UK Accountancy Practices

Accountancy is the first sector in this series whose revenue is contracted by default. A dental practice starts each month having to win it; an accountancy practice starts it holding a fee base that will largely renew whether or not anyone does anything. That single structural difference changes which economic question matters.

What the sector publishes

• ICAEW-registered firms in the UK12,000+Observed
• Entering MTD for Income Tax from April 2026860,000+Observed
• Qualifying income threshold, April 2026£50,000Observed
• Mid-tier firms that have acquired another firm~75%Observed
• Responding mid-tier firms that are private-equity backed~50%Observed
• Modelled collection from recorded time72%Modelled
• Total UK accountancy practices—Not established

The title is not protected, no register is comprehensive, and the unregulated segment sits at exactly the small-client end that Making Tax Digital affects most — so the model marks the sector total Not established rather than assembling a plausible figure from overlapping memberships. The more consequential point is what contracted revenue does to feedback. A practice losing money on a client keeps the client, keeps billing the same fee and absorbs the overrun indefinitely. There is no lost pitch and no drop in turnover; the first visible symptom is staff turnover or a cash squeeze.

Including unregulated practices is deliberate: a small compliance firm losing price-sensitive clients is usually losing them to a competitor with a lower cost base and no professional indemnity or CPD obligation, not to another ICAEW firm. The six archetypes are configurations rather than stages, and each fails in its own direction. The sole practitioner fails on capacity, having no leverage. The small compliance practice fails on realisation, having leverage but no scope control. The established general practice fails on lockup. The consolidator fails on integration, because buying a fee base does not improve realisation on it.

Accountancy has the opposite measurement problem to most sectors in this series. The demand change is precisely known because it is legislated; the market itself is not countable because the title is not protected and no register is comprehensive.

Making Tax Digital is discussed as an opportunity; read carefully it is a capacity shock arriving as a compliance obligation. Client numbers do not change — filing events per affected client rise roughly fourfold, while the supply of qualified staff and the number of productive hours in a fee earner’s year do not move on that timetable. The increase has to be absorbed by higher fees, higher throughput per hour or lower realisation, and only the second improves the business. The segment most affected is also the lowest-fee one.

Revenue begins with purchased fee-earner capacity and improves in quality as more of it is sold against value rather than against time — through defined scope, billed variations and advisory work the client cannot obtain a comparative quote for.

The GRF multiples are published by commercial brokerages and describe practices that came to market; they are not a valuation and not a census. Realised fee per hour is modelled throughout, because no professional body collects rate or fee data.

Almost every other sector in this series transacts on a multiple of earnings; accountancy transacts on a multiple of gross recurring fees, and the convention is informative rather than merely conventional. It says buyers believe they are acquiring an annuity and intend to apply their own cost structure to it — they are explicitly not buying the seller’s margin. Read as a ladder rather than a valuation, 0.8 to 1.7 times is a specification: it states, in the market’s own terms, what has to be true for a fee base to be worth half as much again.

The same sector produces very different economics when realisation, service mix, staff leverage and lockup change. Principal fee-earning time is normalised to a market replacement rate before economic EBITDA, which is what separates a practice from a well-paid job.

Base central model: 1,150 productive hours per fee earner per year · 86% realisation · £81 modelled realised fee per productive hour

Contribution and EBITDA break-even both appear in the management accounts; cash-survival break-even appears nowhere unless somebody calculates it. Lockup is the reason. A practice growing fifteen per cent a year on 110 modelled days is funding a permanently expanding block of unbilled and uncollected value out of its own working capital, which produces the sector’s characteristic failure: a profitable, growing practice that cannot pay salaries. Nothing in the profit and loss account warns of it, because on an accruals basis the practice genuinely is profitable.

Qualified fee-earner capacity binds before client demand in every released archetype, and statute is about to consume more of it per client. Recruitment, delegation depth, review load and records chasing determine whether purchased capacity is economically used.

The gap between 1,750 contracted hours and 1,150 productive ones is mostly legitimate — leave, CPD, training, practice administration. What makes it worth examining is that a meaningful share of the remainder goes on chasing client records, which is fee-earner time producing no billable output at all, and quarterly reporting multiplies exactly that cost: one records request a year becomes four, at the same chase length each time. Records automation is therefore the only rung that recovers hours from the six-hundred block rather than competing for hours inside the eleven-fifty.

Accountancy does not usually have a retention problem. It has a realisation problem, and the losses occur after the client has already been won, inside work the practice has already agreed to perform.

Ask a prospect for these twelve figures. Most practices can produce the fee base, the debtor days and the hours recorded. Realisation by client, write-downs by approver and scope variations billed are the ones that are missing, and they are the three that bracket the entire loss.

The modelled waterfall loses twelve points at write-down, nine to scope creep, five to lockup ageing and two to bad debt, and almost none of it is competitive. Write-down is usually a partner deciding at the point of billing that the recorded time cannot reasonably be charged; the decision is often correct, and it is costly because it is rarely recorded with a reason, so nobody can tell whether the job was mispriced, performed inefficiently or supplied with poor records. Scope creep is the same loss earlier and far more tractable.

Accountancy has an unusually explicit financeability ladder, because the market states its own premium conditions. Practices transact at 0.8 to 1.7 times gross recurring fees, and the reported conditions for the top of that range are a specification rather than an opinion.

Private-equity penetration of the mid-tier now runs to roughly half of responding firms, and the transaction convention states what that capital is buying: an annuity to which the buyer intends to apply its own cost structure. For an independent practice the reading is straightforward — competing on price is unattractive, because cost of capital and central technology spend are both structurally better on the other side, while competing on realisation and niche is viable, because those are operating properties that capital does not confer. Rung three is where that starts.

Modelled effect on normalised EBITDA of a one-standard-step move in each driver, established general practice, widest first. As in every professional-services archetype, the two widest movers are both revenue-side.

Lockup sits fifth by effect on normalised EBITDA and first by effect on whether a practice survives a growth phase; both are true and they are not in tension. Reported profit is calculated on an accruals basis, so work performed but unbilled and fees billed but uncollected both count as earned, and a practice can report a rising margin while its cash position deteriorates every month with no warning in the accounts. The MTD scenarios carry the same shape: absorb, reprice and re-engineer are the only three, and a practice choosing none has chosen to absorb.

Technology creates value in accountancy only when it moves one of the drivers on the previous page. This is not a software catalogue. Each intervention below names the driver it moves, the operator data it requires and the commercial test that decides whether it worked.

Almost every practice can be sold software promising efficiency, and for most that is the wrong first move: efficiency applied to work whose profitability is unknown simply produces the same unknown result faster. Instrumenting realisation is unglamorous and cheap — hours by job, write-downs by approver, variations raised, three fields most practice management systems already hold and few populate with any discipline. Once they exist the practice discovers which clients it is losing money on, which is frequently the most valuable output of the exercise and arrives before anything is built.

The sector model becomes commercially useful at the moment a named practice is scored against it. The eight dimensions below convert everything in this model into a diagnostic that can be completed in a single working session with the practice’s own data.

The eight dimensions are scored against the practice’s own figures rather than against an interview, and in accountancy that distinction is especially sharp: principals are numerate and confident about numbers they have not actually calculated. Where a figure cannot be produced the dimension scores one, which is the finding rather than a penalty. Sequencing then follows the sensitivity ordering with one deliberate exception — where lockup governance scores at level one and the practice is growing, it is addressed first regardless of its fifth-place ranking, because that failure is terminal rather than gradual.