Veracent Intelligence — Research Methodology

Version 1.0 · Effective 30 August 2026 · Published by Veracent Intelligence

This page explains exactly how a Veracent report is produced. It is published so that any reader, and any company we write about, can see the criteria we apply, the sources we rely on, the weightings we use, and the arithmetic behind every figure we publish. We apply this methodology uniformly. Where we depart from it, we say so in the report.

---

1. What we measure, and what we do not

Veracent measures the availability, quality and consistency of publicly verifiable evidence about a financial company or product.

We publish figures describing our evidence. We do not publish a single composite score purporting to rate a company's overall trustworthiness.

This distinction is deliberate and material:

Absence of evidence is not evidence of wrongdoing. A low verification percentage means we could not confirm something from public sources. It does not mean the company is deficient, and it must not be read that way. Many well-run firms disclose little publicly; some poorly-run firms disclose a great deal.

Veracent reports are not: investment advice, a recommendation to use or avoid any product, a solvency or insolvency prediction, an audit, a regulatory opinion, or a statement that any company has broken any law or rule. We are not authorised or regulated by the Financial Conduct Authority and we do not provide regulated advice.

---

2. Scope of every report

Every report states, on its face:

• The legal entity examined — including company registration number and, where applicable, Firm Reference Number. "Trading 212" is not an entity; Trading 212 UK Ltd is. Group companies are examined separately because their regulators, permissions and client protections differ.
• The jurisdiction and regulator the findings apply to.
• The account or product type, where terms vary by product.
• The evidence window — the date range of sources relied on.
• The publication date and last refresh date.

A finding true of one entity in one jurisdiction is not asserted of any other. Where a reader may reasonably confuse two entities, we say so explicitly.

---

3. Source hierarchy and reliability weights

Every source is classified into a tier. The tier determines its weight, denoted w, in the calculations at section 6.

Three rules govern source use:

1. A Tier 5 source is never the sole basis for a published finding. It may corroborate or flag an area for investigation only.
2. Tier 2 outranks Tier 4 for contractual terms. A fee stated in the client agreement is authoritative over the same fee stated on a marketing page. Where they differ, this is recorded as a contradiction (section 7).
3. Every source is canonicalised. Syndicated copies, mirrors and reprints of the same underlying material count once. Republication does not increase confidence.

Each source in a report is listed with its publisher, URL, publication date, tier and weight, so the reader can check our classification.

---

4. Evidence dimensions and criteria

Reports are organised into six dimensions. Each contains a fixed, published list of criteria. Criteria that do not apply to a given entity or product are marked Not Applicable and removed from the denominator — never scored as failures.

1. Regulatory Standing — authorisation status, regulator, permissions held, appointed representative status, trading names, registered address, disciplinary and enforcement history, waivers and exclusions.
2. Client Asset Protection — client money segregation arrangements, custodian identity, depositor or investor compensation scheme coverage and limits, evidence of independent custody assurance.
3. Cost Transparency — headline fees, FX and conversion charges, spreads and implicit costs, account and inactivity charges, withdrawal and transfer costs, conditions that materially alter stated pricing.
4. Security and Operational Controls — published authentication controls, incident history and disclosure, operational resilience disclosures, business continuity statements.
5. Corporate Transparency — ownership and control, group structure, filing currency and timeliness, auditor identity, going-concern statements.
6. Conduct and Track Record — regulatory actions, upheld ombudsman complaints where published, litigation of material significance, documented incidents.

The full criteria list under each dimension is published at [/methodology/criteria].

---

5. Verification states

Each criterion is assigned exactly one state:

"Not Found" is reported neutrally. It carries no negative inference and is excluded from any statement about the company's conduct.

---

6. The arithmetic

All published figures derive from the following. Nothing is estimated, rounded in our favour, or produced by unaudited model judgement.

6.1 Criterion verification strength

For a criterion i supported by one or more sources:

s_i = min( 1.0 ,  w_max × r_i × c_i )

Where:

• w_max = the weight of the highest-tier source supporting the criterion (section 3).
• r_i = the recency factor (section 6.2).
• c_i = the corroboration factor: c_i = 1 + 0.10 × (n − 1), capped at 1.20, where n is the number of independent sources in agreement. Sources sharing an owner, author or underlying origin count as one.

For Not Found criteria, s_i = 0. For Contradicted criteria, s_i = 0 and the conflict is published in full.

6.2 Recency factor

Evidence decays at different rates depending on what it describes:

Stale evidence is either refreshed or the criterion is reset to Not Found. It is never carried forward silently.

6.3 Published dimension figures

For each dimension d with N_d applicable criteria:

Coverage — the proportion of criteria on which we located any qualifying evidence:

Coverage_d = (N_d − NotFound_d) / N_d

Verification Strength — the mean strength of the criteria we did address:

Strength_d = ( Σ s_i ) / (N_d − NotFound_d)

Official Source Ratio — the proportion of addressed criteria resting on Tier 1 or Tier 2 evidence:

OfficialRatio_d = Verified_d / (N_d − NotFound_d)

Coverage and Strength are reported separately and never multiplied into a single figure, because they mean different things: Coverage tells you how much we found; Strength tells you how good it was.

6.4 Report-level figures

EvidenceCompleteness = Σ(N_d − NotFound_d) / Σ N_d          across all dimensions
OverallOfficialRatio = Σ Verified_d / Σ(N_d − NotFound_d)
ContradictionCount   = total criteria in Contradicted state

Every published percentage links to the criteria and sources that produced it. A reader can reconstruct any figure on this page from the report's own source table.

6.5 What we deliberately do not compute

We do not publish a composite trust score, a letter grade, a star rating, or a ranking of one company against another in Version 1. Such a figure would require us to weight dimensions against each other — a value judgement we cannot evidence, and one that would convert a factual statement about our research into an opinion about a named company's character. If we introduce one, we will publish the weightings and the reasoning here first, and version this page accordingly.

---

7. Contradictions and validation

Every report passes through a second, independent validation stage that reviews the report against the underlying evidence only. Its function is adversarial: to identify conclusions the evidence does not support.

It checks each finding for: unsupported inference, contradiction between sources, duplicate evidence inflating apparent corroboration, stale evidence, and reliance on Tier 4 or Tier 5 material where a higher tier was required.

Any finding the validation stage cannot trace to a cited source is removed before publication. Where sources genuinely conflict, we publish the conflict rather than resolving it in either direction.

---

8. Right of reply and corrections

Any company that is the subject of a Veracent report may:

1. Request the underlying evidence for any finding. We will provide the source list for that finding within 5 working days.
2. Submit a correction or additional evidence to corrections@veracentintelligence.co.uk. Submissions are assessed against this methodology like any other source, weighted by tier. Company submissions are Tier 2 where they are binding documents, Tier 4 where they are statements.
3. Have a factual inaccuracy corrected. Where we are wrong, we correct the report, log the change, and state on the report that a correction was made and when.
4. Have a published response attached. Where we disagree, the company may supply a statement of up to 300 words which we publish alongside the finding, unedited.

We aim to acknowledge all correspondence within 5 working days. Correction requests are handled before any escalation is necessary, and we would prefer to hear from a company directly than from its solicitors.

Correction log: every material change to a published report is recorded at [/corrections] with the date, the finding affected, and the reason.

---

9. Refresh cadence

• Reports display their last refresh date prominently.
• Fast-changing criteria are re-checked at least every 90 days on actively published reports.
• Slow-changing criteria are re-checked at least every 12 months.
• Where a report has not been refreshed within its stated window, it is marked Out of Date and its fast-changing figures are suppressed rather than displayed stale.
• Score and figure changes between versions are logged and visible, with the reason for the change.

---

10. When we withhold or withdraw a report

We will not publish, or will withdraw, where:

• Evidence Completeness falls below the minimum threshold for a meaningful report (published per dimension at [/methodology/criteria]);
• The entity cannot be resolved with confidence to a specific registered company;
• A material contradiction cannot be resolved and publishing would mislead;
• A correction request reveals a systematic error in our process, pending review.

---

11. Limitations we want you to know about

Stated plainly, because a methodology that only lists strengths is marketing:

• We only see public evidence. Firms disclose different amounts. Our figures partly measure disclosure practice, not conduct.
• Automated extraction can misread a source. The validation stage at section 7 catches much of this, but not all. This is what the right of reply exists for.
• Entity resolution is hard. Group structures change. We may occasionally attribute a finding to the wrong entity in a group; tell us and we will correct it.
• Our source tiers embed a judgement — that regulators outrank press, and press outranks marketing. That is a defensible judgement, not a neutral fact, and we state it openly.
• Nothing here predicts the future. A firm with complete disclosure today may fail tomorrow.

---

12. When this methodology changes

A criterion or rule is amended only where it is logically incapable of being applied correctly — for example, where the evidence standard it demands could not answer the question it asks, whatever the facts. It is never amended because a particular report fell short of it.

Publication thresholds are not adjusted to admit a specific report. Where any change is made, every affected report is re-assessed under the new version and must pass validation again, and the reason is recorded below.

13. Version history

Questions about this methodology: methodology@veracentintelligence.co.uk

---

Veracent Intelligence Ltd is a research and information publisher. Nothing on this site constitutes financial, investment, legal or tax advice, or a recommendation to enter into any transaction. Readers should conduct their own due diligence and take professional advice appropriate to their circumstances. Veracent Intelligence Ltd is not authorised or regulated by the Financial Conduct Authority.