Short answer
The answer, before the reasoning
GRI requires an organisation to identify actual and potential, negative and positive impacts and prioritise them according to significance. It does not prescribe a universal 1–5 scale, weighting formula or materiality matrix.
A defensible methodology preserves the GRI logic: actual negative impacts are assessed by severity; potential negative impacts by severity and likelihood; severity reflects scale, scope and irremediable character; and severity takes precedence over likelihood for potential human-rights impacts.
LRA Knowledge Hub visual — branded for editorial and learning use.
Source-grounded educational draft. Final LRA technical sign-off is required before publication.
PUBLIC ARTICLE
In practice
FORMAT
| FORMAT | LANGUAGE | VERSION |
|---|---|---|
| Knowledge Card | British English | 1.0 • 1 August 2026 |
Rule
WHO THIS IS FOR
Materiality workstream leads, consultants, reviewers and assurance professionals who need a practical scoring model without turning professional judgement into an unsupported mathematical rule.
What GRI prescribes — and what it leaves to the organisation
GRI 3 prescribes the logic for assessing significance, not a universal calculation template. The organisation should identify actual and potential, negative and positive impacts; assess significance using the relevant criteria; arrange impacts from most to least significant; group related impacts into topics; define and document a reporting threshold; test the resulting material topics; and obtain governance review and approval.
The standard does not prescribe a 1–5 scale, a weighted average, a two-axis stakeholder matrix, a single materiality score, a required number of material topics or a fixed cut-off. A numerical model can be useful as an implementation tool, but it must preserve the underlying GRI distinctions and remain open to professional judgement.
Figure 1. GRI defines distinct significance criteria. A scoring system is a controlled implementation choice, not a standard requirement.
Start with an impact universe, not a list of ESG topics
A scoring model is only as good as the impact statements it receives. “Climate”, “diversity” or “community” cannot be meaningfully rated until the organisation describes the effect, the affected resource or stakeholder, the activity or relationship involved, whether the impact is actual or potential, and whether it is negative or positive.
In practice
| Required impact-register field | Purpose | Illustrative entry |
|---|---|---|
| Impact ID and topic candidate | Provides stable traceability before impacts are grouped into topics. | WAT-017 / water and local communities |
| Impact statement | States the effect on the economy, environment or people. | Dry-season abstraction reduces household access to water in two settlements. |
| Actual / potential | Determines whether likelihood is relevant. | Actual |
| Negative / positive | Determines the significance criteria and prevents offsetting. | Negative |
| Activity / relationship | Identifies where involvement arises. | Owned processing site |
| Affected people/resource and geography | Supports scale, scope and irremediability assessment. | Approximately 1,400 residents and a stressed river basin. |
| Evidence and uncertainty | Makes the rating reviewable. | Metered withdrawals, basin study, community interviews; household data partly estimated. |
| Owner and reviewer | Separates preparation from challenge. | Water lead / human-rights and environmental reviewers. |
In practice
Apply the correct GRI significance logic
| Impact type | GRI significance basis | Practical treatment |
|---|---|---|
| Actual negative | Severity: scale, scope and irremediable character. | Do not apply likelihood — the impact has occurred. Assess the current or realised harm and any continuing effects. |
| Potential negative | Severity and likelihood. | Rate the seriousness of the possible harm and the chance of occurrence. For potential human-rights impacts, severity takes precedence over likelihood. |
| Actual positive | Scale and scope. | Assess how beneficial and how widespread the realised effect is. Do not deduct unrelated negative impacts. |
| Potential positive | Scale, scope and likelihood. | Assess the potential benefit, reach and chance of realisation; distinguish aspiration from evidence-backed potential. |
Rule
HUMAN-RIGHTS SAFEGUARD
For potential negative human-rights impacts, severity takes precedence over likelihood. A low-probability but catastrophic or irremediable human-rights impact must not disappear because a multiplication formula produces a modest score.
A sample 1–5 scale — implementation practice, not a GRI requirement
The following scale is an illustrative model that an organisation can calibrate to its sectors, geography and evidence. The anchors should be rewritten for the organisation’s impact universe. Ratings should be applied at impact level by people with relevant expertise and challenged in a calibration workshop.
These anchors deliberately mix qualitative and quantitative evidence. A rigid global number of people, tonnes or hectares can misstate context: the scale of water withdrawal depends on basin stress; the severity of discrimination depends on the right affected and the consequences; and a small number of people can experience extremely severe harm.
In practice
| Rating | Scale — gravity/benefit | Scope — reach — Irremediable character — negative impacts — Likelihood — potenti |
|---|---|---|
| 1 | Minimal effect; limited change from baseline. | Isolated individual, very small area or negligible resource. — Readily reversible in a short period with routine measures. — Remote under credible scenarios. |
| 2 | Minor effect with limited duration or consequence. | Small group, single localised area or limited resource. — Reversible with modest time and resources. — Unlikely but plausible. |
| 3 | Meaningful effect requiring management attention. | Multiple groups/sites or a material local area/resource. — Remedy is possible but requires sustained action and may leave residual harm. — Possible / could occur in the planning horizon. |
| 4 | Severe effect, rights infringement or major environmental/economic harm; or substantial positive benefit. | Large population, several regions/sites or extensive resource impact. — Difficult, lengthy or only partially effective remedy. — Likely under current controls or exposure. |
| 5 | Catastrophic, life-changing, systemic or fundamental-rights harm; or transformative positive benefit. | Very widespread, cross-border, ecosystem-level or systemic reach. — Irreversible or no effective remedy is realistically available. — Almost certain / recurring / already imminent. |
Do not mechanically average severity
GRI states that any of scale, scope or irremediable character can make a negative impact severe. A simple arithmetic average can therefore dilute one extreme dimension. For example, ratings of scale 5, scope 1 and irremediability 5 average to 3.7, even though the impact may be catastrophic and impossible to remedy for the affected person or group.
Hypothetical scenario
ILLUSTRATIVE CONSERVATIVE RULE
One defensible implementation option is to assign a provisional severity band from the highest dimension, then use a multidisciplinary panel to confirm or adjust the band with a written rationale. Another option is a decision table that treats any rating of 5 as critical and any rating of 4 as at least high. GRI does not mandate either method; the organisation must choose, calibrate and document its approach.
Illustrative only. It shows how the decision is made, not wording that can be copied or relied on.
In practice
| Example ratings | Mechanical average | Conservative assessment question |
|---|---|---|
| Scale 5 / Scope 1 / Irremediability 5 | 3.7 | Does the limited scope genuinely reduce significance when harm is catastrophic and irremediable? |
| Scale 2 / Scope 5 / Irremediability 2 | 3.0 | Does widespread reach create cumulative or systemic significance despite moderate individual gravity? |
| Scale 4 / Scope 3 / Irremediability 3 | 3.3 | What evidence supports a high or critical band, and are vulnerable groups affected? |
Combine severity and likelihood without losing the underlying judgement
For potential negative impacts, the organisation can use a matrix or decision rules to combine severity and likelihood. The output should remain a priority band, not a false precision score. The model should also include an explicit override for severe potential human-rights impacts and, where the organisation chooses, other highly severe low-likelihood impacts.
If the organisation uses multiplication, addition or weights, it should demonstrate that the formula cannot override these safeguards. It should also retain the component ratings and rationale so that reviewers can see why two impacts with the same total score may require different treatment.
In practice
| Severity band | Low likelihood | Medium likelihood — High likelihood — Human-rights override |
|---|---|---|
| Critical | High / critical review | Critical — Critical — Critical regardless of low likelihood where severity warrants priority. |
| High | Medium / high | High — Critical — At least high; document why likelihood does or does not change sequencing. |
| Medium | Low / medium | Medium — High — Review vulnerability, cumulative effects and evidence gaps. |
| Low | Low | Low / medium — Medium — Do not use low data confidence to assume low severity. |
Assess positive impacts separately
Actual positive impacts are assessed by scale and scope; potential positive impacts also use likelihood. Positive and negative impacts should be prioritised separately because they may not be comparable, and negative impacts cannot be offset by positive impacts. A renewable-energy project’s climate benefit does not neutralise an associated displacement or labour-rights impact.
In practice
| Positive-impact check | Reviewer prompt |
|---|---|
| Additionality | Is the benefit an actual effect attributable to the activity, or merely normal business output described positively? |
| Distribution | Who benefits, who does not, and are vulnerable groups reached? |
| Scale and scope | How beneficial is the effect and how widespread is it? |
| Likelihood | For potential benefits, what evidence supports the chance of realisation? |
| Associated harm | What negative impacts could arise from the same activity, and are they assessed separately? |
From impact ratings to a ranked topic list
1. Rate impacts individually and retain component ratings, evidence, uncertainty and reviewer comments.
2. Arrange negative impacts from most to least significant; perform a separate prioritisation for positive impacts.
3. Group related impacts into topics only after impact-level assessment, using stable impact IDs.
4. Set a topic floor so that a severe individual impact cannot be averaged away by many lower-rated impacts within the same topic. This is an implementation control, not a GRI formula.
5. Consider cumulative, systemic and geographically concentrated effects that may not be visible in isolated ratings.
6. Define a transparent threshold or cut-off after reviewing the ranked list and the organisation’s reporting purpose.
7. Record exceptions: topics included below the numerical threshold because of severe human-rights impacts, rapidly evolving evidence, regulatory context or expert challenge.
8. Test the proposed topics against applicable Sector Standards and with relevant experts and information users.
9. Obtain highest-governance-body review and approval, and retain the final decision log.
How to set the threshold
GRI leaves the cut-off to the organisation. The threshold should be set after impacts are ranked relative to one another, not copied from another company or fixed before the impact universe is understood. It should be high enough to focus reporting on the most significant impacts but not manipulated to fit a predetermined report length or available data.
In practice
| Threshold approach | How it works | Control needed |
|---|---|---|
| Priority-band threshold | All critical and high impacts/topics are material; medium items require documented review. | Clear band definitions and exception route. |
| Ranked cut-off | A cut-off is placed after reviewing the ordered list of topics. | Rationale for where the line is drawn and sensitivity testing around the cut-off. |
| Hybrid | Priority bands plus governance inclusion of certain impacts below the cut-off. | Transparent inclusion criteria; no exclusion because reporting is difficult or performance is poor. |
Rule
PROHIBITED SHORTCUT
Difficulty collecting data, absence of a management programme, poor performance or reputational discomfort are not significance criteria. A topic cannot be moved below the reporting threshold because it is hard to disclose.
Calibration workshop
A calibration workshop reduces inconsistent scoring between functions and locations. It is not a vote on topic importance. Participants should review a sample of impact statements before the full assessment, compare ratings, challenge evidence and agree how the scale will be applied.
1. Confirm definitions of impact, actual/potential, negative/positive, scale, scope, irremediability and likelihood.
2. Test the scale on several contrasting impacts, including a severe low-likelihood human-rights case.
3. Compare ratings between environmental, workforce, community and value-chain specialists.
4. Identify where quantitative thresholds help and where qualitative rights- or context-based judgement is necessary.
5. Agree conservative treatment for missing evidence; data gaps should trigger investigation, not automatically low scores.
6. Document disagreements, final calibration decisions and examples to be used in assessor guidance.
7. Recalibrate after the first assessment round if systematic drift or sector/geographic bias is detected.
In practice
Minimum decision log
| Field | What to record |
|---|---|
| Impact and topic IDs | Stable identifiers and version history. |
| Component ratings | Scale, scope, irremediability and likelihood where relevant. |
| Evidence references | Source documents, stakeholder input, expert analysis and data dates. |
| Rationale | Why the ratings and overall priority band were selected. |
| Uncertainty | Missing data, assumptions, confidence and sensitivity. |
| Human-rights logic | Whether severity precedence or a rights-based override was applied. |
| Calibration/challenge | Who prepared, challenged and approved the rating; changes made. |
| Threshold outcome | Included/excluded and the reason, including any exception. |
| Update trigger | Incident, acquisition, new relationship, geography, regulation or evidence that requires reassessment. |
Hypothetical scenario
ILLUSTRATIVE FIGURES
The ratings below demonstrate one calibrated implementation model. They are not GRI-prescribed scores and must not be copied without organisation-specific definitions and evidence.
Illustrative only. It shows how the decision is made, not wording that can be copied or relied on.
In practice
Worked illustrative scoring example
| Impact | Type | Component assessment — Priority conclusion |
|---|---|---|
| Serious injury from contractor work at height | Potential negative; human rights | Scale 5; scope 2; irremediability 5; likelihood 2. — Critical/high priority despite lower likelihood because potential loss of life and irremediability dominate. |
| Dry-season water abstraction reduces household access | Actual negative | Scale 4; scope 3; irremediability 3. — High priority; actual impact assessed through severity only. |
| Affordable renewable-energy product expands access | Actual positive | Scale 3; scope 4. — High positive-impact candidate, assessed separately from negative impacts. |
| Future apprenticeship programme | Potential positive | Scale 3; scope 2; likelihood 2. — Medium/conditional until implementation evidence increases likelihood. |
| Office paper use | Actual negative | Scale 1; scope 1; irremediability 1. — Low relative significance in this impact universe; retain internally and monitor. |
Common methodology errors
• Scoring topic names instead of impact statements.
• Using the same formula for actual and potential, negative and positive impacts.
• Applying likelihood to actual impacts or allowing low likelihood to suppress severe human-rights impacts.
• Averaging scale, scope and irremediability so that one extreme dimension disappears.
• Netting positive and negative impacts into one score.
• Using financial magnitude, stakeholder popularity or media attention as substitutes for impact significance.
• Treating missing data as evidence of low impact.
• Grouping impacts into topics before scoring and then assigning one generic topic score.
• Setting the threshold to produce a desired number of report chapters.
• Retaining no calibration record, decision log or governance approval.
Readiness
Methodology reviewer checklist
- • ☐ The impact universe covers activities and business relationships and includes actual/potential and negative/positive impacts.
- • ☐ Impact statements identify the effect, affected people/resource, location and relationship.
- • ☐ Actual negative impacts are assessed by severity; potential negative impacts by severity and likelihood.
- • ☐ Severity ratings separately consider scale, scope and irremediable character.
- • ☐ Potential human-rights impacts have a documented severity-precedence rule.
- • ☐ Positive impacts use scale/scope and likelihood only where potential.
- • ☐ Negative and positive impacts are prioritised separately and not offset.
- • ☐ Numeric scales are calibrated, evidence-based and explicitly described as implementation practice.
- • ☐ Aggregation cannot dilute a highly severe impact.
- • ☐ Threshold and exceptions are documented and tested against Sector Standards and expert/user input.
- • ☐ Governance review, approval and update triggers are retained.
- • ☐ Disclosure 3-1 accurately describes the method actually used.
Bottom line
A robust GRI scoring methodology uses numbers to organise judgement, not to replace it. It preserves the correct significance criteria, prevents severe impacts from being averaged away, separates positive and negative impacts, makes uncertainty visible and leaves a decision trail that another reviewer can reproduce. The quality of the method is demonstrated by its logic and evidence — not by the sophistication of its spreadsheet.
Official source anchors
The source set below should be rechecked as part of the pre-publication update control. Normative conclusions in this article are based on the current official GRI Universal Standards and official FAQs.
1. GRI 1: Foundation 2021. Requirements for reporting in accordance, reasons for omission, Content Index and Topic Standard selection. Open official source
2. GRI 2: General Disclosures 2021. Official requirements and guidance for the reporting organisation, governance, policies, workers and stakeholder engagement. Open official source
3. GRI 3: Material Topics 2021. Official guidance and disclosures for determining, listing and managing material topics. Open official source
4. GRI Universal Standards 2021 FAQs. Official clarifications on Topic Standard disclosure selection, reasons for omission, materiality and boundaries. Open official source
5. GRI Standards — English language. Official access point for current GRI Standards and publication versions. Open official source
Take it with you
The checklists as a working spreadsheet
Every checklist and table on this page, with empty status, owner and evidence columns for your team to fill in and keep.
✓ LRA AI Assistant · Human-in-the-loop
Ask about this guide
It answers from this page, and reaches into the linked disclosure cards when your question is about the standard itself. Your first two answers are free without signing in.
Go deeper · GRI
GRI Standards Certified Training
A full reporting cycle with a mentor: impact inventory, threshold, Topic Standard selection, Content Index and assurance readiness.
Available as Guided Flex, Live Cohort, 1:1 Expert Mentorship or Corporate Programme.
