play icon for videos

CSR Metrics: Indicators, KPIs, Scoring & Scorecards

Reference library for CSR metrics — the indicator families, scoring formulas, and rating methodologies (MSCI, Sustainalytics, EcoVadis) behind every scorecard.

Updated
July 17, 2026
360 feedback training evaluation
Use Case

What are CSR metrics?

CSR metrics are the defined indicators that measure a company's social and environmental performance: representation and pay equity, community investment and people reached, emissions and waste, supplier standards, and the stakeholder sentiment behind them. Each metric is a formula with a unit, allowed values, and a source — the building block every CSR scorecard and rating is assembled from. A metric is only as good as its definition.

The gap most CSR programs hit is not a missing indicator; it is that the same indicator is defined three different ways across departments, so the scorecard cannot be compared or defended. And under the ESRS standards the hardest part is upstream of the number entirely: collecting stakeholder input aligned to the standard is, in practitioners' words, the weak link.

Key takeaways

  • A CSR metric is a defined indicator — a formula, a unit, allowed values, and a source — not a number someone typed into a slide. Undefined indicators break comparability the moment two departments roll up.
  • The weak link is the stakeholder data, aligned to the standard. Under ESRS, a metric is only defensible if the input behind it was collected to the standard and can be traced to who gave it.
  • Sopact calls the reference layer the Metric Dictionary: every CSR indicator defined once — formula, unit, allowed values, and the stakeholder source it draws from — so a scorecard number means the same thing across the company and across years.
  • Scores and ratings are composites of defined metrics. MSCI, Sustainalytics, and EcoVadis weight indicators into a rating; the weighting is only as credible as the definitions underneath it.
  • Output metrics count activity; outcome metrics measure change. The distinction, and the definitions on both sides, live here; the measurement discipline over time lives on the CSR performance page.

A metric is a definition, not a number.

The reason a CSR scorecard is hard to defend is that its indicators were never written down once. One team counts community reach as people served, another as touchpoints, a third as dollars granted; each is reasonable, and together they make the rolled-up number meaningless. A definition — reach means unique individuals who completed a program, counted once — is what makes two business units finally comparable.

Sopact calls that reference layer the Metric Dictionary: every indicator defined once, with its formula, unit, allowed values, and the stakeholder source it draws from, on one record. Change a definition and past periods re-score to match instead of leaving two incompatible histories. The disclosure that consumes these metrics is on the CSR reporting page, and the measurement discipline over time on the CSR performance page.

The definition also fixes the ESRS weak link. When the stakeholder question that feeds an indicator is written from the dictionary and collected clean at the source, the metric arrives already aligned to the standard, rather than being reconstructed from a survey nobody designed against the requirement.

The eight CSR indicator families.

Most CSR metrics fall into eight families: workforce and DEI, community investment, environmental, supply chain, health and safety, governance and ethics, customer and product, and stakeholder sentiment. Naming the family helps with two decisions: which instrument collects it, and which rating methodology consumes it.

The table below pairs each family with an example indicator and the source it draws from. Ratings agencies such as IRIS+-aligned funds, MSCI, and Sustainalytics build composite scores by weighting indicators across these families; the composite is only as defensible as the definitions and sources beneath it.

Eight indicator families, one metric dictionary.

CSR indicators sort into eight families, and each draws from a specific source — an HR system, a meter, a supplier audit, or a stakeholder survey. Read the last column: the families that draw from stakeholder surveys are the ESRS weak link, and the ones that most need a defined dictionary.

The eight CSR indicator families
Indicator familyExample indicatorDraws from
Workforce & DEIRepresentation, pay equity, retentionHR system + employee survey
Community investmentGrants, volunteer hours, people reachedGrant records + beneficiary survey
EnvironmentalEmissions, energy, waste, waterMeters + facility data
Supply chainSupplier audits, labor standardsSupplier survey + audit
Health & safetyIncident rate, training completionOperations and HR records
Governance & ethicsBoard diversity, code-of-conduct trainingGovernance records
Customer & productAccess, satisfaction, product safetyCustomer survey
Stakeholder sentimentTrust, material-topic salienceStakeholder survey (the ESRS weak link)

The families that draw from meters and records are the easy half; the families that draw from stakeholder surveys are where CSR metrics usually break, because the input was never collected to a standard. Defining each indicator once in the Metric Dictionary, and collecting its source clean, is what closes that gap.

A scorecard is refreshed each quarter. The Loop keeps the definitions holding.

A metric is only reproducible if the analysis behind it holds still: the same scoring guide applied the same way to the same stakeholder data returns the same figure. If the number moves when someone re-checks it, the scorecard is a judgment call, not a metric. That reliability is part of the Loop, Sopact's method for continuous impact intelligence: collect clean at the source, analyze the moment data arrives, improve while you can still act.

The Loop is what keeps a CSR metric defensible over time. Every indicator traces back to the stakeholder responses beneath it, so when a rating agency or an auditor questions a score, it resolves to its source. That standard has its own chapter in reliability and reproducibility.

One method, three moves that never stop

1 · CollectClean at the source; each indicator's stakeholder input, aligned to the standard.
2 · AnalyzeOn arrival; the same scoring guide, so a metric is reproducible.
3 · ImproveIn time to act; fix a drifting definition before it reaches the scorecard.

Then the cycle runs again, a little sharper each quarter. Read the method: the Loop methodology →

Build the metric dictionary this week

The fastest way to make a scorecard defensible is to define one indicator and collect its source clean. Each prompt below pastes into Sopact Sense's Assistant, or reasons through with your team; the arrow above each links the Academy walkthrough that shows the expected output and the tips.

Academy walkthrough → Build the metric dictionary

Turn the CSR indicators below into a metric dictionary: [PASTE INDICATORS]. For each, give the field name, a one-sentence definition, the formula, the unit, the allowed values, and the source it draws from. Flag any indicator two teams would compute differently and propose the stricter reading. Return a table: Indicator / Definition / Formula / Unit / Allowed values / Source.

Academy walkthrough → Collect the stakeholder input to standard

From this metric dictionary and the ESRS/GRI topics I must cover [PASTE], draft the stakeholder survey that collects the input clean at the source: the questions, the required fields, the allowed categories, and the identifier that links each response to a stakeholder. Flag any indicator whose source I do not yet collect.

Academy walkthrough → Cut to the few metrics that matter

From the indicator families below, propose the smallest set of CSR metrics that covers my material topics: [PASTE FAMILIES / MATERIAL TOPICS]. Write each as who or what is measured / what changes / by when, and mark OUTPUT or OUTCOME. Drop indicators that duplicate. Return a table: Metric / Definition / Output-Outcome / Material topic.

Academy walkthrough → Trace a scorecard cell to its source

For each cell in this CSR scorecard, build a source row: the score, the indicator and its definition, the stakeholder responses behind it, the weighting applied, and the one-line calculation. If a source is missing, write MISSING SOURCE. Return a table: Cell / Indicator / Source / Weighting / Calculation. Scorecard: [PASTE]

Learn the how-to in the Academy

Each walkthrough is short and practical: what to do, the prompt to run, the output to expect, and the tips that keep it reliable.

Watch: what's actually changing in impact and CSR measurement, and why defined metrics beat documented ones.

Frequently asked questions

What are CSR metrics?

CSR metrics are the defined indicators that measure a company's social and environmental performance — workforce and DEI, community investment, environmental, supply chain, health and safety, governance, customer, and stakeholder sentiment. Each is a formula with a unit and a source. In Sopact's framing they live in a Metric Dictionary, defined once so a scorecard number means the same thing everywhere.

What is an example of a CSR metric?

Examples by family: workforce representation and pay equity, community people-reached and volunteer hours, greenhouse-gas emissions and waste diverted, supplier audit pass rate, safety incident rate, board diversity, customer access, and stakeholder trust. Sopact defines each in the Metric Dictionary with its formula, unit, allowed values, and the stakeholder or system source it draws from.

What is the difference between a CSR metric and a KPI?

A CSR metric is any defined indicator of social or environmental performance; a KPI is the small subset a company holds itself accountable to. The discipline is choosing a few KPIs and defining them precisely. Sopact treats every KPI as a dictionary entry so it stays comparable across business units and years rather than drifting into near-synonyms.

How are CSR scores and ratings calculated?

A CSR score is a composite: indicators are normalized and weighted into an overall rating, the approach MSCI, Sustainalytics, and EcoVadis each use with their own weighting. The score is only as defensible as the definitions and sources underneath it. Sopact holds those in the Metric Dictionary, so a composite score can be traced back to the indicators and stakeholder responses that produced it.

What is the difference between output and outcome metrics in CSR?

An output metric counts what the company did — dollars granted, hours volunteered, people trained. An outcome metric measures what changed for the people or environment reached — income gained, emissions avoided, retention improved. Sopact keeps both defined in the Metric Dictionary; the measurement discipline that produces outcome metrics over time is on the CSR performance page.

What CSR metrics does ESRS or CSRD require?

ESRS spans twelve standards — environmental, social, and governance — and requires metrics on the topics a company's double-materiality assessment identifies as material, each backed by stakeholder input collected to the standard. Sopact aligns each indicator's definition and its stakeholder source to ESRS, so the metric arrives disclosure-ready rather than reconstructed at filing.

Why is collecting CSR metric data so hard?

The system metrics — emissions, headcount — are the easy half; the metrics that draw from stakeholder surveys are the weak link, because the input is often collected without a standard or an identifier, so it cannot be trusted or traced. Sopact fixes this by designing the stakeholder survey from the Metric Dictionary and collecting it clean at the source.

What is the difference between CSR metrics and CSR reporting?

CSR metrics are the indicators, formulas, and scoring; CSR reporting is the disclosure artifact that consumes them, covered on the CSR reporting page. Sopact keeps the metric definitions and the disclosure on one record, so the report's numbers and the metric library are always the same numbers.

How do I build a CSR scorecard?

Define each indicator once, collect its source clean, choose the weighting, and compute the composite — then keep every cell traceable to its inputs. Sopact builds the scorecard from the Metric Dictionary, so a score can be defended to a rating agency by resolving it to the indicators and stakeholder responses beneath it, not by re-deriving it from spreadsheets.

Next: put the metrics into a disclosure on the CSR reporting page, or measure them over time on the CSR performance page.