Overview
Every Standard ESG certificate carries a single public number from 1 to 10. That number is the end product of a defined, multi-step calculation — not a subjective impression — and this guide walks through exactly how it's built, from a single questionnaire answer up to the number printed on the certificate, with worked examples along the way.
The Big Picture
The score is built bottom-up, through five layers:
Indicator (0–100) → Criterion (weighted mean of its indicators) → Subject (weighted mean of its criteria) → Pillar (weighted mean of its subjects) → Composite (0–100, weighted mean of the three pillars) → Public score (1–10).
Each layer is a weighted mean of the layer below it, so a single weak indicator doesn't collapse an otherwise strong subject any more than a single strong indicator can paper over a genuinely weak one — the arithmetic mirrors how a careful human reviewer would actually weigh evidence. Sitting outside this arithmetic entirely are the gates, described later in this guide, which can deny certification regardless of what the math produces.
Step 1: The Indicator Score
Every questionnaire item — an indicator — is answered and scored 0–100 according to a defined scoring rule attached to that indicator. The rule depends on the answer type:
- A boolean indicator ("Does the company have a written anti-corruption policy?") might score 100 for yes, 0 for no.
- A percentage indicator ("What share of electricity comes from renewable sources?") might map directly, or against a curve appropriate to the industry.
- A numeric indicator (accident rate, energy intensity) is typically scored against a defined threshold or benchmark for the industry.
- A file-reference indicator (e.g., "Upload your GHG inventory") scores based on whether the required evidence is present and — at Level 2+ — verified (see below).
Each indicator also carries a weight relative to the other indicators in its criterion, and an applicability tag. An indicator that doesn't apply to a given company or industry — say, a biodiversity indicator for a pure software company — is excluded from the calculation entirely. It is not scored as zero, and it does not drag down the average; it simply isn't counted.
Step 2: Criterion, Subject, and Pillar Scores
From here, the aggregation is mechanical and repeats the same pattern three times:
- Criterion score = the weighted mean of its applicable indicator scores.
- Subject score = the weighted mean of its criteria scores. (Subjects are the fifteen core subjects — E1 through E5, S1 through S5, G1 through G5 — described in the Protocol Overview.)
- Pillar score = the weighted mean of its subjects' scores, giving a 0–100 score for each of Environmental, Social, and Governance individually.
At this point you already have something useful even before a composite is calculated: three separate pillar scores, which is why a company can be strong on Governance and weaker on Environmental, and both facts remain visible rather than being blended away immediately.
Step 3: The Composite Score and Pillar Weights
The composite score (0–100) is the weighted mean of the three pillar scores, using default weights of:
- Environmental: 40%
- Social: 35%
- Governance: 25%
These defaults can be overridden per industry template where a sector's risk profile warrants a different balance — a resource-intensive industry might weight Environmental even higher; a services industry might shift weight toward Governance. Whatever the weights, they are fixed as part of the industry template and stamped, with the rest of the template, by the content hash described in the Protocol Overview — so the weighting a company was scored against is always traceable and never retroactively adjustable.
Step 4: Mapping to the Public 1-10 Score
The composite score maps linearly onto the public score printed on the certificate: score10 = max(1, round(composite / 10)).
In words: divide the composite by ten, round to the nearest whole number, and never go below 1 — even a very low composite still receives a score of 1 rather than 0, because the scale itself is 1–10, and separately, seriously disqualifying conduct is handled by the gates rather than by driving the score to zero (see the gates section below).
The same formula, and the same 1–10 scale, is used at every certification level. A score of 7 means the same thing about performance whether it comes from a Level 1 self-assessment or a Level 3 on-site assessment — what differs between them is how much independent checking sits behind the number, which is exactly what the level on the certificate communicates (see The Three Certification Levels Explained).
A Complete Worked Example
Take a simplified, illustrative company — real templates have far more indicators than this example uses, but the arithmetic is identical.
Environmental pillar — suppose its five subjects score: E1 = 80, E2 = 70, E3 = 60, E4 = 75, E5 (not applicable, excluded). With equal weighting among the four applicable subjects: (80 + 70 + 60 + 75) / 4 = 71.25.
Social pillar — S1 = 90, S2 = 85, S3 = 70, S4 = 65, S5 = 80. Equal weighting: (90 + 85 + 70 + 65 + 80) / 5 = 78.
Governance pillar — G1 = 75, G2 = 60, G3 = 85, G4 = 70, G5 = 80. Equal weighting: (75 + 60 + 85 + 70 + 80) / 5 = 74.
Composite, at default weights (E 40% / S 35% / G 25%): composite = (71.25 × 0.40) + (78 × 0.35) + (74 × 0.25) = 28.5 + 27.3 + 18.5 = 74.3.
Public score: score10 = max(1, round(74.3 / 10)) = max(1, round(7.43)) = 7.
This company's certificate would show a score of 7 out of 10 — provided none of the gates in the next section are triggered.
The Gates: When the Score Doesn't Matter
However the arithmetic above comes out, certification is denied outright, at any level, if any of the following are true:
- Credible evidence or admission of child labour or forced labour.
- An active, material legal sanction for environmental crime or corruption that was undisclosed at the time of assessment.
- Any single pillar scoring below 20, even if the composite is otherwise acceptable.
- Failure to complete all applicable mandatory indicators.
These gates exist because averaging can conceal a serious failure: a company could theoretically post a composite of 74 in the example above while having, say, a Social score dragged down by a genuine forced-labour finding buried in one subject. The gates make sure that scenario denies certification rather than being smoothed over by strong Environmental and Governance numbers. No pillar can buy back a disqualifying fact in another.
Verification Adjustments by Level
The scoring pipeline above describes the base calculation from declared answers. Two further adjustments apply depending on certification level:
- At Level 2, any indicator whose required supporting document is missing, expired, or rejected during document verification is discounted to 50% of its declared value, or flagged for clarification rather than accepted. A company that declared a strong answer but can't substantiate it does not keep the full-strength score for that indicator.
- At Level 3, where an on-site finding contradicts a declared answer — for example, the declared accident rate doesn't match the incident log an auditor reviews on site — the on-site finding overrides the declared value, and the indicator is rescored using the corrected figure. The discrepancy itself is also logged as part of the assessment record.
This is why the same underlying performance can, in principle, produce a lower published score at a higher level: not because the company got worse, but because inflated or unsubstantiated answers get corrected as verification deepens. A rigorous Level 3 score of 7 is a stronger claim than an unverified Level 1 score of 7 precisely because it has survived this correction process.
Why This Design, and What It Protects Against
Three deliberate choices are worth naming explicitly, because each answers a specific failure mode seen elsewhere in the ESG ratings landscape:
- A published, fixed formula — rather than an opaque proprietary model — means anyone can, in principle, reproduce how a given composite was reached, addressing the divergence problem that plagues ESG ratings more broadly (different providers' scores for the same company have been found to correlate as weakly as 0.38).
- Exclusion rather than zero-scoring for non-applicable indicators avoids penalizing companies for questions that were never relevant to their industry in the first place.
- Gates that sit outside the weighted average ensure that no combination of strong pillars can mathematically offset a genuinely disqualifying fact — the score can never be used to launder a serious violation.
Frequently Asked Questions
Standard ESG (standardesg.org) publishes the same scoring methodology described here for every industry template. To see how the level attached to your score changes what it proves, read The Three Certification Levels Explained; to check a specific certificate, see Verify a Certificate.
Was this page useful?