The S in ESG
is the one
nobody can
prove

Ask a sustainability team for their carbon number and you get a figure with a boundary, a method and a standard behind it. Ask for their social number and you get a paragraph.

This is not because environmental teams work harder. It is because the environmental pillar solved a problem the social pillar has not: it agreed on a unit.

Why E works and S does not

Carbon has the Greenhouse Gas Protocol. Emissions reduce to tonnes of CO₂ equivalent, scopes are defined, boundaries are specified, and two companies in different industries on different continents can be compared on the same axis. It took decades to build and it is imperfect, but it exists.

Social has no equivalent. There is no widely adopted standard for measuring social impact in the way the GHG Protocol standardises emissions, and social factors are inherently context-specific and frequently qualitative — labour conditions, community impact, access, dignity.12 A training programme in one country and a water intervention in another do not reduce to a shared number, and pretending otherwise produces false comparability rather than real insight.

The social pillar is the easiest to report on vaguely and the hardest to report on well.

What gets reported instead

In the absence of a unit, organisations fall back on proxies: training hours delivered, audit completion rates, policy coverage, people reached, sessions run. These are easy to track, easy to aggregate, and they reflect activity rather than whether social conditions actually improved.2

That distinction is the entire argument of this paper, so it is worth stating without hedging: a training hour is an input. A policy is an intention. An audit completion is a process. None of them is an outcome, and reporting them in an outcome-shaped slot is how the social pillar acquired its reputation.

Figure 1

Why the three pillars are not equally provable

Pillar	Common unit	Typical reported metric	Is it an outcome?
Environmental	tCO₂e, m³, tonnes	Emissions by scope, water withdrawn, waste diverted	Largely yes
Governance	Structural facts	Board independence, audit committee, policy existence	Verifiable, if not an outcome
Social	None	Training hours, people reached, sessions delivered	Usually not

Why this is about to become expensive

For as long as sustainability reporting was narrative, a weak social section carried no penalty. That is changing, and the Gulf is a good place to watch it change.

In Saudi Arabia, 94 companies listed on the Saudi Exchange published sustainability reports in 2024, and roughly 65% of the top 100 by revenue disclose ESG information — largely voluntarily.4 On the environmental side there is a national number to report against: the Saudi Green Initiative commits the Kingdom to reducing emissions by 278 million tonnes annually by 2030 and to net zero by 2060.5

Note what that produces. A Saudi corporate reporting today can point at a quantified national environmental target and show its contribution in the same unit. On the social side it has a genuine mandate — Vision 2030 is unusually explicit about quality of life and human capability — and no comparable unit in which to evidence delivery against it.

As disclosure moves towards assurance, that gap stops being a presentational problem. An assured figure needs a stated method, a defined boundary and traceable source records. "We delivered 14,000 training hours" satisfies none of those requirements as an outcome claim, however accurately the hours were counted.

The fix is not a new global standard

It would be convenient to conclude that the sector needs a GHG Protocol for social impact. We do not think one is coming, and we are not sure it should. Social outcomes are context-specific in a way emissions are not; a single global unit would either be so abstract as to be meaningless, or would flatten differences that matter.

The practical fix is narrower and available now. It is not to make social outcomes globally comparable — it is to make each one individually defensible.

A social result carrying those four properties can be assured. It will not be comparable to another company's social result — but comparability was never the thing that made a number credible. Traceability was.

What a company should ask its delivery partners

Most corporate social programmes are executed by implementing partners, and the quality of the eventual disclosure is decided by those partners' measurement practice rather than by the reporting team. Four questions, asked at contracting rather than at reporting, settle it:

• What is the baseline, and when will it be taken? If the answer is after delivery starts, there will be no defensible outcome claim, and no amount of reporting effort will create one later.
• Which named target does each indicator map to? An SDG target number or a national framework pillar. If the indicators are bespoke, the results will need translating and something will be lost in the translation.
• How will attrition be recorded? Not whether — how. Participants will be lost, and a partner who has not thought about it will quietly report on the survivors.
• Who signs off the result? If it is the same team that delivered the programme, the number is a self-assessment. That is not fatal, but it must be disclosed as one.

None of this requires a new standard, a new framework or a new acronym. It requires deciding, before the money is spent, that the social pillar is going to be evidenced to the same standard as the environmental one — and then contracting for it.