Reporting year 2025 · Published Q1 2026

Annual Transparency Report 2025

What we did, how often we said no, how often we were wrong, and what we are changing. Published whether or not the numbers flatter us — that is the only version of this document worth issuing.

Signed by

  • Chief Executive Officer
  • Head of Verification

Both signatories are accountable for the accuracy of these figures. Figures in this prototype are illustrative and pending sign-off.

Headline

1,842 verifications. We said no in 210 of them.

1,842Verifications completed
11.4%Failed — a red line was triggered
23.7%Qualified — conditions attached
98.6%Delivered within stated SLA

Outcome distribution

Verified Premium (90–100)
14.1%
Verified Plus (80–89)
31.2%
Verified (70–79)
19.6%
Qualified (60–69)
23.7%
Not verified (<60)
11.4%

Volume by tier

Essential
612
Standard
903
Comprehensive
327
With on-site audit
219

219 files combined a desk verification with a physical site visit. On-site audits were conducted in 14 provinces.

Refusals

Why we said no.

The 210 failures broke down as follows. A single file can trigger more than one red line; the primary reason is counted.

Red lineReasonFilesShare
R2Fabricated or altered certification8641.0%
R4Entity does not exist / licence revoked4722.4%
R6Obstruction of an audit2411.4%
R3Sanctions or denied-party match2110.0%
R1Forced or child labour finding188.6%
R5Undisclosed safety-critical subcontracting146.6%

The single most common reason we refuse a supplier is a certificate that does not exist at the issuing body. Not a missing certificate — a forged or altered one. It is why we authenticate rather than copy.

Where the work was

Verification volume by province.

Concentration reflects China's manufacturing geography rather than our marketing reach. Where we have thin coverage, we say so.

Guangdong
41%
Zhejiang
22%
Jiangsu
12%
Shandong
8%
Fujian
7%
Other provinces
10%

Who commissioned the work

  • Retailers & importers — 46%
  • Brand owners — 21%
  • Government & public procurement — 14%
  • Marketplaces & platforms — 9%
  • Industrial & component buyers — 7%
  • M&A and legal advisers — 3%

Largest single buying account represented 19% of revenue — below the 25% disclosure threshold.

Being wrong

Appeals, corrections and refunds.

The numbers that matter most are the ones where we were wrong.

Appeals
41

Lodged by suppliers. 6 upheld, 9 partially upheld, 26 rejected.

We lost
15

Files changed after appeal — 36.6% of appeals resulted in a change to the report.

Corrections
9

Material factual corrections issued, all within the 5-working-day SLA.

SLA refunds
26

Files delivered late. Fee refunded in full, per our own terms.

Where we were wrong: 4 of the 6 upheld appeals concerned capacity conclusions drawn from registered data without on-site triangulation. In 2026 we now flag any capacity conclusion above a materiality threshold as audit-recommended rather than stating it as verified.

Independence in practice

Conflict declarations, attestation and revenue.

100% attestation

All staff and auditors completed the annual independence attestation. No refusals.

2 declared conflicts

Both resulted in recusal from the affected file and reassignment. Neither file was delayed beyond SLA.

3 concerns raised

Raised via the independence channel. Two unsubstantiated; one resulted in a process change around gift handling.

0 supplier-side revenue

No listing, badge, placement, audit-travel or consultancy income from any assessed supplier.

219 audits, 0 subcontracted

Every site visit was performed by a salaried Business Verify auditor.

19% revenue concentration

Largest single buying account, disclosed here as required above 15%.

For 2026

What we committed to change.

Capacity conclusions must be labelled by evidence grade
Following four upheld appeals. A capacity conclusion is now stated as "registered", "assessed" or "observed" so a buyer can see immediately whether anyone stood on the floor.
Faster turnaround on Essential at panel scale
Target: 3 working days for Essential within a multi-supplier panel, down from 3–5. Requires staffing we committed to in Q1.
Publish refusal rates by province
Buyers asked whether certain clusters carry systematically higher risk. We will publish the distribution, with a caveat about selection bias in who gets verified.
Extend appeals transparency
From 2026, upheld appeals will be described individually in the annual report rather than only in aggregate.

This report covers the calendar year 2025. Figures in this prototype are illustrative placeholders pending internal sign-off and external assurance.

Assurance over these numbers.

An independent firm examines our controls annually, including the conflict register and the data behind this report.