Accounting practice benchmarks: 9 numbers that tell the truth

Most owners of small firms run the business on two numbers: revenue, and whatever is in the bank. Both are lagging. Neither tells you why this quarter felt twice as busy as last year for the same result, or why cash is tight in a month with record invoicing.

The nine numbers below fix that. Each one can be calculated from the practice-management system and ledger you already run. For each: what it measures, how to calculate it, what moves it, and the decision it should trigger. There is a one-page monthly dashboard at the end.

People searching for accounting practice benchmarks usually want a single figure to compare against. Be careful with that. The published studies each survey a different population, and most of the detail sits behind a paywall or a membership login:

• The Crunch (Australia). An annual benchmark report for small and medium Australian firms from The Benchmarking Group, in its third year in 2026. It covers financial performance, productivity, team structure, service mix and pricing, and participants receive a report comparing them with that year’s cohort. Our earlier piece explains what it is and who it suits.
• The PCPS CPA.com National MAP Survey (US). Run by the AICPA’s Private Companies Practice Section with state societies such as TXCPA, every two years. The 2025 results are available to PCPS members, who can filter by firm size and region.
• The Rosenberg MAP Survey (US). An annual study of CPA firms; the 2025 edition, its 27th, drew 296 firms and reports realisation, utilisation, staff-to-partner ratios, staff turnover and income per partner by revenue band. It skews toward firms well above the 2-30 person range.
• UK. We could not find a current, widely published benchmark for UK practices of this size. ICAEW’s Evolution of mid-tier accountancy firms research surveys much larger firms (11 to 249 principals; 35 responded in February and March 2026).

We don’t reproduce fee or charge-out figures from any of them. More useful than any outside median is your own number, calculated the same way every month, so the trend is real. Use a survey to ask a better question, never as a target.

More useful than any outside median is your own number, calculated the same way every month, so the trend is real.

1. Revenue per full-time equivalent

Measures: how much revenue each person in the firm supports. Calculate: trailing 12 months of revenue from your ledger, divided by average full-time equivalent headcount over the same period (count a three-day-a-week employee as 0.6, and count yourself). Moves it: automation that removes hours, better-scoped jobs, a service mix that leans on higher-value work, or hiring ahead of demand (which pushes it down, temporarily and on purpose). Decision: if it falls for two quarters without a planned hire behind it, find where the hours went before you add another person. Our revenue-per-employee piece covers the tooling side.

2. Lockup days (WIP days plus debtor days)

Measures: how long, on average, work sits between being done and being paid for. This is the number that explains a tight bank balance in a good month. Calculate: WIP days = unbilled WIP at month end, divided by trailing 12 months of revenue, times 365. Debtor days = accounts receivable at month end, divided by the same revenue figure, times 365. Add them together. Xero Practice Manager (XPM) has WIP reports and a KPI dashboard, and FYI has a WIP lockup report. Link Academy, which trains firms on the product, warns that its debtor figures can be unreliable, so take receivables from the Xero ledger itself. Moves it: invoicing on completion rather than at month end, collection terms, and how firmly you chase. Decision: rising WIP days means a billing habit problem; rising debtor days means a collection problem. Getting paid before you send the draft attacks both.

3. Realisation

Measures: for firms that record time, the share of the recorded value that actually gets billed and collected. Calculate: amount invoiced for a job (or collected, if you want the stricter version) divided by the value of time recorded against it at your standard internal rates. Karbon’s analytics and the Xero Practice Manager KPI dashboard both report it. For fixed-price work, run the same arithmetic with hours: the fixed amount divided by hours logged gives you an effective return per hour you can compare between jobs and clients inside your own firm. Moves it: scope creep, jobs quoted before the records were seen, rework, and a team that records time honestly (which will lower the number before it helps it). Decision: sort clients by realisation, lowest first. The bottom ten are your repricing or rescoping list for the next renewal.

The bottom ten are your repricing or rescoping list for the next renewal.

4. Utilisation

Measures: the share of paid hours spent on client work. Calculate: chargeable (client) hours divided by available hours (contracted hours minus leave and public holidays), per person and for the firm. Every practice-management system with timesheets can produce this. Moves it: admin load, internal meetings, training, onboarding, and seasonality. Decision: read it next to realisation. High utilisation with low realisation means the team is busy on work that does not pay. Consistently very high utilisation for one person means that person is the bottleneck, and possibly a resignation risk.

High utilisation with low realisation means the team is busy on work that does not pay.

5. Fixed-price overrun rate

Measures: how often fixed-price jobs take longer than the hours you budgeted. Calculate: jobs completed this month where logged hours exceeded the budget, divided by all jobs completed. Any system that records a budget against a job will show this (Kloud Connect’s Power BI dashboards and Xero Practice Manager’s job reports are two routes). Moves it: late or messy client records, unclear scope, and junior staff on work that needs a senior. Decision: if one job type overruns repeatedly, fix the scope or the process before the next cycle.

6. Client retention

Measures: whether the client base is holding. Calculate: two versions. Count retention = clients at the start of the year still active at the end, divided by clients at the start (exclude new ones). Revenue retention = revenue this year from clients who were with you last year, divided by what those same clients paid last year. A client export from AccountKit, FYI or your ledger’s customer list is enough. Moves it: service experience, deliberate offboarding of poor-fit clients, and price changes. Decision: falling count with rising revenue retention can be healthy (you let small clients go on purpose). Falling revenue retention needs an exit conversation with every client who left.

Falling revenue retention needs an exit conversation with every client who left.

7. Client concentration

Measures: how exposed the firm is to a handful of relationships. Calculate: revenue from your top five and top ten clients as a share of total revenue, from a sales-by-customer report in Xero or QuickBooks. Moves it: growth in one relationship, or loss of smaller clients. Decision: if one client could leave and take a large slice of revenue with it, plan for that now: document the work, spread the relationship across two people, and grow elsewhere.

8. Wage-to-revenue ratio

Measures: how much of every dollar of revenue goes to the people who deliver it. Calculate: total wages plus on-costs (superannuation, payroll tax or employer contributions, contractors doing client work) divided by revenue, from the profit and loss report. Include a notional wage for yourself if you do client work and don’t pay yourself one. Moves it: pay rises not matched by revenue, hiring ahead of demand, and offshore or contract support. Decision: a rising ratio with flat revenue per full-time equivalent means productivity has not kept pace with payroll. That points back to numbers 3, 4 and 5.

9. Profit after a notional owner wage

Measures: whether the firm is a business or a well-paid job. Calculate: net profit, minus what it would cost to employ someone to do your role. Divide by revenue for a margin. Moves it: every number above. Decision: if this is thin or negative, the firm cannot run without you, and it will be hard to sell. Treat that as the first problem to solve, ahead of growth.

Pull these on the same working day each month, from the same reports, and keep the last 12 months side by side. The trend matters more than any single reading.

Number	Source	Watch for	Decision it triggers
Revenue per FTE (trailing 12 months)	Ledger plus headcount	Two falling quarters	Find the lost hours before hiring
WIP days	PM system WIP report	Rising	Invoice on completion
Debtor days	Ledger aged receivables	Rising	Tighten terms and chasing
Realisation (or effective return per hour)	PM time and invoices	Bottom ten clients	Rescope at renewal
Utilisation	Timesheets	Gap from realisation	Cut non-paying work
Fixed-price overrun rate	Job budgets versus hours	Repeat job types	Fix scope or process
Client and revenue retention	Client list, sales by customer	Revenue retention falling	Exit conversations
Top-ten concentration	Sales by customer	One client dominating	Spread the relationship
Wage-to-revenue	Profit and loss	Rising with flat productivity	Review 3, 4 and 5
Profit after owner wage	Profit and loss	Thin or negative	Fix before growing

If you are about to make your first hire, our first-employee guide shows which of these to watch most closely in year one.

This is general information about managing a practice, not financial or legal advice.

How often should a small firm review its KPIs?

Monthly for the dashboard, pulled on a fixed day so every month is comparable. Quarterly, look at the 12-month trend with anyone who shares the decisions. Weekly reviews mostly chase noise.

Do these numbers work for a firm that doesn’t record time?

Most do. Lockup, retention, concentration, wage-to-revenue and profit need no timesheets. Realisation and overruns need some record of hours, even a rough one on sampled jobs. One workable approach is to time a sample of jobs each quarter purely to keep those two numbers honest.

Why not just compare myself with an industry average?

Because the average belongs to a different set of firms. Survey populations differ by country, size, service mix and whether the owner draws a wage, and the definitions vary between studies. If you do buy a benchmark report, check how it defines each measure and recalculate yours the same way before comparing.

What is the difference between WIP days and debtor days?

WIP days measure the delay between doing work and invoicing it, which is inside your control. Debtor days measure the delay between invoicing and being paid, which depends on your terms and follow-up. Splitting them tells you whether to fix your billing habits or your collection process.

Which single number should I start with?

Lockup days, because it links directly to cash and needs only two reports. Once you trust that figure, add realisation (or effective return per hour) to see which clients and job types are eating your time.