Benchmark
How US CPA firms sound
Thirty-five firms, six pages each. Level with Britain, and a different phrasebook.
Thirty-five American CPA firms, read by the same code on the same day as their British counterparts, six pages each. The two countries' accountants score almost level. They do not write alike, and the difference is not the one you would guess.
What was measured
Six pages a firm, nobody named
Fifty firms were listed, from the national networks below the Big Four down to the regional practices. Seven block automated readers and are not here; one address had gone, four could not be reached on the day, and three had too little continuous prose across six pages for the method to score. Thirty-five remain. That is the market below the largest firms, and it is said here so nobody has to work it out.
Each was read the way a Wordprint reads a customer: the homepage, then up to five of its own about, services and story pages, measured together as one body of prose, 3,150 words for the median firm. Everything below comes from the published method, which scores four things and reports eighteen more. No firm is named, no page text was kept, and the UK benchmark this one is set against was run on the same instrument.
Same code, same day, both sides of the Atlantic.
The headline
A median of 73, and level with Britain
The sector averages 71.9 and the median firm scores 73. The bottom tenth is at 59, the top tenth clears 84, and one firm reaches 91 and the established band. Twenty-one firms are in the seventies and eighties, thirteen are in the middle band where a voice exists and is buried, and none falls below 40.
Set that against the British figures, 72.3 and 71.7, and the two countries' accountants are the same sector on this instrument. Both show their working: the median American firm puts 312 figures, dates, names or results into every ten thousand words, the British 345, and not one firm on either side is short of evidence. Both have good rhythm. Sentence-length variance runs at 8.6 here against 8.8, on sentences of seventeen and eighteen words. The professions read as people who are precise for a living, and that is worth as much in Ohio as in Oxfordshire.
Where they differ is furniture. Only 60% of the words on a median American firm's site sit in continuous sentences; the rest is cards, tiles and headings. The British figure is 76%. American accountancy sites are built like product sites. More boxes. Less talking.
The failure
Nothing lands: twenty-three firms out of thirty-five
The commonest failure is the British one. Landing is the share of sentences at five words or fewer, the ones that carry the point after the long sentence has done the explaining. Real voices measured between 15% and 29%. The median American firm manages 6.4%, better than Britain's 3.8% and still a third of the floor, and twenty-three of the thirty-five are under 8%.
The second failure is where the countries part. Eighteen American firms of thirty-five cross the line of sixty phrasebook terms per ten thousand words; the median is 62, the top tenth 134. The British median is 42, with nine firms over the line. American accountants write from a fuller phrasebook. It is also a different book.
The British firm is on a journey. The American one is a trusted advisor in a landscape.
Same profession, two phrasebooks, one missing short sentence.
The shared phrasebook
The words twenty of thirty-five firms share
These are the phrases used by a fifth or more of the firms measured, counted as firms rather than occurrences. The list is longer than Britain's and the top of it reaches further.
| Phrase | Firms using it |
|---|---|
| landscape | 20 of 35 |
| empower | 19 of 35 |
| excellence | 18 of 35 |
| leverage | 16 of 35 |
| trusted advisor | 15 of 35 |
| journey | 14 of 35 |
| holistic | 13 of 35 |
| streamline | 13 of 35 |
| best practice | 12 of 35 |
| thought leader | 9 of 35 |
| seamless | 9 of 35 |
| unlock | 9 of 35 |
| robust | 8 of 35 |
| tailored solution | 8 of 35 |
| innovative solutions | 8 of 35 |
Two of these are American in a way the British list is not. Both are titles the firm awards itself. A buyer who reads three shortlisted sites will meet them on two, and the third firm will have said something specific by accident. That firm scored 91.
Who is speaking, and how
The firm, on twenty-seven sites, with a long dash
On twenty-seven of the thirty-five sites the firm outnumbers the client: "we" and "our" at a median of 474 per ten thousand words against "you" and "your" at 245. Four sites let a person speak as "I", the same share as in Britain. Contractions run at 147 per ten thousand words. The British run at 120.
One measurement needs stating carefully, because it would be easy to misread. Long dashes appear on thirty of thirty-five American sites at a median of nineteen per ten thousand words. The British figure is under one. In the AI consultancy benchmark a rate of ten was worth remarking on, because British business English rarely uses the mark. American business English does: it is house style in the standard American manuals, and a dash in an American paragraph says nothing about who or what typed it. The method is not an AI detector. This is the sector that shows why a punctuation mark cannot be one.
What the top decile does
Three habits that separate 84 from 73
- It says what the firm does, in nouns. The top firms name services, industries and the size of client they suit. The middle describes commitment and partnership, which is the phrasebook by another route.
- It lands. The top tenth puts one sentence in six at five words or fewer. The bottom tenth manages one in forty.
- It writes more sentences and fewer tiles. The best-scoring firms are the ones whose pages are three quarters prose; the sector as a whole is three fifths.
None of that needs a rebrand. Read the services page aloud. Highlight the phrasebook. Let one sentence in six stop early.
Limits
What this benchmark is not
Not a ranking, and no firm is named here or in the data behind it, which stores numbers only. Not a quality judgement: the score cannot tell whether the advice on a page is right. Not an AI detector, as the dash figure above shows. And the largest firms are absent by their own choice, so what is measured is the sector that lets itself be read.
Measured 9 September 2026. Thirty-five firms, six pages each, numbers only.
Where does your firm sit?
Paste a page you have published into the free test, choose the sector, and get the same four measures beside the sector median in about a minute. A Wordprint turns the result into rules the whole practice, and the tools it drafts in, can follow.
Written by Ian, with Claude assisting, and scored by the smarticulate method at 98.2 out of 100 before it went up. Ian spent twenty-odd years turning thinking, his own and other people's, into something a board would act on, and built smarticulate to put numbers on it. His practice is ConsultingPartner.io: enterprise experience, right-sized. Who is he? →
More guides: the UK accountancy benchmark, the UK agency benchmark, how to measure brand voice.