Every two-provider practice owner eventually stares at the same spreadsheet at 9 p.m.: keep the biller you have, or hand the whole revenue cycle to an outside service that quotes a percentage of what they collect. The pitch decks make it look obvious in both directions. An in-house biller "only" costs a salary. An outsourced service "only" takes a small cut. Both framings are wrong, and the gap between them is where a small practice quietly loses tens of thousands of dollars a year.
The honest way to settle the outsourced medical billing vs in-house cost question is to build both numbers the same way: fully loaded, with every line item that actually leaves your bank account. For a practice running two physicians and roughly $1.2M-$1.5M in annual collections, the two totals land closer together than either vendor wants you to believe. What separates them is not the sticker price. It is who absorbs the risk when a claim gets denied, a biller quits, or the payer changes a rule in the middle of the quarter.
The Salary Line Is the Smallest Part of an In-House Biller
Start with the number you already know. A full-time medical biller or biller/coder in a two-provider primary care practice earns somewhere between $44,000 and $58,000 depending on your market, with $52,000 a fair midpoint. That is the figure that sits on your P&L, and it is the figure that fools most owners into thinking billing is a cheap function.
The payroll load is where the real money hides. Employer-side costs on a W-2 employee run 25% to 40% on top of base wages, and for an administrative healthcare role the middle of that band is realistic:
- FICA and Medicare payroll taxes: roughly 7.65% of wages, about $3,980.
- Health insurance contribution: $6,000-$9,000 if you cover a meaningful share of a single plan.
- Retirement match, if you offer one: 3% is about $1,560.
- Workers' comp, unemployment insurance, and payroll processing: $1,200-$1,800.
- Paid time off: two weeks plus holidays is roughly 4% of salary you pay for zero claims work, about $2,080 in value that still has to be covered.
Stack those on a $52,000 base and the all-in personnel cost lands between $67,000 and $73,000. Call it $70,000. That is a 35% load, right in the middle of the band, and it exists before the biller touches a single claim.
Then come the tools the job requires. A clearinghouse seat, a claim-scrubbing or coding-assist subscription, and the billing module of your practice management system add $3,600-$7,200 a year. Continuing education and coding-update materials to keep certifications current run another $500-$1,200. Now you are near $75,000-$80,000 all-in for one in-house biller, and you have one biller, which means one point of failure.
What the Percentage Model Actually Costs at Your Volume
Outsourced billing companies almost universally charge a percentage of collections, typically 4% to 9% for a small practice, with 5% to 8% being the common range once you are past the smallest tier. The percentage sounds harmless until you multiply it by real revenue.
A two-provider primary care practice collecting $1.4M a year at a 6.5% rate pays $91,000. At 5% it is $70,000; at 8% it is $112,000. So on the raw number, a mid-priced outsourced service and a fully loaded in-house biller are in the same neighborhood, roughly $70K-$110K, which is exactly why the decision is hard and why cost-per-hour comparisons mislead.
The structural difference matters more than the point estimate. The percentage model is variable: it flexes down in a slow month and up in a strong one, and it includes coverage you would otherwise pay separately for.
flowchart LR
A[Patient encounter] --> B{Billing model}
B -->|In-house| C[One biller<br/>fixed 70K to 80K]
B -->|Outsourced| D[Service takes<br/>5 to 8 pct]
C --> E[Coverage gap<br/>when biller is out]
D --> F[Scales with<br/>collections]
E --> G[Claims stall<br/>AR climbs]
F --> H[Denials worked<br/>by a team]
G --> I[Revenue leak]
H --> J[Cleaner cash flow]Read the two paths carefully. The in-house branch is cheaper on paper in a high-revenue year, but it concentrates your entire revenue cycle in one person's inbox. The outsourced branch costs more when you are collecting well, but a service spreads the work across a team, so a single vacation or resignation does not freeze your cash flow. Neither branch is automatically right; the right answer depends on how much denial and follow-up work your payer mix generates and how much downside risk you can tolerate.
The Line Items Nobody Puts in the Quote
Both models carry costs that never make it into the comparison spreadsheet, and both of them favor the option the salesperson is not selling.
For in-house billing, the invisible cost is coverage and continuity. When your one biller takes two weeks off, gets sick for a week, or resigns, claims stop going out. Timely-filing windows keep running while your outbox sits still. A two-week gap on a $1.4M practice is roughly $54,000 of collections that get delayed, and some slice of it ages past filing deadlines and never comes back. Replacing a biller who quits costs another 20%-30% of salary in recruiting, onboarding, and the three-to-six-month ramp before a new hire hits full clean-claim speed. Our survival plan for when your only biller quits walks through how fast that cash-flow cliff arrives.
For outsourced billing, the invisible cost is scope and control. Many contracts bill the percentage only on primary claim submission and charge extra, or simply do not chase, secondary claims, patient balances, and appeals on denials. If your service quietly writes off anything that takes real effort, your net collection rate drops and the "cheap" percentage gets expensive in aggregate. You also lose real-time visibility: when a patient calls asking about a bill, your front desk is now a middleman to a vendor.
The metric that actually decides which model makes you money is not the fee. It is your clean-claim rate and your days in AR. A biller or service that lands 95%+ first-pass clean claims and holds days in AR under 40 is worth paying more for than a cheaper option running 85% clean and 55 days in AR. On $1.4M in charges, moving from a 12% denial rate to 6% is worth far more than the entire difference between the two staffing models. If you have never calculated it, our breakdown of days in AR against MGMA benchmarks shows where the red lines sit.
Where the Front Desk Quietly Decides Your Billing Cost
Here is the part both vendors leave out: a large share of denials are born at the front desk, not in the billing office. Eligibility never verified, a wrong subscriber ID keyed at check-in, a missing referral, an authorization that was never captured. Whether you bill in-house or outsource, those errors flow downstream and land as rework, denials, and days in AR, and rework is the single most expensive activity in the revenue cycle.
This is where the staffing math connects back to the phones. When your front desk is buried answering calls, verifying insurance between check-ins, and booking appointments, verification gets rushed and the error rate climbs, which inflates your billing cost no matter which model you chose. Take the call volume off their plate and they have room to get the front-end data right.
That is the role CallSphere plays before a claim is ever created. The AI front desk answers 100% of calls, books appointments, and captures accurate patient and insurance details at the point of scheduling, so the information feeding your biller or your billing service is clean from the start. On the back end, the hands-off billing and claims capability drafts and submits claims and follows up on denials automatically, which is exactly the work an outsourced service charges its percentage for and the work an overloaded in-house biller runs out of hours to do. You can see how the front-desk and billing pieces fit together on the features page, and the flat, headcount-independent pricing is the structural opposite of a percentage that grows every time you have a good month.
flowchart TD
A[Call comes in] --> B[AI front desk answers]
B --> C[Insurance captured<br/>at booking]
C --> D[Clean intake data]
D --> E[Claim drafted<br/>and submitted]
E --> F{First pass}
F -->|Accepted| G[Paid faster]
F -->|Denied| H[Auto follow-up<br/>and resubmit]
H --> GThe point is not that software replaces your billing decision. It is that clean front-end data lowers the denial rate that makes billing expensive under either model, so you are choosing between an in-house biller and an outsourced service on a smaller, cleaner claim volume to begin with.
Running the Decision for Your Own Two-Provider Numbers
Skip the vendor math and build your own. Pull four figures: your annual collections, your current clean-claim rate, your days in AR, and your realistic all-in cost for one biller including the 35% load. Then answer three questions honestly.
First, can one person actually carry your claim volume plus denials, appeals, and patient balances? For two full-time providers with a normal primary care payer mix, one biller is usually stretched, and "stretched" shows up as unworked denials, not as a complaint you will hear. Second, what happens the week that person is out? If the answer is "claims stop," you are already paying an unpriced continuity risk that an outsourced team or automated submission removes. Third, is your denial rate above 8%? If it is, front-end data quality, not the choice between models, is your most expensive problem.
For many two-provider practices the sober conclusion is that the models cost roughly the same, and the tiebreaker is risk tolerance and denial management, not sticker price. If you have a strong, tenured biller and a backup plan, in-house keeps control and knowledge in the building. If your biller is a single point of failure or your denial rate is creeping up, outsourcing or automating the submission-and-follow-up layer buys you continuity that a solo hire cannot.
What to Do Before You Sign Anything
Do not decide on the salary line versus the percentage line. Those are the two least reliable numbers in the comparison. Build the fully loaded in-house figure at roughly $70,000-$80,000, price the outsourced service on your actual collections at their real all-in percentage, and then weight both by the metric that moves cash: first-pass clean-claim rate and days in AR.
Before you commit either way, tighten the front end, because that is where denials are manufactured regardless of who works the claims. Every call answered and every insurance detail verified at booking is one less denial your biller or your billing service has to rework later. Fix that, recalculate on the lower denial volume, and the in-house versus outsourced choice gets a lot less nerve-wracking, because you are no longer trying to out-hire a problem that started three steps upstream at the phone.