Key takeaways
- Build the ROI from your own after-hours call log, sorted by call type, not from a vendor's estimate of your volume.
- Only some calls create revenue: new-patient inquiries and clinical calls that end in an office visit. In one pediatric call center, 52% of triaged calls ended with a visit recommendation[2].
- Count only the visits you would not have kept anyway; voicemail callbacks rescue some, but only 19% of adults generally answer unknown numbers[3].
- Price the service at your real minutes, monthly minimum included, then test how the answer changes when the capture assumption halves.
Run the numbers on your own calls first
Most answering-service decisions get made on a feeling: the physician is tired of the night phone, or the morning voicemail pile is out of hand. Both are good reasons. Neither tells you whether the service pays for itself, and a vendor's sales sheet will not either, because it does not know your call mix. A month of your own after-hours calls run through a short worksheet does.
Patients do not stop needing care at 5 pm. In 2022, 56.3% of US emergency department visits, about 87.5 million, arrived outside Monday-to-Friday 8 am to 5 pm[1]. Some of those patients called a practice first. The worksheet below asks a narrower question: of the calls your practice gets after hours, how many turn into visits you keep only if somebody answers, and what is that worth against what the coverage costs?
The example practice throughout is fictional: three providers in a suburban family practice, open 8 to 5 on weekdays, with voicemail and an on-call physician's cell number today. Replace every figure with your own as you go. If you have not yet decided whether a small office needs coverage at all, answering service for a small practice covers that question; this article assumes you want the ROI line itself.
Step 1: count and sort a month of after-hours calls
Pull four weeks of inbound calls from the phone system, every call that arrived while the office was closed, including the ones that rang out with no message. Then listen to or read the voicemails and sort each call into one bucket. Hang-ups with no message go in their own row; you will decide how to treat them in step 4.
| Call type | Calls in four weeks | Creates revenue? |
|---|---|---|
| New-patient inquiry | 22 | Yes, if it books |
| Established patient: clinical question | 25 | Sometimes, if it ends in an office visit |
| Appointment change or cancellation | 31 | Protects a visit already booked |
| Refill request | 18 | No, but costs staff time |
| Office information: hours, address, plans accepted | 16 | Indirectly |
| Hang-up, no message | 20 | Unknown |
Two buckets carry most of the money. New-patient inquiries become first visits. Clinical questions from established patients often end with a recommendation to come in: in a Denver pediatric after-hours call center, 21% of 32,968 triaged calls were advised to be seen urgently, 27% the next day and 4% later, while 48% got home-care advice[2]. Your adult or specialty mix will differ, but the pattern holds: a meaningful share of clinical calls turns into a visit, and where that visit happens depends on whether the call was answered.
- Seen urgently 21%
- Seen next day 27%
- Later visit 4%
- Home care 48%
Source: Pediatrics (American Academy of Pediatrics), via PubMed, 2006[2]
Show the data
| Recommendation | Share of calls |
|---|---|
| Seen urgently | 21% |
| Seen next day | 27% |
| Later visit | 4% |
| Home care | 48% |
Step 2: decide which calls you would lose
A call only counts toward ROI if the current setup loses it. New patients are the clearest case. They are often calling several practices, and the one that answers tends to get the booking. In a Texas audit of primary care offices, 72.4% of simulated new-patient callers who reached a scheduler were offered an appointment[4]; a caller who reaches voicemail is offered nothing that night.
Clinical calls lose differently. An unanswered evening call does not disappear; the patient decides alone, and some go to urgent care or the emergency department. That visit is lost to the practice, and it costs the patient far more. After-hours calls that end in the ER works through that side; for this worksheet, leave clinical calls out of the revenue line unless you have evidence of patients going elsewhere. It keeps the result conservative.
Step 3: put a value on a booked visit
Use what you collect, not what you charge. If you do not have your own figure handy, a national anchor is the average cost of an office evaluation and management service for people in large employer plans: $125 in 2021, against $556 in the emergency department[5]. Your payer mix may put you above or below it.
A new patient is rarely worth one visit. Americans averaged 320.7 physician office visits per 100 people in 2019, about half of them to primary care physicians[6]. A new patient who stays a year might reasonably bring two or three visits. The example uses two visits, $250, which undercounts patients who stay for years but avoids pretending every caller becomes a lifelong patient.
Step 4: subtract the visits you keep anyway
This is the step most ROI sheets skip, and it is the one that decides the answer. Some after-hours callers leave a message, get a call back in the morning and book anyway. Those visits are not created by an answering service, so they cannot be counted as its return.
Callbacks connect less often than offices assume. Only 19% of adults say they generally answer calls from unknown numbers[3], and a practice callback at 9 am usually arrives from exactly that. Look at your own last month: of the new-patient voicemails, how many became booked visits within a week? The rest, plus a fair share of the hang-ups, is your recoverable pool. In the example, 35% of new-patient inquiries become bookings that voicemail would have lost. If you have no data at all, run the numbers at 15% and at 35% and see whether the decision changes.
Step 5: price the coverage at your real minutes
Quotes come per minute, per call or as a monthly package, and each one hides a minimum somewhere. Public contracts show the shapes. One bidder on an Arkansas state solicitation priced 24-hour answering at $1.25 for the first minute and $1.15 for each additional minute[7]. A Michigan county health department contract set $895 a month for up to 1,200 minutes, with extra minutes at $0.78[8]. A university nurse triage contract charged $1.22 a minute for non-clinical calls, $19.37 per registered nurse call and a $370 monthly minimum[9]. These are reference points, not quotes for your practice.
Example: 112 calls a month averaging three minutes is 336 minutes. At the Arkansas bid's per-minute rates, that is $3.55 a call, about $400 a month. Under a package with a higher minimum, the same calls might cost $900. Use your actual quote, and read answering service pricing models before comparing two of them, because rounding rules and per-message fees move the total more than the headline rate.
Step 6: the ROI line, with a stress test
Now put it together. Monthly value is the extra bookings times the visit value, plus any staff time the service saves. Monthly ROI is that value minus the coverage cost, divided by the cost. In the example, the morning voicemail routine shrinks when messages arrive as clean summaries; call it three and a half hours a month at the $36.73 hourly total compensation for office and administrative support workers[10], roughly $130.
| Input | Example |
|---|---|
| After-hours calls a week | 26 |
| Share that are new-patient inquiries (%) | 20% |
| Extra bookings voicemail would lose (%) | 35% |
| First-year value of a new patient | $250 |
| After-hours calls a month | 113 |
| From new customers | 23 |
| Bookings if answered | 7.9 |
| Monthly after-hours value | $1,970 |
How it is worked out: After-hours calls a week × 4.33 weeks × new-customer share × booking rate × average value.
Assumptions
- Two office visits at $125, the 2021 average cost of an office evaluation and management service in large employer plans.[5]
- Office visits averaged 320.7 per 100 people in 2019, so two visits in the first year is a conservative figure.[6]
- The 20% new-patient share and 35% extra-booking rate come from the example month in step 1 and step 4; replace both with your own.
Example. Clinical calls, appointment changes and staff time are left out of this calculator, so it shows the new-patient line only.
| Scenario | Extra bookings a month | Visit value | Staff time saved | Coverage cost | Monthly ROI |
|---|---|---|---|---|---|
| Conservative | 3.4 | $844 | $130 | $900 | 8% |
| Middle | 7.9 | $1,970 | $130 | $900 | 133% |
| Per-minute quote | 7.9 | $1,970 | $130 | $400 | 425% |
The stress test is the useful part. In the example, at 35% extra bookings, the service pays for itself several times over at either price. At 15% and the higher package price, it barely breaks even. That is the honest answer for many practices: the decision rests on how many new patients you are really losing at night, which is why step 4 deserves an afternoon with the voicemail log rather than a guess.
Mistakes that flatter the result
Most inflated ROI sheets make the same handful of errors. Check yours against this list before it goes to the partners.
- Counting every after-hours caller as a lost new patient. Most are established patients, and many of them call back in the morning on their own.
- Using charges instead of collections. A visit billed at one figure and paid at another is worth what you collect.
- Counting one person twice. A hang-up at 7:02 and a voicemail at 7:05 from the same number are one caller; match numbers before you total the log.
- Leaving out the morning work. Someone still reads every message, returns the calls and books the visits. A service that sends cleaner messages saves part of that time, not all of it.
- Pricing an average month. Flu season and the week after a holiday weekend are when coverage earns its keep and when per-minute bills peak, so price your busiest month too.
Write the assumptions on the same page as the result. A year from now, when someone asks whether the service is still worth it, you will want to rerun the same worksheet with fresh call data rather than argue about what was assumed.
What changes the answer
- Your call mix. A practice whose after-hours calls are mostly refill requests and appointment changes will struggle to show ROI on revenue alone; the case rests on staff time and the physician's sleep.
- The monthly minimum. Low-volume practices often pay for minutes they never use, which is the single most common reason the math disappoints.
- Who books. If the service only takes messages, every booking still waits for a morning callback, so part of the gain depends on how fast staff call back.
- Holidays and weekends. A long weekend can double a month's after-hours calls; check the quote's holiday terms.
- Caller
- SI Employee
- System
- Your team
01Caller
New patient calls, 7:40 pm
Asks whether you take her plan and have openings for a physical.
02SI Employee
Answered, not voicemail
Confirms new patients and her plan from the list you approved.
03SI Employee
Message taken
Name, callback number and preferred times. No health details.
04System
Summary emailed
Lands in the front-desk inbox with the recording and transcript.
05Your team
Booked at 8:05 am
First callback of the day books the visit in your scheduling system.
Example. The booking itself still happens in your own system when staff call back.
Coverage does not have to be all one thing. An SI Employee can answer the office-question and new-patient share of the after-hours line, from information the practice signs off, while anything patient-specific goes to the on-call or answering-service number you choose. It takes no health details and gives your 911 wording to anyone describing an emergency. That split changes the call profile you ask a clinical service to price; run both halves through the same worksheet. To hear how an after-hours call sounds before you model it, place a test call.
For a larger practice, the orthopedic after-hours line example models the same arithmetic for a five-physician group taking about 85 calls a week outside office hours. And when the quotes arrive, what a medical answering service costs explains each line item.
Frequently asked questions
What is a good ROI for an answering service?
Any result above zero after a conservative stress test. If the service only pays for itself when you assume most new-patient callers would otherwise be lost, treat the case as weak and check your voicemail data first.
Should clinical after-hours calls count toward ROI?
Only if you have evidence that unanswered clinical calls are ending up at urgent care or the emergency department instead of your office. Many do end in a visit recommendation[2], but the safest worksheet leaves them out of the revenue line and treats safety as its own reason.
What visit value should I use?
Your own average collection per visit. Without one, the 2021 average for an office evaluation and management service in large employer plans, $125, is a reasonable anchor[5]; multiply by the visits a new patient usually brings in their first year.
How long should I track calls before deciding?
Four weeks is the minimum, and avoid a holiday month unless you price holidays separately. Eight weeks smooths out a quiet or a flu-heavy month.
References
CDC National Center for Health Statistics (2022). National Hospital Ambulatory Medical Care Survey: 2022 Emergency Department Summary Tables (Table 3) (opens in a new tab) ↑
Pediatrics (American Academy of Pediatrics), via PubMed (2006). How safe is triage by an after-hours telephone call center? (Pediatrics 2006;118(2):457-63) (opens in a new tab) ↑
Pew Research Center (2020). Most Americans don't answer cellphone calls from unknown numbers (opens in a new tab) ↑
JAMA Network Open (via PubMed Central) (2020). Association of Simulated Patient Race/Ethnicity With Scheduling of Primary Care Appointments (JAMA Network Open) (opens in a new tab) ↑
Peterson-KFF Health System Tracker (2021). Outpatient visits billed at increasingly higher levels: implications for health costs (opens in a new tab) ↑
CDC National Center for Health Statistics (2019). FastStats - Physician office visits (opens in a new tab) ↑
Arkansas Department of Human Services (2024). Revised Official Bid Price Sheet: 710-24-058 Telephone Answering Service (opens in a new tab) ↑
Ingham County Board of Commissioners (2013). Resolution #13-190: Resolution to authorize an agreement for call center services at the Health Department (opens in a new tab) ↑
University of Wisconsin–Madison Purchasing Services (Internet Archive copy) (2020). Telephone Nurse Triage Services, contract 20-5393 (contract last updated 02/14/2020) (opens in a new tab) ↑
U.S. Bureau of Labor Statistics (2026). Employer Costs for Employee Compensation, Table 4: private industry workers by occupational and industry group, June 2026 (opens in a new tab) ↑