted per minute rather than as a subscription: Answer United's own pricing guide quotes $1.00 to $3.50 per minute and monthly medical and healthcare plans of $500 to $3,000, while Direct Line Answers publishes per-minute billing of $0.70 to $1.20. Those are two vendors' list prices, not a market rate. Get a written quote for your own call volume and make it include setup fees, holiday premiums and overage, which is where the advertised number tends to come apart.

Hire a part-time receptionist. Worked example, assumptions stated: Job Bank puts the median hourly wage for a medical receptionist in Ontario at $20.50. Suppose twenty hours a week, all year — that is 1,040 hours, or $21,320 in gross wages. Ontario's minimum vacation pay for an employee with under five years' service is 4%, adding $853. Employer CPP is 5.95% on wages above the $3,500 basic exemption, so 0.0595 x ($21,320 - $3,500) = about $1,060. Employer EI is 1.4 times the employee rate of 1.63%, so 2.282% of $21,320, or about $487. Total: roughly $23,700 a year, a little under $2,000 a month — a few dollars more on a real payroll, where vacation pay attracts CPP and EI too, and before statutory holiday pay, training, and the cost of covering their sick days. A good receptionist handles the difficult conversations well — and goes home at five, which is where gema picks up.

Use an AI front desk. This is the category gema is in.

gema Front Desk is $299 CAD a month per clinic location, with no setup fee, no booking commission and no per-practitioner charge. It answers inbound calls around the clock with no hold queue and no voicemail box, and handles SMS, web chat and email on the same line of business. It books, cancels and reschedules directly, because it owns the calendar rather than sitting on top of one, and it sends every booking and cancellation straight into Aspire, Cliniko, Juvonno, IntakeQ/PracticeQ or Practice Better. A caller who wants a person is put through to your office line during your open hours, and flagged for your team after hours. The plan includes a pool of credits every month, shared by SMS, email and chat as well as voice — a pool, not an unlimited allowance, and usage past it is billed monthly at the published rate. Records are stored in Canada, and it is built for PHIPA and PIPEDA. If what you actually need is to stop losing bookings rather than to answer the phone, gema Booking at $149 a month gives you online booking and the patient portal that patients use themselves; it does not answer the phone.

When a call needs your team

gema hands off cleanly rather than guessing. During your open hours, a caller who needs a person — someone upset, or something outside a normal booking — is transferred straight to your office line. When you are closed, gema tells the caller the office is closed and flags the call for your team to follow up as soon as you open. Either way, the whole conversation, transcript included, is waiting in your activity log. And at eleven on a Sunday night, a caller who just wants an appointment gets one on the spot — the hour a closed front desk cannot cover at all.

Where to start

Export a month of call records this week and count the unanswered ones. Split them by hour. Tag a sample of voicemails by type. Work out your own new-patient value and your own drift rate, with the assumptions written down where you can argue with them later. Then decide what to change — a shift pattern, a call-back rule, a booking page, a phone setting, a person, or a product. Most clinics find the real problem is narrower and more fixable than the frightening annual figures suggest. That is good news, and it is also why the frightening figures exist.