Write ESG review findings that distinguish evidence from uncertainty

Fictional Site R example: the 27 September summary reports seven actions complete; the 30 September register labels five closed and two open. These are document statements, not verified outcomes. The manager’s explanation was uncorroborated at the 2 October cut-off. Actual completion and effectiveness remain unestablished: reconcile the sources and obtain supporting evidence before relying on the summary.

A useful ESG review finding lets the reader trace each statement to its source and see where the evidence stops. Missing records, inconsistent documents and management explanations can all matter. Each supports a different kind of statement.

This guide helps teams write findings from an agreed document review. It offers a practical writing method, not an OECD or IFC reporting format, a compliance determination or an assurance conclusion.

Start with the question the finding must answer

Name the entity or site, period, review question and information cut-off. Then identify what was examined: a complete register, selected records or a management interview. A claim about a whole business needs a different evidence base from a description of three supplied documents.

IFC Performance Standard 1 calls for adequate, accurate and objective evaluation and presentation by competent professionals in identifying risks and impacts. See paragraph 19, printed page 6. That principle informs the writing discipline here; citing it does not establish that the standard applies to a particular business or that this review meets it.

Use wording that shows what supports the statement

Apply these four descriptions to individual statements. One finding may contain all four:

• Supported within the reviewed scope: state the narrow fact the evidence establishes. “The supplied register labels five entries closed” describes the register. It does not establish that five actions were successfully completed.

• Attributed account: identify who said what and when. “The operations manager stated on 1 October that all actions were complete.” Say whether that account was corroborated.

• Unresolved inconsistency: identify both sources, their dates and the precise difference. Preserve the disagreement until there is a supported explanation; do not silently choose the more convenient account.

• Information unavailable or unexamined: describe the limitation accurately. Distinguish a requested record not provided by the cut-off from a record received but outside the agreed review. Neither establishes that the underlying activity never occurred. A missing record alone does not establish noncompliance.

Use verbs such as “records”, “states”, “shows” and “could not establish” with a named subject. Avoid “verified” unless the finding explains the verification performed. A document's existence, its contents and the real-world condition it describes are separate propositions.

A fictional seven-action inconsistency

Suppose a desk review asks whether management's status summary is supported by the supplied records for Site R. The information cut-off is 2 October 2026. The following records and people are entirely fictional:

• D1: a management summary dated 27 September, page 2, reports all seven environmental actions complete.

• D2: the action register, version 3 dated 30 September, rows A1–A7, labels A1–A5 closed and A6–A7 open.

• I1: in an interview on 1 October, the operations manager says A6 and A7 were completed on 26 September and the register has not been updated.

• Review limitation: completion records requested for A6 and A7 were not provided by the cut-off. No site work or effectiveness testing was included.

The document inconsistency is established. The reason for it and the actual completion status remain unresolved. The manager's explanation is plausible, but this review has not corroborated it.

A traceable finding: “F-04: The supplied records do not consistently support the seven-action completion statement for Site R. D1 (27 September, page 2) reports seven complete; D2 (version 3, 30 September, rows A1–A7) labels five closed and two open. In I1 (1 October), the operations manager attributed the difference to an outdated register and stated that A6–A7 were completed on 26 September. Requested completion records for those two actions were not supplied by 2 October. Actual completion and effectiveness were not established through this desk review. Before relying on D1 for the management decision, ask the operations owner to reconcile the sources and provide dated support for the proposed status.”

This wording preserves the supported observation, attributed explanation, limitation and follow-up. It avoids declaring either that two actions failed or that all seven succeeded. The five closed labels also remain register entries, not five independently confirmed outcomes.

Explain why the uncertainty matters

Connect the finding to the decision: here, management cannot yet rely on the summary as a reconciled account of the seven actions. Name the specific question that remains open and the person responsible for answering it. Record an agreed response date rather than inventing one.

Keep the strength of the evidence separate from the potential significance of the underlying issue. Counting documents received or register entries marked closed does not, by itself, justify a low-risk rating. If a rating is used, state the agreed criteria, supporting reasoning and limitations alongside it.

The OECD guidance says information or claims about actual or potential adverse impacts need not be fully verified before prompting further assessment. See Question 22, printed page 65. Uncertainty should therefore shape the follow-up, rather than automatically lowering its priority. A potentially urgent concern needs the appropriate escalation route while the evidence is assessed.

Check the wording before issuing the finding

1. Can another reviewer locate every source and the exact passage or entry used?

2. Does each sentence distinguish a source's content from a conclusion about actual practice?

3. Are conflicting accounts and uncorroborated explanations still visible?

4. Are missing information, exclusions, dates and sample limits precise?

5. Does the conclusion stay within the work performed, with a specific follow-up where needed?

6. Does the management summary retain the qualifications in the detailed finding?

For help organising an agreed review of records and business practices, a findings summary, an evidence-based risk and gap register, and follow-up questions, see Nitium's ESG due diligence support or discuss your review. Legal opinions and independent assurance are outside this advisory scope. Questions requiring specialist judgement can be organised using the specialist handoff guide.

Primary references checked 13 October 2026. The evidence descriptions, example and wording checks are illustrative working methods.