UK Private Physiotherapy & Multidisciplinary MSK Clinics

The private physiotherapy market is large in underlying need, fragmented in delivery and poorly described by official business statistics. Commercial performance is better understood through operator-level throughput, episode economics, revenue quality and operating maturity.

What the sector publishes

• HCPC-registered physiotherapists76,910Observed
• CSP members in private practice11 in 100Observed
• People with a UK MSK condition>20mObserved
• Sample median initial / follow-up£78.75 / £65Observed sample
• Spare attended capacity in base archetypes26–40%Modelled
• Recurring or contracted share, single-site to platform22–44%Modelled
• Founding acquisitions in the July 2026 platform launch3Observed

The gap between 76,910 registered physiotherapists and the number of physiotherapy businesses is the central measurement problem in this sector: registration counts people, not trading entities, and the official classifications that would count entities fold private clinics in with a much wider set of activities. Any market-share claim built on a national denominator is therefore unfalsifiable, and this model refuses to publish one. What can be established is the shape of the economics inside an operator — attended visits, what each realises, and how much of it is contracted forward.

NHS departments and hospital rehabilitation units deliver clinically similar care but do not price it, market it or collect it the way an independent clinic must; including them would blend two incompatible revenue identities. Digital-only platforms are excluded from the operating core for the same reason and retained as payers, substitutes and competitors, because they change what a private patient will pay and compete for the same discovery moment. Running across all six archetypes is an axis the labels do not capture — whether the configuration is measured. Two clinics with identical rooms, clinicians and revenue can sit two rungs apart on that alone.

The need substrate is large and the provider landscape is fragmented, but the private market is not cleanly measured. This edition shows known demand anchors and buyer mechanisms rather than a false-precision national TAM.

High information asymmetry alongside medium-to-high substitutability describes a buyer who cannot assess clinical quality before purchase and who has visible alternatives at a similar price. In that regime proof and friction do more work than tariff: review count and named-clinician visibility behave like price, because a patient who cannot judge the clinical work substitutes the signals they can judge. The practical test is not whether the tariff is competitive but whether the path from discovery to a confirmed appointment can be completed without a human having to intervene.

The sector begins with transactional appointments but can evolve into treatment episodes, packages, memberships, insurer activity, employer contracts and specialist pathways. Revenue quality improves as future income becomes more visible.

List price is not realised price. Geography, appointment duration, clinician seniority, payer mix, insurer fees, packages, discounts, cancellations and initial-to-follow-up mix all change the revenue earned per unit of clinician time. The captured observations are a sample frame rather than a census: they support a statement about the shape of the corridor, not a national average, and the model does not publish one.

An initial assessment observed between £40 and £140 against a sample median of £78.75 is not evidence of irrational pricing. It is evidence that the sector sells several different products under one label — appointment length, clinician seniority, location cost base and whether the fee includes a treatment plan all vary underneath a single word. Benchmarking a clinic against the median is close to meaningless without controlling for those four. What is meaningful is the internal spread: whether the same 60-minute appointment is sold at two different prices through two different channels, the most common recoverable pricing defect in the sector.

The same clinical market can produce very different economics when utilisation, staffing model, episode design, revenue recurrence and overhead absorption change. Owner clinical labour is normalised to a market rate before EBITDA.

Established single-site base: 4.5 visits per episode · approximately £327 total revenue per episode · approximately £170 gross profit per episode

Owner clinical labour is normalised to a market rate before EBITDA, which is why the founder-led corridor resolves to approximately break-even: revenue that looks profitable on a drawings basis is largely the owner’s own clinical time, costed at what an associate would be paid to deliver it. That is not a criticism of the model but the correct starting point for any conversation about transferable earnings. Between £666k and £1.039m of revenue, modelled EBITDA moves from £57k to £355k — a six-fold change on a 56% revenue change, because overhead is close to fixed across the band.

In the current archetype configurations, clinician time binds before room time. Demand capture, diary design, attendance and clinician productivity determine whether existing infrastructure is economically used.

Every archetype carries between 26% and 40% spare attended capacity. That is not idle equipment; it is contracted clinician time, leased floor space and staffed opening hours already paid for and not producing revenue — the closest thing the sector has to free inventory. Because clinician time binds before room time the constraint is almost never physical: a clinic that believes it needs another treatment room usually needs a fuller diary in the rooms it has. Three of the four levers that close the gap are administrative rather than clinical, which is why capacity recovery is a systems project rather than a hiring one.

Each transition changes the number of people who reach an attended episode, the acquisition cost allocated to that episode and the probability that value continues through completion, referral, reactivation or membership.

Join source, booking, attendance, revenue and future cohort data so interventions can be measured rather than asserted.

Six transitions sit between an enquiry and an expanded relationship, and each multiplies rather than adds: a clinic converting 70% at every stage keeps about one enquiry in eight through to expansion, which is why single-stage improvements produce disproportionate results and why measuring only the first and last number hides the problem. The two transitions carrying most of the loss are enquiry-to-booking and booked-to-attended. Neither is a clinical variable, and both are recoverable without adding demand.

The market contains local independents, premium specialists, multidisciplinary clinics, regional groups, digital substitutes and emerging consolidators. Financeability depends on repeatability, visibility and management depth rather than brand size alone.

The two percentages on each rung move together, and neither is a marketing outcome. Recurring or contracted share rises from 0% to 44% because the operator installs maintenance offers, membership cadence, insurer panels and employer contracts; owner dependence falls because the same installation removes the founder from the processes that generate that revenue. The ladder is not a size progression — rungs 2 and 3 are reachable at a single site with no additional clinicians, because what changes is governance rather than scale.

The model is most exposed to variables that determine the economic yield of clinician time. Membership attachment, attendance and clinician remuneration are material secondary levers.

Utilisation and realised price sit at the top of the tornado because they act on the same quantity from two directions — one changes how many attended hours the configuration produces, the other what each hour earns. Together they account for roughly £165k of modelled movement against a £224k base, which is to say the base case is a midpoint between two very different businesses rather than a stable number. The four scenarios are not optimism settings: the margin progression from 9% to 39% is driven entirely by how much revenue a fixed cost base is asked to carry.

A modern interface, chatbot or practice-management system is not itself an economic outcome. Axial maps each intervention to the revenue, cost, capacity or information variable it is intended to change.

The interventions most often described as AI here — intake triage, provider matching, missed-call recovery, reminder and reactivation sequences — all sit in the capture and productivity columns. They work on administrative transitions, not clinical ones, which is why they can be adopted without changing scope of practice. That placement sets a ceiling and a floor: no capture technology raises the price a patient will pay, but a clinic carrying 26–40% spare attended capacity has enough unfilled inventory that recovering part of it changes the EBITDA corridor materially.

Axial uses eight dimensions to identify whether a clinic is informal, functional, repeatable, integrated or genuinely scalable and financeable.

The dimensions are scored independently and deliberately do not average into a single headline number. An operator can sit at level 4 on capacity utilisation and level 1 on data visibility, and that asymmetry is the finding — it says the clinic runs well on instinct and cannot yet prove it. Level 3, measurable and repeatable, is the hinge, and it is reachable at any size including a single site with one clinician. Growth adds revenue; maturity makes revenue transferable, and a clinic can double turnover while moving no distance up this scale.