Clinician Compensation Structures for Therapy and ABA Practices

Most compensation structures weren’t designed. They accumulated. A starting rate set years ago, a raise at an anniversary, a bump to keep someone from leaving. Each one made sense at the time. But reimbursement rates didn’t move with them, and nobody was adding up the total. That’s usually the story behind a P&L that keeps getting tighter while the practice keeps growing.

Clinician pay is almost all of what it costs to deliver the care. That makes compensation a margin question before it’s an HR question. The gross margin piece covers how to read a P&L (the profit and loss statement) that way. This piece covers the pay structures themselves and what each one does to margin over time.

Compensation models in therapy and ABA practices

One thing to know up front: nobody formally surveys how these practices actually pay people. The APA and the counseling associations track what clinicians earn, not how the pay is structured. So the ranges below come from accounting firms and vendors who work inside these practices, checked against what owners themselves report. I’d treat them as consensus, not verified benchmarks.

Fee splits, session rates, and salary in mental health group practices

Mental health group practices tend to use four structures, sometimes in combination:

• Fee split. The clinician gets a percentage of what the practice collects, typically 50 to 70 percent, with 60/40 cited most often. Check what the split applies to: 60 percent of collections is a different number than 60 percent of billed charges, because billed charges include write-offs the practice absorbs first.
• Per-session flat rate. A set dollar rate keyed to the billing code: one rate for a standard 45-minute session, longer codes at a multiple of that base.
• Straight salary. Now mostly limited to leadership and supervisory roles, paying for oversight rather than session volume.
• Salary plus productivity. A base salary with a bonus once the clinician clears a caseload threshold.

Pre-licensed clinicians, under any structure, typically earn 10 to 20 percentage points less than licensed peers. The gap covers the supervision the practice provides.

BCBA and RBT compensation structures in ABA practices

ABA practices structure BCBA and RBT pay differently:

• BCBAs: salary plus billable-hour tiers. The most common structure in job postings and owner-reported data is a base salary, commonly $75,000 to $95,000, paired with a billable-hour expectation of 25 to 30 hours a week and a per-hour bonus above that threshold. Straight hourly pay exists too, though less commonly.
• RBTs: hourly. Almost universally, and classified as non-exempt, meaning the practice must pay them for all hours worked, not just billable ones. The wage rules behind that are in the legal section below.

The one credible survey here is BHCOE’s annual benchmark report, drawn from 87 organizations across 32 states.

The fully loaded cost of a clinician: payroll taxes, benefits, supervision

A clinician costs more than the number in the offer letter. The practice pays FICA (7.65 percent), unemployment tax (usually another 1 to 3 percent), and workers’ comp. If there’s health insurance, a retirement match, and PTO, that’s usually another 20 to 30 percent of base pay, though ABA practices tend to run leaner. And clinicians get paid for hours that don’t bill: documentation, meetings, training, supervision. Put it all together and a clinician usually costs somewhere between 1.25 and 1.4 times their base pay.

Why clinician compensation rises faster than reimbursement rates

Reimbursement hasn’t kept up with inflation. Medicare is the one payer that publishes its rates over a long window, so it’s the cleanest place to see the trend.

Medicare sets its rates with something called a conversion factor, the dollar multiplier behind every code it pays. In 2020 that number was $36.09. In 2026 it’s $33.40, a 7.5 percent cut before inflation, and inflation ran about 27 percent over the same stretch (per BLS CPI data). Some psychotherapy codes got value increases along the way that clawed part of that back, but nobody publishes a series that nets it all out.

Medicaid ABA rates are a state-by-state story with no national trend. New York cut its technician-delivered rate 25 percent between late 2025 and early 2026. Alabama has been flat since at least 2023. Nebraska raised its schedule in 2025 after a market rate study. And commercial rates are negotiated contract by contract; nobody publishes how they’ve moved over time.

Private-pay practices aren’t off the hook; the mechanism is just different. The practice sets its own rates, so there’s no payer to blame. But raising fees fast enough to keep up with compensation runs into what clients will pay, and the drift shows up the same way: pay growing faster than the fee increases the owner is actually willing to take.

Here’s the simple math. A raise this year becomes next year’s base; it doesn’t reset. Say clinician compensation is 55 percent of revenue and grows 3.5 percent a year while revenue per clinician stays flat. That share climbs about two points a year. Nobody made a decision that looked wrong at the time, and the number moved anyway.

Compensation drift on a real practice P&L

A group practice I work with saw this play out over two years. Year one, revenue grew about 20 percent and payroll grew about half that. Healthy. Year two, revenue grew about 5 percent, payroll grew about 16 percent, and net margin (what’s left after every cost, not just the cost of care) swung about eight points, from a mid-single-digit profit to a small loss. When we compared the two years employee by employee, the single biggest driver wasn’t new hires or people working more hours. It was rate increases on existing staff: the same hours at new rates.

No single raise was large. Most were 50 cents to 2 dollars an hour, granted at anniversaries, often after years with none. There was no annual raise budget, so nobody could see the total until it showed up on the P&L.

I’ve seen the same pattern in ABA. After 2020, holding onto BCBAs in a competitive market pushed compensation conversations faster and higher than any plan would have set them. In both cases, nothing could show the decisions in total until the margin had already moved.

Three numbers that show whether clinician compensation is drifting

Three numbers show whether compensation is drifting and how fast.

1. Fully loaded clinician compensation as a percent of revenue, today. Wages or splits, employer payroll taxes, and benefits, for clinical staff only, over the trailing three to twelve months. Keep supervisor and BCBA pay in; it’s already clinical payroll. Use collections as revenue if the books are cash-basis. If benefits aren’t split by staff type, allocate by clinical payroll’s share of total payroll.
2. The same number for an earlier period. Three years back is ideal, but year over year works if that’s what the books can produce. The trend matters more than the level, and nobody publishes this ratio for behavioral health, so it has to come from the practice’s own books. If older numbers aren’t recoverable at all, start the baseline now and compare from here forward.
3. The comparison to benchmark, in the next section. For ABA, compare wages plus taxes and benefits against their separate rows. The mental health range is all-in.

A practice at 54 percent that was at 54 three years ago is stable. The same 54 percent that was at 48 three years ago is drifting at two points a year.

Clinician compensation benchmarks for therapy and ABA practices

Here’s where that number lands against what’s published.

Sources: ABA figures from Flychain’s 2026 benchmark report, drawn from more than 350 practices; the combined row is this article’s sum of the two Flychain ranges, not a Flychain figure. Mental health figures are a convergence of guidance from CPA firms serving this segment, not a formal survey, and should be read with that weight.

A quick note on how these connect to gross margin (revenue minus the cost of care; the gross margin piece treats anything under 40 percent as the warning zone). For mental health, the math is almost automatic: if all-in clinician compensation is above 60 percent of revenue, gross margin is under 40. For ABA, watch the combined load. Wages, taxes, and benefits together above roughly 60 percent of revenue put a practice in the same zone, even when the wage-and-tax piece alone looks fine. And notice that the top of the combined range in the table already crosses that line. A practice can be inside the typical range and still be past the threshold; typical isn’t the same as sustainable.

Fee split vs. salary: how the compensation model changes margin drift

The same market pressures create different drift depending on how a practice pays.

In the practices I work with, the split percentage itself has generally held steady. The squeeze comes through what the retained share has to absorb, and nothing in a split structure catches that automatically.

Under salary, pay can rise at every review while the weekly session count stays flat, and nothing catches that gap either. The fix depends on the model. On a split, price the full employer load into the percentage instead of assuming the split protects the retained share. On salary, tie increases to delivered hours, not just tenure.

How to fix compensation drift: raise budgets, hour tiers, new-hire structures

Fixing drift is more practice-specific than measuring it, and I have less of a settled playbook here than for the diagnostic above. What follows is what I’ve seen work, and where it stops working.

Set an annual raise budget before any individual conversation. Decide from the P&L what total dollars the practice can add to compensation this year. Then push allocation down to clinical leadership, with parameters for higher and lower, but the total stays inside that number. People talk, so a consistent and transparent process matters almost as much as the amount. The practice in the earlier story had no such number, which is how the increases went untracked until they hit the P&L.

Tie increases to delivered hours. One ABA practice I work with answers requests for higher pay by pairing the raise with a larger billable-hour expectation: one rate tier unlocks at 25 hours a week, the next at 30. A clinician who wants to earn more delivers more, and the increase is funded by the hours behind it. Tracking has to stay simple enough to run consistently, and transparent enough that clinicians trust it.

The tier structure can land badly with clinicians who chose group practice partly to get away from productivity pressure; some will read it as a quota. The version that has held up is opt-in: nobody’s existing arrangement changes, and the higher tier exists only for the clinician who asks for more. Some will still leave over the message underneath, and a practice introducing this should expect that self-selection.

Apply the new structure to new hires first. Repricing people already on staff is the hardest version of this change, and usually stalls. Setting the go-forward structure for new hires stops the drift from compounding without reopening existing arrangements.

I haven’t seen a practice cleanly unwind accumulated drift by resetting existing pay downward. It resolves slowly instead, through attrition, and through raised expectations that some clinicians meet and others don’t. Prevention is the realistic goal.

Legal issues in clinician compensation: 1099 status, fee-splitting laws, wage-and-hour rules

None of this is legal advice. It’s a map of where compensation structures create exposure, to put in front of an attorney before finalizing a structure.

Worker classification. The IRS applies a control-based test to decide whether someone is an employee or a contractor, and the Department of Labor’s 2024 rule applies a six-factor economic-reality test on similar ground. In one published IRS determination letter involving a therapist, the worker was found to be an employee despite a signed contractor agreement. A practice that controls a contractor therapist’s schedule, documentation standards, and caseload carries real classification risk, and penalties scale with payroll.

State fee-splitting statutes. Roughly twenty states bar psychologists from paying or receiving compensation for referrals, and several extend the rule to other licensed professions. APA Services’ guidance draws the line at payment for the referral itself, not ordinary compensation for services rendered, so a standard revenue split for a treating clinician is generally distinguishable. The rule is state-specific.

ABA ownership rules. New York requires ABA professional entities to be owned by licensed behavior analysts, and Illinois has enacted the same rule, effective 2027. It’s an ownership question rather than a compensation question, but it interacts with any structure that routes a percentage fee to a non-clinical management entity.

Wage-and-hour rules. Non-exempt hourly employees, which covers most RBTs, must be paid for all hours worked, including documentation and training, plus overtime past 40 hours a week. Per-session pay for staff who should be hourly can fall below minimum wage in a light week if non-billable time isn’t separately paid.

A diagnostic to run this week

Pull the last three months of P&L. Add fully loaded clinician compensation: wages or splits, employer payroll taxes, and benefits, with supervisory pay left in. Divide by revenue for the same three months. Then repeat the calculation for the same three months for the last three years, and compare the numbers to the benchmark ranges above.

Lined up side by side, those numbers tell the story: either the structure still fits the practice, or it’s been drifting away from what the practice can sustain. The earlier that shows up, the more options stay open. An annual raise budget is easy to set before drift has accumulated, and much harder to introduce once every clinician expects a raise on request.

If the books can’t produce either number cleanly, that’s the first problem to solve, before any of the rest of this applies. What a bookkeeper should be producing for a practice covers what that looks like.