ESRS Disclosure Requirements, Datapoints and Application Requirements: How to Read the Standards

Prepared in British English as a practitioner Knowledge Card Package: answer, explanation, application, evidence, connections and publishing layer.

Many implementation projects begin with a datapoint workbook. That is understandable: a workbook can be filtered, assigned to owners and loaded into software. The risk is that the project starts treating each row as a self-contained legal requirement. Context is then lost.

A single datapoint may depend on the objective of the DR, a condition in the preceding paragraph, an AR, a defined term, a materiality conclusion, a phase-in, a relief or information elsewhere in the sustainability statement. Conversely, a single sentence may contain several distinct datapoints that require different owners and evidence. The job of the reporting team is therefore to preserve the legal and conceptual context while making the requirements operational.

Figure 1. Anatomy of an ESRS requirement. Branded educational visual by London Reporting Academy.

A disciplined reading sequence has nine steps.

1. Confirm the applicable legal text and reporting period. Determine whether the organisation is applying the 2023 ESRS or, once legally effective, the revised 2026 ESRS. Do not mix numbering or wording silently.

2. Identify the material topic and relevant IROs. A topical DR is not read in isolation from the double materiality assessment.

3. Read the DR objective. The objective explains the user need the disclosure is intended to meet. It is a powerful completeness test.

4. Read every operative paragraph in the DR. Note “shall disclose”, “shall report”, “shall describe” and “shall explain”.

5. Atomise the text into datapoints. Separate each distinct fact, method, boundary, period, assumption, explanation or amount.

6. Read all linked ARs. They may change the inputs, method, context or presentation expected.

7. Check glossary terms and cross-references. Defined terms may determine boundary, classification or measurement.

8. Apply materiality, reliefs and presentation rules. A prescribed datapoint is still subject to information materiality and any specific conditions.

9. Test the disclosure as a whole. Ask whether the statement achieves the DR objective and contributes to fair presentation rather than merely filling rows.

A DR commonly includes:

• a title and identifier;

• an objective explaining what users should understand;

• one or more operative “shall” paragraphs;

• conditions, exceptions or “if applicable” language;

• related ARs;

• cross-references to ESRS 1, ESRS 2 or a topical standard;

• links to policies, actions, metrics, targets, financial effects or specific topical content.

The objective is not decorative. Suppose a team has supplied every listed field but the disclosure still prevents a reader from understanding the scope, method or significance of the information. The objective signals that the drafting is incomplete even though the spreadsheet appears green.

DRs are not all applied in the same way

A cross-cutting DR may apply across all material topics or at undertaking level. A topical DR is generally activated by a material topic and may contain further conditions. Some information may be mandatory irrespective of the materiality assessment under the applicable text, while other information is subject to the materiality of the topic and the information. Current and revised ESRS differ in detail, so the project should maintain a version-controlled applicability rule for each DR.

A datapoint is the smallest distinct information item that can be requested, evidenced, reviewed and mapped to a report location. It is not necessarily a number.

Examples include:

• a policy name and approval date;

• the scope of a target;

• a calculation methodology;

• an explanation of why quantitative information is not useful;

• a metric value and unit;

• whether a target is required by law;

• the date on which the materiality assessment was last updated;

• a description of significant assumptions;

• a statement that no measurable outcome-oriented target has been set and how effectiveness is tracked instead.

A good data dictionary therefore classifies datapoints by type and control needs. Narrative datapoints need source evidence and drafting controls just as quantitative datapoints need calculations and reconciliations.

The revised ESRS explain that ARs are mandatory, support the application of the main requirements and have the same authority as the main body. ARs use “shall consider” to identify issues, resources or methodologies the undertaking is expected to take into account. They can also provide presentation options.

This means “shall consider” should not be converted into “shall disclose every listed item”. The organisation must demonstrate that it considered the specified matters in preparing the disclosure. The resulting reported information still depends on the operative DR, materiality and facts.

For project purposes, each AR should be tagged as one or more of the following:

• interpretation rule - clarifies the meaning of a requirement;

• method rule - specifies or constrains measurement or assessment;

• input rule - identifies evidence, sources or factors to consider;

• presentation option - permits narrative, tabular or other presentation;

• context rule - requires supporting information for understanding;

• example or illustration - helps interpretation but does not create a new universal fact pattern.

The phrase “subject to materiality” is not a licence to remove inconvenient information. The organisation needs a documented information-materiality judgement based on whether omission, misstatement or obscuring could reasonably be expected to influence user decisions or informed assessments.

ESRS disclosures are structured into topics, which may be disaggregated into sub-topics. The level of granularity should match the disclosure objective and the undertaking's facts. A material IRO may sit below a broad topic heading and require more specific analysis or entity-specific information.

Defined terms matter because they can control:

• whether an item is an impact, risk or opportunity;

• whether the reporting boundary includes own operations or value-chain information;

• how short, medium and long term are understood;

• whether an action is a target, policy, remediation or transition plan;

• what counts as a metric, estimate, affected stakeholder or business relationship.

In the revised ESRS, glossary-defined terms are signalled typographically in the standard. In an internal data request, preserve the official term and add a plain-English explanation rather than replacing it with a looser synonym.

The revised ESRS make fair presentation explicit. The sustainability statement should disclose relevant information about material IROs and represent them faithfully through a complete, neutral and accurate depiction. Information should also be comparable, verifiable and understandable. Entity-specific information is required when applying ESRS is insufficient for users to understand material IROs and their management.

Fair presentation applies to the sustainability statement as a whole, not as a mechanical assertion that each individual datapoint is “fair”. The reporting team should therefore perform two reviews:

• datapoint completeness review: have all applicable and material information items been addressed?

• statement-level fair-presentation review: does the combined narrative, metrics, limitations and cross-references convey a balanced and coherent picture?

This is why a report can fail even with a high spreadsheet completion percentage. It may contain boilerplate, hide significant limitations, use inconsistent boundaries or omit entity-specific context.

The revised ESRS 2 GDR-T requires disclosure of measurable, time-bound, outcome-oriented targets related to material IROs. Paragraph 51 contains nine distinct information elements, while paragraph 52 addresses the situation where no measurable outcome-oriented target has been set. ARs then clarify progress reporting, outcome orientation and scientific evidence.

Step 1 - capture the requirement in a controlled specification

Step 2 - issue a data request that preserves context

A weak request says: “Please provide the target and baseline.” A controlled request says:

The request specifies the material matter, boundary, period, methodology, governance and evidence. It also prevents the owner from supplying a corporate ambition that does not meet the target definition.

Step 3 - validate the response

The reviewer checks:

1. the target is approved and current;

2. the boundary matches the reported impact and metric;

3. the baseline can be reproduced;

4. units and absolute/relative basis are unambiguous;

5. methodology and assumptions are documented;

6. current performance is reconciled to the metric disclosure;

7. legal and scientific claims have specialist support;

8. limitations and changes are reflected in the draft.

Step 4 - draft the report disclosure

Why the wording is stronger

• it identifies the material impact and target relationship;

• the value, unit, baseline, boundary and period are clear;

• the method and estimate are transparent;

• the action link and current performance are visible;

• the limitation is specific and has an improvement path.

What must not be copied mechanically

The percentage, sites, baseline, exclusions, method and progress are illustrative. A real undertaking must use its own approved facts, apply the applicable ESRS edition and determine whether additional topical or entity-specific information is material.

Figure 2. From ESRS text to a controlled data request and disclosure. Branded educational visual by London Reporting Academy.

EFRAG IG 3 for the 2023 ESRS lists the universe of possible sector-agnostic datapoints and helps teams organise data collection. EFRAG describes it as non-authoritative implementation support. It is not a checklist for identifying material matters or IROs, and it does not replace ESRS 1, the DR objectives, ARs or the materiality assessment.

A safe workflow is:

1. import the IG 3 inventory into the data dictionary;

2. preserve source standard, DR, paragraph, AR and datapoint identifiers;

3. add applicability and materiality status fields;

4. assign owners only after the reporting boundary and topic assessment are understood;

5. keep a local change log for corrections and standard updates;

6. do not use the 2023 list as if it were an official datapoint list for the revised 2026 ESRS.

1. Starting with the datapoint list rather than the DR objective. Add the objective and paragraph context to every data request.

2. Treating each spreadsheet row as an independent requirement. Store parent DR, condition, AR and cross-reference fields.

3. Assuming “shall consider” means “shall disclose all examples”. Record evidence of consideration and report the material output required by the DR.

4. Marking unavailable information as not material. Separate materiality, availability, estimation, relief and omission statuses.

5. Ignoring narrative controls. Require evidence, owner approval, period/boundary checks and balanced drafting for narrative datapoints.

6. Using unofficial paraphrases as the legal source. Keep the official paragraph as the source anchor; use paraphrase only for operational instructions.

7. Mixing 2023 and 2026 identifiers. Maintain an edition field and a controlled transition mapping.

8. Calling a completed workbook “ESRS compliance”. Complete the statement-level fair-presentation, cross-reference and assurance review.

• ESRS 1: drafting conventions, fair presentation, double materiality, information materiality, entity-specific disclosures and presentation.

• ESRS 2: cross-cutting DRs and general disclosure requirements for policies, actions, metrics and targets.

• EFRAG IG 3: non-authoritative list of 2023 ESRS datapoints and explanatory note.

• EFRAG IG 1: non-authoritative materiality assessment guidance for the 2023 ESRS.

• ESRS digital taxonomy: a related but technically distinct digital-tagging layer; the human-readable datapoint list is not a substitute for taxonomy validation.

Is every “shall” sentence automatically disclosed?

It is prescribed when the DR is applicable, but the organisation must still apply the relevant materiality, condition, relief and presentation rules. A missing-data problem does not make the information immaterial.

Is an AR optional because it sits outside the main paragraph?

No. ARs support application of the requirements and are mandatory. The exact effect depends on the wording: “shall consider” requires the matter to be taken into account but does not necessarily require every example to be disclosed.

Is a datapoint always a metric?

No. Datapoints include narrative explanations, methods, boundaries, status statements, governance information, limitations and quantitative metrics.

Can we use IG 3 as our master data dictionary?

Yes, for the 2023 ESRS it is a useful starting inventory, provided the organisation adds applicability, materiality, owners, evidence, controls, report locations and version management. It should not be treated as authoritative or automatically reused for the revised 2026 ESRS.