Group Therapy Practice Profitability: The Math That Actually Determines Owner Income

Modeled estimate: a group therapy practice running at steady utilization earns an operating margin of roughly 12–14% before owner income tax, with owner profit scaling from about $63,000 per year at 5 clinicians to about $277,000 at 20. That answer comes from the transparent model on this page — stated inputs, arithmetic shown — not from a survey or a client sample.

Published commentary on this question quotes owner incomes of $60,000–$200,000 and margins anywhere from 7% to 40%. Those figures come from coach and accounting blogs and are anecdotal; the ranges are wide mainly because the writers rarely publish their inputs. When the inputs are explicit — utilization, compensation split, overhead — the plausible range narrows sharply and every disagreement becomes checkable arithmetic rather than competing anecdotes.

One definition matters before the numbers. “Owner profit” here means the practice’s operating profit before income tax, and it excludes any clinical work the owner personally performs. An owner who also carries a caseload earns their own clinician compensation on top of the figures below.

Four levers set almost the entire result: utilization, compensation split, overhead, and rates. Everything else — branding, niche, location prestige — only matters to the extent it moves one of these four.

Utilization is the number of sessions each clinician actually holds per week, after cancellations and no-shows. It is the strongest lever because a filled slot adds revenue against costs that are mostly already paid: the office exists, the admin team is staffed, the software subscription does not change. In this model, each additional weekly session per clinician is worth about $1,656 per clinician per year in owner profit.

The compensation split is the share of collections paid to the clinician who delivered the session. Published commentary typically describes W-2 splits of 50–60% of collections plus payroll taxes and benefits, and 1099 splits of 60–70% with no payroll load. This model uses a 60% fully loaded W-2 share — an assumption, stated as one, and roughly equivalent to a 1099 practice paying 60–65% directly.

Overhead is everything that does not scale with a single session: rent, admin payroll, software, insurance, accounting, marketing. Billing losses behave like overhead in reverse — unworked denials and write-offs quietly remove 3–5% of gross revenue in this model’s assumption. Rates are the average actually collected per session across the payer mix, which is usually lower than any single fee schedule suggests.

The ranking matters. A practice cannot cut its way to health through overhead alone: the entire software line at 10 clinicians is smaller than the modeled value of one weekly session per clinician across the group. Fill calendars first, negotiate rates second, and treat overhead as the discipline that keeps the first two from leaking.

Start with one clinician. At 22 sessions per week for 46 working weeks, a clinician holds 1,012 sessions per year. At $100 average collected per session, that is $101,200 in annual revenue. At a 60% loaded compensation share, the clinician costs $60,720, leaving $40,480 per clinician per year to cover shared overhead and owner profit.

Quotable model output: in this model, a 10-clinician practice at 22 sessions per clinician per week and a 60% loaded compensation share nets the owner $142,640 per year — a 14.1% operating margin. Every input behind that sentence is in the methodology section.

Scaling from 5 to 20 clinicians, the model steps overhead the way group practices typically experience it: roughly one loaded admin FTE per five clinicians, rent that steps up with office capacity, a software line that grows with headcount under per-clinician pricing, and billing losses that scale with revenue. Margin stays in a narrow 12–14% band; the dollars scale with size.

More than any other input, and it is worth seeing precisely. Hold the 10-clinician practice constant — same compensation split, same overhead — and move only average sessions held per clinician per week. Compensation (60%) and billing losses (4%) scale with revenue; the remaining $221,680 of overhead does not. So roughly 36 cents of every marginal session dollar reaches the owner.

Quotable model outputs: moving average utilization from 20 to 25 sessions per clinician per week raises modeled owner profit from $109,520 to $192,320 — an $82,800 swing on the same headcount and the same overhead. Across the full table, utilization alone moves the modeled operating margin from 7.7% at 17 sessions to 18.1% at 27.

This is why schedule density, cancellation recovery, and intake speed are profit work, not admin trivia. A practice that fills two more slots per clinician per week has done more for owner income than any overhead cut available to it.

No government agency publishes group-practice owner income, which is why this page is a model. What public data does exist anchors two of the model’s inputs: clinician labor cost and session rates.

On labor: the U.S. Bureau of Labor Statistics reports a median annual wage of $59,190 for substance abuse, behavioral disorder, and mental health counselors and $63,780 for marriage and family therapists (May 2024). Sanity check: a modeled W-2 clinician producing $101,200 in collections at a 52–56% base split earns roughly $53,000–$57,000 in base wages before benefits — right at the BLS medians. The model’s compensation assumption is ordinary, not exotic.

On rates: CMS’s Calendar Year 2026 Physician Fee Schedule lists a national non-facility payment of about $167 for CPT 90837 (60-minute psychotherapy), with 90834 (45 minutes) lower; the CY 2026 final rule also exempts time-based behavioral health services from its new efficiency adjustment. Commercial reimbursement for masters-level clinicians frequently sits below Medicare psychologist rates, and self-pay varies by market — so this model’s $100 blended average collected per session is a deliberately conservative, explicitly editable assumption rather than a claim about your payer mix. The startup playbook on this site works its launch-stage compensation example at a $140 illustrative contracted rate; the difference between the two figures is the stated assumption — conservative blended average here, single contracted-rate illustration there — not a disagreement about payer data.

What remains genuinely unpublished anywhere: utilization distributions across group practices, real overhead ratios, and denial-rate distributions. Anyone quoting precise industry averages for those is estimating — this page just says so out loud.

Admin payroll is the larger line — about 9–10% of revenue in this model ($96,000 at 10 clinicians) — but it is also the line that buys utilization: intake speed, schedule density, cancellation backfills, and clean claims are admin work. Understaffing admin below capacity tends to show up as empty calendar slots, and in this model one lost weekly slot per clinician costs $16,560 per year across a 10-clinician group — more than most admin savings.

Software is a smaller line with one distinctive property: under per-clinician pricing it is the only overhead item that automatically grows every time you hire. Quotable model output: at 10 clinicians, per-clinician practice management software at published July 2026 rates costs $4,680–$11,880 per year, versus $1,788 on a flat $149-per-month per-practice plan — a difference worth roughly half a weekly session per clinician in owner profit.

Billing losses outweigh both when left unmanaged. The model assumes 4% of gross revenue ($40,480 at 10 clinicians); every additional point of unworked denials costs $10,120 per year at that size. Software and admin choices interact here — this is the connection most profitability commentary skips: the point of the overhead stack is to protect utilization and collections, not to be minimized in isolation.

Copy the P&L structure above and replace the assumptions, in this order. First, measure real utilization: sessions held — not booked — per clinician per week, averaged over at least eight weeks. Second, compute your true average collected per session from remittances, not from your fee schedule; the gap between billed and collected is usually the first surprise. Third, load your compensation correctly: base split plus payroll taxes, benefits, and paid time for W-2 staff.

Then fill in the overhead rows from your actual books — rent, admin payroll, software, insurance, accounting, marketing — and take billing losses from your aging report rather than accepting this model’s 4%. The result will not match this page’s numbers, and it should not: the model’s job is to make the structure of the answer checkable, so your version of it is too.

For the software row, the interactive cost calculator runs the per-clinician versus flat-rate math for your exact clinician count and current platform, with published prices as editable defaults.

Every number on this page is either cited to a public source or listed below as an assumption. The model excludes: the owner’s own clinical caseload (add your clinician-level earnings on top), income taxes, debt service, one-time startup costs, and data-migration or training costs when switching systems. It models a steady-state year, not a launch year.

Corrections are invited. If you own or advise a group practice and believe an input is wrong, email hello@clinicpro360.com with the correction and its basis. Material changes will be logged on this page with dates. Inputs are refreshed when BLS releases new occupational wage data and when CMS publishes each year’s fee schedule — roughly semiannually.

Update log: model published July 14, 2026, with inputs reviewed the same day. No revisions logged yet.