Triple Bottom Line Index

Methodology

This page is generated from the same constants the scoring code uses, so it cannot drift from what actually produced the numbers. Methodology version 1.0, pipeline version 1.0.0.

The short version. Three pillars, equally weighted. People comes from the government's statutory gender pay gap register. Planet comes from the Science Based Targets initiative's validated-target dashboard. Profit comes from reported financials. A company is ranked only if all three pillars can be scored.

Why not use an existing ESG rating?

Because they disagree with each other. Correlation between the major ESG raters sits around 0.5; for credit ratings it is about 0.99. When two agencies looking at the same company reach materially different conclusions, at least one of them is not measuring what it claims to — and none of them will show you the workings.

This index does the opposite. Fewer inputs, all of them public and statutory, every one of them shown.

People

Source: the Gender Pay Gap Service. Every UK employer with 250 or more staff must file each year, and the service publishes the full dataset as a CSV.

Group handling. A listed plc is usually a holding company that employs almost nobody directly — its people work for UK operating subsidiaries which file separately. We match a constituent to its subsidiaries by Companies House number and by name stem, then aggregate, weighting each filing by the midpoint of its reported employer size band. Every match was reviewed by hand; unrelated employers that share a name stem — schools, care homes, similarly-named businesses — are excluded by an explicit list in the source code. The entities matched to each company are listed in its detail panel.

MetricWeightScores 100Scores 0
Median gender pay gap50%0 → 10030 → 0
Seniority gap35%0 → 10040 → 0
Median bonus gap15%0 → 10060 → 0

Median gender pay gap

The difference between the median hourly pay of men and women. 0% or better scores 100; 30% or worse scores 0.

Seniority gap

How much more of the workforce's lowest-paid quarter is female than its highest-paid quarter, in percentage points. Measures whether women are concentrated at the bottom. 0pp scores 100; 40pp scores 0.

Median bonus gap

The difference between the median bonus paid to men and to women. 0% scores 100; 60% or worse scores 0.

These are compared in absolute terms, not against sector peers. A 30% pay gap means the same thing in banking as it does in retail.

Planet

Source: the Science Based Targets initiative target dashboard. The SBTi independently validates corporate emissions targets against the Paris Agreement and publishes the register.

What this measures, and what it does not. It measures ambition that has survived independent validation, and the deadlines attached to that ambition. It does not measure tonnes of CO2 emitted, or whether emissions are actually falling. That is the largest known weakness in this index. Closing it means extracting reported Streamlined Energy and Carbon Reporting figures from annual reports, which is the next piece of work on the roadmap.

Commitment status

StatusPointsMeaning
Target Validated45Emissions targets independently validated by the SBTi.
Committed25Has committed to setting targets, but they are not yet validated.
Removed8Made a commitment, then failed to get a target validated in time and was removed from the SBTi register.
None0No commitment or validated target on the SBTi register.

Removed scores above none but well below committed. A company that made a public commitment and then failed to get a target validated in time has demonstrated intent and missed it — worse than a live commitment, better than never having engaged at all.

Target quality

Deadlines are scored, not just the existence of a target. Without that, a company targeting 2030 and one targeting 2035 score identically and the pillar collapses into a near-binary has-a-target flag.

Scope 3 — supply chain and product use — is the majority of a typical corporate footprint. A target that excludes it is a materially weaker commitment.

Where a validated target is aligned to less than 1.5°C, the whole pillar score is multiplied by 0.85.

Profit

Source: reported company fundamentals — revenue, margins, returns and headcount as stated in the latest annual accounts.

MetricWeightHow it is scored
Operating margin35%Ranked against index members in the same sector.
Return on equity25%25% or better scores 100; zero or negative scores 0.
Revenue growth20%+15% or better scores 100; −10% or worse scores 0.
Revenue per employee20%Ranked against index members in the same sector.

Operating margin and revenue per employee are ranked within sector, because absolute comparison across sectors is meaningless: a supermarket and a software company do not share a margin scale. Return on equity and revenue growth are scored absolutely. Figures reported in a foreign currency are converted to sterling at the build-date rate.

Combining the pillars

The overall score is the mean of the three pillar scores, equally weighted. There is no defensible objective basis for weighting people against planet against profit, so we do not pretend to one. The full data file is published; anyone who prefers different weights can recompute from it.

Missing data is never an advantage

A company is ranked only when all 3 pillars can be scored. This matters more than it looks. Pillar medians differ — Planet scores higher across the index than People does — so averaging over "whatever we happen to have" would hand a company that discloses less a better headline score than one that discloses more. Companies below the bar keep their individual pillar scores and are published separately, clearly marked.

Within a pillar, a missing component is handled the same way: the remaining weights are renormalised rather than the missing value being treated as zero, so a company is scored on what it discloses rather than penalised twice.

Known limitations

Corrections

If a number here is wrong, it is wrong in public and in the open. The pipeline, the matching rules and the exclusion lists are all in the repository. Open an issue and it will be fixed in the next build.