Revision notice, 16 October 2026
The verdict on Claim 6 (words in company names) has been corrected. This page previously gave a register-wide figure of 47.9%, which cannot be right: the five words listed together cover at most 42.8% of companies. The conclusion has been updated to match. The conclusion and verdict wording were also revised on 15 October 2026. See the Corrections Log.
Data used: the complete Companies House register for SIC 96020 (Hairdressing and other beauty treatment), all 158,433 companies ever incorporated, every status, pulled via the Companies House Advanced Search API on 24 August 2026.
Where We Stand
This is not a claim that illicit finance in cash-intensive sectors doesn't exist
Nothing in this piece argues that money laundering never happens through small cash-handling businesses, or that Companies House abuse isn't a real problem — it plainly is, which is exactly why Parliament passed the Economic Crime and Corporate Transparency Act, and why Companies House has removed 151,000 registered office addresses from the register in the year to 31 March 2026 as part of tackling address misuse (Companies House Annual Report and Accounts 2025 to 2026).
What this piece argues is narrower and, we think, more important: a blanket statistical narrative about an entire sector is not the same thing as evidence against any specific business, and should not be treated as such. If real bad actors are operating in this sector — and in a population of 158,433 companies spanning over a century, some almost certainly are — the correct response is targeted investigation by the bodies with the power and the duty to carry it out: the National Crime Agency, Companies House's own enforcement function, HMRC. Not a population-level statistical inference published by a compliance software vendor, picked up by trade press, and left to attach itself to every business in the sector by association.
The data deserves to be interrogated vigorously in either direction. That's what the rest of this page does. And if it turns up something that looks like a genuine, specific case worth investigating, the right next step is to hand it to someone with the authority to investigate it properly — not to publish a list ourselves.
The Claims Being Tested
The report's hairdressing-relevant claims, as published:
- "Suspected shell companies" in the data have an average lifespan of 170 to 194 days, against almost five years for the wider register of dissolved UK companies
- Suspect incorporations rose more than 340% between 2016-2018 and 2023-2025
- A single area of Cardiff contains 119 suspected companies on its own, across both sectors
- A single London formation agent address hosted 17 separate hairdressing and beauty companies between 2018 and 2023, "all with near-identical lifespans and dissolution patterns"
- Almost half of the suspected hairdressing companies use one of just five generic words in their name
- Conservative modelling suggests £310m-£464m may have moved through the 3,097 identified companies; extrapolated to over £1bn across five sectors over a decade
Every claim above that can be tested against the public register has been tested. The one that can't — the financial modelling — is addressed on its own terms in that section.
Claim 1: The 170-194 Day Lifespan
| Metric | Value |
|---|---|
| Dissolved SIC 96020 companies with a valid lifespan | 79,549 |
| Median lifespan | 755 days (~2.1 years) |
| Companies falling exactly in the claimed 170-194 day window | 486 (0.61%) |
| Companies dissolved within 194 days (the upper bound) | 2,519 (3.17%) |
Not representative of the sector. The report's figure describes its own unpublished list of suspected companies, so this is context rather than a direct refutation. The sector's real median lifespan is well short of a multi-year average — hairdressing is a genuinely high-churn, low-capital sector — but nowhere near the claimed 170-194 day shell window. That window accounts for a small minority of all dissolutions, not a representative pattern.
Claim 2: 340% Growth in Incorporations
| Period | Incorporations |
|---|---|
| 2016-2018 | 20,377 |
| 2023-2025 | 47,417 |
| Growth | 133% |
Not true of the sector as a whole. Real sector-wide growth over this period is 133% — substantial and real, but well under half the claimed 340%. We don't know what population the 340% figure refers to; it isn't the sector as a whole.
Claim 3 & 4: Address Clustering as a Shell Signature
If short-lived companies were disproportionately shell registrations, they should cluster at a small number of addresses more than genuine businesses do.
| Group | Companies | Share address with 3+ others |
|---|---|---|
| Short-lived dissolved (≤194 days) | 2,519 | 10.1% |
| Longer-lived dissolved (>194 days) | 77,030 | 15.7% |
Runs backwards. Longer-lived, genuine businesses cluster at shared addresses more than short-lived ones — the opposite of what a shell-formation pattern predicts. The largest address clusters in the sector are large, well-known, mass-market company-formation-agent addresses used by tens of thousands of companies across every UK sector, not specific to hairdressing.
The specific "17 companies, near-identical lifespans" claim
The report doesn't name the address, so we tested every London address hosting a comparable number of hairdressing/beauty companies incorporated in that window. Four addresses matched on count (14-20 hairdressing and beauty companies incorporated there between 2018 and 2023). The table below shows every company ever registered at each address, not only that 2018-2023 cohort, so it is context for the claim rather than a like-for-like replication. None showed the described pattern:
| Address (type) | Lifespan range | Still active today |
|---|---|---|
| 60 Millmead Business Centre (shared office space) | 151 to 2,217 days | 4 of 19 |
| "Demsa Accounts", Green Lanes (named accountancy practice) | 428 to 4,344 days | 18 of 26 |
| 7 Bell Yard (formation agent / accountant) | 126 to 869+ days | 13 of 30 |
| 27 Old Gloucester Street (formation agent / accountant) | 106 to 3,079 days | 11 of 22 |
No matching address found. Every candidate shows lifespans varying by an order of magnitude within the same address, genuine distinct trading names (named stylists, specific barbershop brands, aesthetics clinics), and a large share still actively trading years later. This is what an accountant's or formation agent's registered-office client list looks like — not a "near-identical lifespan and dissolution pattern."
Claim 5: "A Single Area of Cardiff" — 119 Suspected Companies
This is the most specific geographic claim in the report, and testing it turned up something the report itself doesn't appear to account for.
Discovery: the Companies House default address
Searching the SIC 96020 register for Cardiff addresses, one postcode area — CF14 — accounts for 1,289 of the sector's 2,084 Cardiff-registered companies. That concentration is not one suspicious address; it's hundreds of ordinary Cardiff streets (North Road, Whitchurch Road, Llanishen, Llandaff North), each hosting one or two genuine local businesses.
But within CF14, 1,144 SIC 96020 companies — nearly ten times the report's claimed Cardiff total — share a single entry: "[Company number] — Companies House Default Address, Cardiff, CF14 8LH."
This is not a real business address. It is a standard, automated administrative placeholder that Companies House assigns to any company, in any sector, anywhere in the UK, when its registered office is judged not to be an appropriate address (for example after returned mail, a successful dispute, or an unauthorised-address finding), a duty and power strengthened by the Economic Crime and Corporate Transparency Act. Published explanations: Crunch and Gateley's ECCTA guide. Because a default address can be followed by strike-off, companies carrying it are likely to skew towards dissolved.
What this means for the claim
If an address-clustering methodology does not specifically exclude the Companies House default address as a known artifact, it will count every company whose real address was successfully removed under ECCTA as if it had chosen to cluster in Cardiff. Given the scale here — 1,144 companies in our sector alone, against a claimed total of 119 across two sectors — this artifact is large enough to be the dominant explanation for a "Cardiff cluster," not a marginal one.
We are not claiming to have proven the report's 119 figure is entirely this artifact — the report's methodology isn't published, so we can't trace their number directly. What we can say is that a far larger, well-documented, sector-agnostic administrative phenomenon exists at exactly the location named, dwarfs the claimed figure, and has nothing to do with hairdressing, Cardiff as a place, or coordinated shell-company address selection. Any claim about a Cardiff "cluster" needs to rule this out before it means anything.
Claim 6: "Almost Half... Use One of Just Five Generic Words"
Across the sector's 158,433 companies, the five most common words in company names are:
| Word | Companies | % of sector |
|---|---|---|
| BEAUTY | 26,094 | 16.5% |
| HAIR | 18,255 | 11.5% |
| AESTHETICS | 8,818 | 5.6% |
| NAILS | 7,577 | 4.8% |
| BARBER | 7,098 | 4.5% |
Together these five words appear in 67,842 company names, which is at most 42.8% of the register (fewer if some names contain more than one of the words). We have not yet counted how many companies contain at least one of them.
Not a like-for-like test. The report's "almost half" describes its own unpublished list of suspected companies, so it cannot be tested directly; across the whole register the share is at most 42.8%, so "almost half" is in the same range. The framing is the larger problem. These five words are not generic in any sense that implies evasion — they are the literal names of the services a hairdressing or beauty business provides. A company called "Sarah's Hair & Beauty Ltd" or "City Barbers Ltd" is not hiding behind vague language; it is doing the most basic, sensible thing a small business can do, which is name itself after what it sells. Presenting this as a suspicious pattern is equivalent to observing that many law firms use the words "law," "legal," or "solicitors."
The £310m-£464m Figure
Cannot be tested. This is not a claim we can confirm or refute from Companies House data, because it requires transaction-level banking information that the public register does not contain and that the report does not publish. The report states the modelling draws on "published estimates from the National Crime Agency and the National Economic Crime Centre," but does not show the calculation from those estimates to this specific figure.
This is a different kind of problem from the claims above, which could at least be compared with the register. This one simply has no public working to check. A number with no visible methodology is not evidence in either direction — it's an assertion, and should be weighted as one.
Conclusion
Of the six hairdressing-relevant claims in this report, five can be compared with the register. None of the five is confirmed: the lifespan and growth claims do not hold for the sector as a whole; the "17 companies" address could not be matched; the Cardiff figure cannot be traced and sits on a far larger default-address artifact; and the naming claim describes the report's own unpublished 3,097-company list, so it can only be set against the register-wide share, which is at most 42.8%. Two additional important findings emerged directly from testing: address clustering runs in the opposite direction to the one predicted, and the report's single most specific geographic claim sits on top of a well-documented government administrative artifact roughly ten times its own stated size.
None of this means the underlying concern — that cash-intensive small businesses can be misused for money laundering — is fabricated or unimportant. It means this specific report does not provide reliable evidence for that concern as it applies to hairdressing, and that a reader relying on it would be relying on numbers that don't hold up to a straightforward check against the primary source the report itself is built on.
The honest, responsible response to genuine concern about a handful of bad actors is not a sector-wide statistical narrative that puts every hairdresser, barber, and beauty salon owner in the country under an unstated cloud. It's for the National Crime Agency, Companies House, and HMRC — who hold the actual investigative powers and the actual transaction data this report doesn't — to pursue whatever specific, individual cases the evidence supports. That list, if one is warranted, isn't ours to draw up. It's theirs.
Methodology
Primary data
- Companies House Advanced Search API, full historical register, SIC 96020, all statuses, snapshot 24 August 2026 (158,433 companies)
What we tested and how
- Lifespan: days between incorporation and dissolution dates, all 79,549 dissolved companies with valid date pairs
- Growth: exact incorporation counts by year, full historical register (not an active-only snapshot, which would understate older years through survivorship bias)
- Address clustering: exact string match on registered office address field
- Cardiff address: postcode-area extraction (CF prefix) plus address-string search for "Companies House Default Address"; the default-address mechanism checked against two published explanations (Crunch; Gateley's ECCTA guide)
- Naming: tokenised word-frequency count across all 158,433 company names, legal suffixes (Ltd/Limited/LLP/PLC) stripped
What we could not test
- The £310m-£464m financial modelling, and its extrapolation to £1bn+ across five sectors — no published methodology to check against
- Whether the report's specific 3,097-company "suspected shell" list overlaps with what we've tested here, since that list itself isn't published