How to Choose a Medical Billing Company in 2026 (Without Regretting It in Six Months)

Almost every practice that switches billing vendors tells the same story. The demo was excellent. The first month was fine. By month four, nobody could say why the accounts receivable over 90 days had doubled, the assigned account manager had changed twice, and the monthly report was a PDF that answered none of the questions anyone actually had.

Choosing a medical billing company is not really a software decision or even a price decision. It is a decision about who will be accountable for your cash flow when a payer changes a policy on a Tuesday and nobody sends an email about it. This guide walks through how to evaluate that, using the criteria that separate vendors who work claims from vendors who merely process them.

First, decide what you are actually outsourcing

The phrase "billing company" covers at least four different service scopes, and confusion here causes most of the disappointment later. Be explicit about which one you are buying before you talk to anyone.

• Claims submission only — the vendor takes your coded charges, scrubs them, and files them. Denials, appeals, and patient balances stay with you. Cheapest, and the least useful for a struggling practice.
• Full revenue cycle management — charge entry, submission, denial work, appeals, payment posting, patient statements, and reporting. This is what most practices mean when they say outsourcing.
• Coding plus billing — the vendor's certified coders assign codes from your documentation. Adds cost, removes a large source of denials, and shifts compliance exposure that you still ultimately own.
• Credentialing and enrollment bundled in — payer enrollment, CAQH maintenance, and recredentialing handled alongside billing. Worth paying for if you hire providers regularly.

Write your scope down in one paragraph before the first sales call. If a vendor's proposal describes a different scope than your paragraph, that is not a negotiation detail — it is the whole contract.

Understand the three pricing models and what each one incentivizes

Pricing is where incentives live. Every model rewards a particular behavior, and you want the reward pointed at collections, not at volume.

Percentage of collections is the dominant model in the United States, typically landing somewhere between four and nine percent of net collections depending on specialty, claim volume, and average charge. Its virtue is alignment: the vendor gets paid when you get paid. Its weakness is that it quietly rewards chasing large, easy claims and letting a tail of small balances rot. Ask directly how small-balance claims are worked and whether there is a dollar threshold below which appeals stop.

Per-claim flat fees look transparent and suit high-volume, low-complexity specialties. The risk is the opposite of the percentage model: the vendor earns on submission, not on payment. A denied claim that gets refiled is revenue for them and a delay for you. If you consider per-claim pricing, insist on a first-pass acceptance guarantee in writing.

Full-time-equivalent or fixed monthly staffing is common for larger groups that want dedicated people and predictable cost. It works when your volume is stable and your internal management is strong enough to direct the team. It fails when you assume the vendor will self-manage.

The numbers you should ask every vendor for

Vendors will happily talk about technology. Push past it and ask for performance data from clients in your specialty, ideally with the client names redacted but the specialty and state named. Any established vendor can produce this in a day. Hesitation is information.

• Clean claim rate, also called first-pass acceptance — the share of claims accepted on first submission without human rework. Strong performers are consistently above 95 percent.
• Net collection rate — collected dollars as a share of what contracts actually allow. Below 95 percent means money is being written off that could have been recovered.
• Days in accounts receivable, plus the percentage of AR over 90 and over 120 days. The average matters far less than the tail.
• Denial rate by root cause, not just overall. A vendor who cannot break denials into eligibility, coding, authorization, and timely-filing buckets is not analyzing them.
• Average turnaround from date of service to claim submission, and from remittance to payment posting.
• Appeal overturn rate — of denials appealed, how many were paid. This one number reveals whether appeals are real work or a form letter.

Questions that separate strong vendors from polished ones

Sales teams rehearse the obvious questions. These are the ones that get unrehearsed answers.

Who specifically works my account, where are they located, and what is their caseload in claims per person? A named team with a stated caseload is a commitment. "Our team" is not.

What happens when a claim is denied for the third time? Listen for a described escalation path with a human owner and a timeline. If the answer is that it gets written off after a threshold, ask what the threshold is and who approves it.

Show me the actual monthly report you send clients, not a marketing mock-up. The right report shows collections against expected, AR aging, denial root causes with dollar values, and payer-level performance. A report that only shows charges, payments, and adjustments is a bank statement, not management information.

How do you handle payer policy changes? A good answer names specific payers, describes who monitors bulletins, and gives an example of a change caught this year and what the vendor did about it.

What is your credentialing capability? Even if you are not buying credentialing today, a vendor who understands enrollment will spot the denials that are really credentialing failures in disguise — the out-of-network rejections that appear when a provider's payer file lapses.

What is your process on day one of transition, and what does the runout of old claims look like? Transitions are where practices lose the most money, and a vendor who has done many of them will describe the runout plan without being prompted.

Contract terms worth reading twice

Billing contracts are usually short, which lulls people into skimming them. The clauses below determine what happens on your worst day, so read them on your best one.

• Term and termination. Anything longer than twelve months for an initial term deserves a discount. Look for a termination-for-convenience clause with 60 to 90 days' notice, not just termination for cause.
• Data ownership and export. Your claims, remittances, patient balances, and denial history are yours. The contract should say so, name the export format, and set a delivery deadline after termination.
• Runout handling. When you leave, who works the claims already in flight, at what rate, and for how long? Silence here is expensive.
• Performance standards with teeth. A stated clean claim rate or days-in-AR target that carries a fee credit if missed is meaningfully different from an aspiration in the marketing deck.
• The business associate agreement. Required under HIPAA, and it should name breach notification timelines, subcontractor obligations, and where protected health information is stored and processed.
• Fee base definition. Percentage models must state clearly whether the fee applies to net collections, gross charges, or all deposits including patient payments the practice collects at the desk. This single definition can change your annual cost by tens of thousands of dollars.

Red flags that reliably predict a bad year

None of these is automatically disqualifying on its own. Two or more together are a pattern.

• A quoted rate far below the market range for your specialty. Billing is labor. Unusually cheap labor gets spread thin across too many accounts.
• No named account manager, or an account manager who changes during the sales process.
• Reluctance to share references in your specialty and state.
• Reports that cannot be customized, or a refusal to show a real one before signing.
• No written transition plan.
• Vague answers about where work is performed and how PHI is secured.
• Pressure to sign before the end of a month or quarter.

Common questions practices ask before switching

How long does a transition take? Plan on 30 to 60 days from signature to full operation. Clearinghouse enrollment, payer EDI agreements, and electronic remittance setup all involve third parties with their own queues, and none of them move faster because your contract has a start date. Practices that plan for 30 days and get 55 experience it as a failure; practices that plan for 60 and get 45 experience the same vendor as excellent.

Will we lose money during the switch? Some short-term slowdown is normal, and the size of it depends almost entirely on how the runout of existing claims is handled. Agree in writing who works claims already in AR, at what rate, and until what date. The worst outcome is the gap where the old vendor has stopped caring and the new one has not started.

Do we have to change our practice management system? Usually not. Most billing companies work inside the client's existing system, which is preferable because your clinical workflow and historical data stay put. Be cautious with vendors who require their proprietary platform: it can be excellent, but it also means your data lives with your vendor, and leaving becomes materially harder. If you go that route, the data export clause is no longer boilerplate — it is the most important paragraph in the agreement.

What about compliance? Your billing partner is a business associate under HIPAA and needs a signed business associate agreement before any protected health information moves. Ask where data is stored, whether any work is performed offshore, what subcontractors are involved, how access is logged, and how staff are trained. Ask when their last security assessment was and what changed as a result. A vendor who answers these crisply has been asked before by someone serious.

Who owns compliance risk for coding? You do. A vendor's coders can be excellent and the practice still carries the exposure for what is submitted under its tax ID. That is an argument for periodic independent coding audits regardless of who does the coding, not an argument against outsourcing.

Run a real 90-day evaluation instead of a gut check

Once you sign, resist the urge to judge the vendor on feeling. Set a baseline in the first week using your own historical numbers — days in AR, AR over 90, net collection rate, denial rate by cause, and monthly collections for the same months last year. Then review at 30, 60, and 90 days against those figures.

Expect a dip in month one. Transitions always cost something: enrollment mapping, clearinghouse setup, and the learning curve on your documentation habits. What you are watching for is direction. By day 60, first-pass acceptance should be climbing. By day 90, AR over 90 days should be flat or falling and denial root causes should be identified, not merely counted.

Two things make this review honest. First, agree on the metric definitions in writing before day one, because "net collection rate" has more than one arithmetic in the wild. Second, schedule the three reviews as calendar meetings during onboarding. Reviews that are not scheduled do not happen, and unmonitored vendor relationships drift by default.

In-house, outsourced, or hybrid?

Outsourcing is not automatically right. A single-provider practice with a strong, tenured biller and a clean payer mix often collects better in-house than it would with a mid-tier vendor. The math tips toward outsourcing when you face staff turnover you cannot easily replace, when you are adding providers or locations, when your payer mix is complex, or when AR over 90 days has crossed 20 percent and internal capacity is the binding constraint.

The hybrid arrangement is underrated. Keep front-desk eligibility, authorization, and patient collections in-house where the patient relationship lives, and outsource claim submission, denial work, and appeals where scale and payer specialization pay off. Practices that do this well tend to hold on to the parts of the cycle that depend on local knowledge and buy the parts that depend on volume.

A short scorecard you can use this week

Score each finalist from one to five on the following, then compare totals rather than impressions: specialty experience, published performance metrics, named team and caseload, denial and appeal methodology, reporting quality, transition plan, contract flexibility, credentialing capability, technology fit with your EHR, and reference quality.

Weight specialty experience and denial methodology double. Those two predict outcomes better than anything else on the list, including price. A vendor charging a point more but recovering four points of net collection rate is not the expensive option.

Whatever you decide, put the review cadence in place before the contract starts. The practices that get good results from billing partners are not the ones that picked perfectly — they are the ones that kept measuring after the ink dried.