Corrections Log

Every change to a published figure or statement: date, what it was, what it is now, and why.

← Back to Archive Home

Superseded versions remain available through web archive captures. Corrections are listed newest first and are never removed.

Tribunal figures corrected and unfinished sections removed

Was'Under 2%' of cases involved employment status disputes; '530 cases analysed'; unfinished editorial notes visible in the sections on conversations with HMRC and ACAS and on HMRC enforcement lists
NowOver 500 cases reviewed; 363 logged in the tracker, of which 28 (7.7%) are flagged as employment status disputes; the two unfinished sections removed
WhyThe 2% figure was not supported by the tracker (28 of 363 is 7.7%). The notes were published by mistake, and the statements in those two sections were not yet backed by a record.
EvidenceTribunal case tracker (363 cases logged, 28 flagged as employment status)

Naming-claim figure corrected and verdict changed

Was'47.9% of all companies use at least one of these five words'; verdict 'the number checks out'; conclusion said one claim checks out numerically; growth described as 'post-pandemic'; canonical address pointed to the old filename
NowThe five words together cover at most 42.8% of companies (counts sum to 67,842 of 158,433); verdict is 'not a like-for-like test'; conclusion says none of the five testable claims is confirmed; 'post-pandemic' wording removed; canonical address corrected; dated revision notice added
Why47.9% was arithmetically impossible: it exceeds the sum of the five individual shares. The 'post-pandemic' wording was inconsistent with the company-demographics page.
EvidenceThe page's own table: 26,094 + 18,255 + 8,818 + 7,577 + 7,098 = 67,842, against 158,433 companies (Companies House snapshot of 24 Aug 2026)

Rounding, CBILS dates and wording of the loan figures made accurate

WasTotal lent £1,035.7m; repaid £524.4m; CBILS final payments '2027-2028'; outstanding balances rated 'Very High' confidence and described as operating businesses; claims settled described as failed businesses; last updated 'October 2025'; page said an FOI request for Pay As You Grow uptake data had been submitted
Now£1,035.8m; £524.5m; CBILS end dates described as not given in the FOI response; outstanding balances rated 'High' and described as an estimate; claims settled described as guarantee claims paid by government; revision notice and last-updated date added; FOI response linked; statement about a pending FOI request removed (none is outstanding)
WhyTotals are now rounded from the exact FOI row totals. The 2027-2028 CBILS dates conflicted with a 6-year term from 2020-21 drawdown and are not in the FOI. The Bank holds no write-off data, so the outstanding balance is an estimate and 'failed businesses' overstated what the data shows.
EvidenceBritish Business Bank FOI 25-002 row totals, re-added
2026-10-16 · index.html

Broken links, VAT threshold and independence statement corrected

WasLinks to high-street-laundromat-tested.html and faq.html, neither of which exists; COVID card listed a regional breakdown; VAT-registered tier said 'above £90,000 (VAT threshold)' for 2023-24 data; independence statement said 'no funding from… commercial entities' without declaring the author's own commercial interest; background given as 'salon industry software developer'
NowLinks to highstreet-laundromat-claims.html and sources/data-sources-faq.html; COVID card lists what the page contains; threshold stated as £85,000 until 31 March 2024 and £90,000 from 1 April 2024; independence statement now says the archive is produced through Salon Logic Ltd, which also develops SalonLogic Pro, and is not funded by any outside organisation
WhyFound when checking the live site; the card described content the page does not have; the VAT threshold rose from £85,000 to £90,000 on 1 April 2024, so £90,000 was wrong for most of the 2023-24 data year
EvidenceLink check of the live site, 16 Oct 2026; GOV.UK VAT registration threshold

UK consumer spending on hairdressing and personal grooming revised to the ONS release of 30 September 2026

Was2025: £10.26bn nominal; real 2025 9.7% below 2019; gap to formal turnover £4.16bn
Now2025: £9.29bn nominal; real 2025 17.3% below 2019 (2024 prices); gap £3.19bn
WhyONS revised the CCRE and CCHP series. Statements that no longer hold (four consecutive quarterly declines; nominal falling three straight quarters; nominal back at 2019 level by 2023) were withdrawn.
EvidenceONS series CCRE and CCHP, release 30 Sept 2026; dated revision notices on each page
2026-10-15 · index.html

Homepage brought into line with the business-data and company pages

WasVAT tier 2022-23 (£3.15bn, 14,900); formal tier labelled March 2024; FOI described as pending
NowVAT tier 2023-24 (£3.42bn, 14,480); formal tier March 2025; one-year gap described; company register card updated to the 19 Sept 2026 snapshot (159,404 companies, 78,304 active; was 158,433 and 78,005)
WhyPages had drifted out of sync; one unbalanced HTML tag removed
EvidenceHMRC FOI2026/87008; ONS AH1765/AH1788

Outstanding-balance totals corrected so they equal the sum of their own rows

WasBBLS £206.8m, CBILS £6.9m, total £213.7m, 79.4% resolved
NowBBLS £236.8m, CBILS £7.9m, total £244.7m, 76.4% resolved
WhyPublished totals did not equal the sum of the band rows above them (differences of £30.0m and £1.0m)
EvidenceRow sums re-added and checked against British Business Bank FOI 25-002 (15 Oct 2026)

Summary cards, repayment dates and metadata aligned

WasClaims settled £266.1m (25.6%); CBILS lending from £500k+ businesses 71%; BBLS final payments 'May 2027'; dataset metadata referring to furlough/SEISS and HMRC
Now£266.6m (25.7%); 73%; final payments May 2026 to March 2027 (6-year term from drawdown, 12-month holiday included); metadata describes BBLS/CBILS and the British Business Bank
WhyCards disagreed with tables; BBLS term is six years including the 12-month holiday; page holds no furlough or SEISS data
EvidenceGOV.UK Pay As You Grow announcement; FOI 25-002

Errors and inconsistencies corrected

Was'25,000 PAYE jobs' lost in 2020-21; cross-check section numbered 6 but cited as Section 8; summary line giving UK 52% self-employed and a 34-58% range; 'covers complete population, not samples'; canonical link to employment-analysis.html
NowPAYE fell 26,000 (total workforce 25,000); section renumbered and cross-references updated; economy-wide 13.4% only (World Bank/ILO); SPI described as sample-based; canonical link to this page
WhyInternal inconsistencies and one incorrect description of the Survey of Personal Incomes
EvidencePage's own tables; HMRC SPI documentation

Formation-trend and ranking statements corrected

Was2025 'highest full year on record'; 2026 'on pace to exceed 2025'; growth described as post-pandemic; age bands described as established businesses
Now2023 was the peak (15,953), 2025 second; 2026 run-rate about 15,100; most growth before 2020 and flat since 2022; age bands labelled as time since incorporation, including dissolved companies
WhyThe page's own table contradicted the statements
EvidenceCompanies House snapshot of 19 Sept 2026

Test wording and sourcing made accurate

Was'None of six claims replicate'; 'does not replicate' for claims about SmartSearch's unpublished list; '14-20 companies each' with an all-time table; '151,000 unauthorised addresses' unsourced; 'eight independent sources' unlisted; tracker attributed an address-sharing fall to a denominator effect
NowNone of the five testable claims is confirmed (see the 16 Oct 2026 entry on the naming claim); verdicts say what was tested; table described as all-time; 151,000 sourced to the Companies House annual report; two sources named; unexplained fall flagged as unexplained
WhyVerdicts and sourcing overstated what the work shows
EvidenceCompanies House Annual Report 2025-26; Crunch and Gateley published explanations