What a Front Desk Actually Costs, Beyond the Salary

Ask a practice owner what their front desk costs and you get a salary. It is the number on the payroll report, it is the number in the budget line, and it is the number every staffing conversation gets anchored to. It is also the least interesting of the four figures that matter, because it is the only one that does not move when the phone gets busier.

The other three are harder to see and larger in combination. None of them require a consultant to work out. All of them come from data you already have, and the point of assembling them is not to argue for cutting anyone. It is that any decision about phone coverage, extra hours, or software gets compared against the wrong number until you have all four.

The salary is the number you already know#

A front-desk salary is public information and the easiest of the four figures to establish. You already know yours exactly, which is the point of starting here: it is the one number in this article you do not have to go and calculate, and it is the number every other cost gets quietly compared against.

For a sanity check against the market, national wage data puts the median for medical secretaries and administrative assistants at $22.08 an hour, or $45,930 a year, as of 2025 [1]. That is worth knowing if you are modelling a role you have not filled yet, or checking whether your own number sits above or below where candidates in your area expect it to be.

Treat this as the anchor and nothing more. It is the only one of the four numbers that a payroll report gives you directly, and it is the only one that stays flat whether the phone rings 40 times a day or 400. Everything expensive about a front desk is downstream of that ringing.

Two adjustments make the anchor honest before you go further. Add employer taxes and benefits, which a practice manager can pull in a few minutes. Then divide by the hours that are genuinely spent on the phone rather than on rooming, intake, checkout, and everything else the role absorbs, because otherwise you are attributing the whole salary to one of its several jobs.

Salary is the only one of the four numbers that does not move when the phone gets busier.

Turnover is the number that moves the most#

Replacing a front-desk hire costs recruiting time, onboarding time, and a stretch of reduced output while the new person learns your practice. It is the largest single swing in the four, and unlike salary it varies enormously between practices in the same town.

The good news is that the market has stopped getting worse. In MGMA polling, around 70% of practices reported turnover the same as or lower than the prior year, split evenly between the two, while 29% said it had risen [2]. Stabilised is not the same as solved, and front office remains one of the roles most often named as a hotspot.

Recruiting difficulty tells the same story from another angle. Asked which role is hardest to recruit, 47% of practice leaders named medical assistants, roughly triple the 15% who named nurses, ahead of billers at 10% and coders at 9% [2]. Front-desk and clinical support roles compete for overlapping candidates in most markets, so a hard MA market is rarely an easy receptionist market.

What practices actually do about it is worth knowing before you assume money fixes it. Asked for their most effective retention tactic for front-desk staff specifically, leaders split evenly: 33% said higher pay and 33% said engagement [3]. That is a genuinely even split, and it means half the practices that solved this did not solve it by paying more.

The number you want for your own model is simple. Take the months a front-desk seat sat empty or half-productive in the last two years, and multiply by what that seat produces in a month. Most practices have never calculated it and are startled by the result.

Coverage is a structural number, not a scheduling one#

A week has 168 hours. A practice open eight hours a day, five days a week, is staffed for 40 of them. That is not a rounding error around the edges of the business, it is 24% of the week, and calls keep arriving through the other 128 hours whether anyone is there or not.

This is the number that resists being solved by hiring, which is what makes it structural. Adding Saturday mornings adds four hours. Adding a late clinic one evening a week adds three. Both are real operational commitments and neither meaningfully changes the shape of the week.

The reason it belongs in a cost model rather than a scheduling conversation is that the uncovered hours have a price even though nothing appears on a payroll report for them. Whatever share of your inbound demand arrives in those 128 hours is being handled by voicemail, by a callback the next morning, or by nobody.

Sizing it does not need new software. Your phone system already records call times, and one export sorted into open hours and closed hours gives you the split for your specific practice. The worked version of that exercise takes an afternoon and produces a number you can defend.

Attention is the cost nobody puts a figure on#

The fourth cost is the one practice owners describe constantly and quantify never. A receptionist handling a checkout while the phone rings is doing two jobs badly at once, and the patient standing at the desk is usually the first person to notice it happening. Nobody bills for that, and no report records it.

It shows up as small failures rather than a line item. A caller who waits on hold and hangs up. A message taken accurately but never actioned because the person who took it got pulled into something else. A recall list nobody has worked in three weeks because there was never a quiet twenty minutes to work it.

What makes this cost slippery is that every individual instance looks trivial and forgivable. One hold that ran long is not a problem. One message that sat until Thursday is not a problem. The cost lives in the rate rather than in any single event, which is exactly why it never appears on a report and why the people closest to it describe it as a feeling rather than a figure.

Patients have a view on this that is worth taking seriously, because it cuts against the usual assumption. In a 2024 survey of 1,000 US adults, 42% said they were comfortable with AI scheduling routine appointments, and among patients dealing with sensitive health issues, 67% said they would be more comfortable booking through an online chatbot than with a person [4]. The same survey found 81% still wanted a human for actual medical advice, which is the line worth respecting.

Half the practices that fixed front-desk retention did not fix it by paying more.

The practical version of this cost is a question rather than a formula: how many times a day does your front desk get interrupted mid-task, and what is the failure rate on the task that got interrupted? Nobody has a source for that. You can watch it for a week.

What each option is actually being compared against#

Once the four numbers exist, the options in front of a practice stop looking like they belong in the same category. A part-time hire, an answering service, and an automated front office solve different subsets of the four, and a comparison that treats them as three prices for one thing is the most common way this decision goes wrong.

A part-time evening hire moves the coverage number and nothing else. It adds staffed hours, which is real, but it also adds a second exposure to the turnover number and does nothing about interruption during the day, because the interruption happens at 11am when the new person is not there. Practices frequently discover the fourth cost is the one they most wanted fixed.

An answering service moves coverage and leaves the rest alone. Calls get picked up outside hours and come back as messages, which means the work still lands on the same desk the following morning, competing with the same walk-ins. It converts a missed call into a task, and a task is not the same as a resolution.

Automated coverage moves coverage and interruption together, because a call that never reaches the desk is not an interruption at the desk. What it does not move is turnover, and it does not remove the need for a person, since anything clinical or ambiguous still has to reach one. Any vendor implying otherwise is describing a product that does not exist.

The reason to lay the three side by side is that the honest answer for a lot of practices is a combination, or nothing at all. A single-location practice with light evening demand and a stable front desk has no problem worth spending on. A practice with 128 uncovered hours, a seat that has turned over twice, and a desk that gets interrupted forty times a day has three separate problems, and buying one solution for all three will disappoint on two of them.

Putting the four together#

None of the four numbers is hard to produce, and none of them requires software to establish. Two come out of records you already keep, one needs a single phone-system export, and the last needs a week of paying attention. Assembled, they answer a different question from the one a salary line answers on its own.

The reason to assemble all four is not to build a case for replacing anybody. It is that three of them move with call volume and the one everyone quotes does not, so a practice that only knows its salary line is comparing every coverage decision against the one number that stays still.

It is also worth being honest about the precision available here. Two of the four are estimates and will stay estimates, because no practice runs a controlled experiment on its own vacancy months. A range you can defend beats a decimal you invented, and the decision this feeds is usually directional rather than fine-grained.

That comparison is where most decisions in this category actually go wrong. A practice weighing an evening receptionist, an answering service, or an AI front office against a $45,930 salary line is answering a different question from the one it thinks it is asking. Against all four numbers, the arithmetic frequently reverses.

None of this says the answer is software. Some practices work the four numbers and conclude their coverage is fine, or that the fix is a rota change, or that the real problem is a recall list rather than a phone. That is a legitimate outcome and a cheaper one. What is not legitimate is deciding either way on a salary line alone.

If you want the shorter version of this exercise aimed specifically at the calls you are missing, the cost of a missed call works the same four-number logic in the other direction.