A carbon number on its own tells you how big a company's footprint is. It does not tell you whether that footprint is reasonable, improving, or better or worse than anyone else's. For that you need a denominator - and the moment you choose one, you have made a claim about what the carbon was for.
That choice is not a technicality. Divide the same emissions by revenue, by operating cost, by headcount or by capital and you get four different stories about the same company, each defensible, each answering a different question. Most organisations pick one, usually without deliberating, and then wonder why the comparison feels unfair.
In defence of intensity
Before the taxonomy, the objection - because carbon intensity targets have a reputation problem.
They are widely treated as the soft option: the target you set when you are not prepared to commit to an absolute reduction. We will get more efficient per unit, the argument runs, while continuing to grow, and the atmosphere will be no better off. As a critique of intensity targets used instead of absolute ones, that is entirely fair. The atmosphere responds to total tonnes, not to ratios.
But the reputation is doing too much work, and recent events should make us look again. When the SBTi changed its rules in April 2026 to let companies spread reductions later along the path to 2050 - roughly 21% by 2030 where the previous rules required around 42% - it did so because a great many companies were simply not going to reach the targets they had already committed to. Nobody set those targets expecting them to move. What the change exposed was not a flexible rule; it was a large number of absolute commitments made without a credible route to them. An absolute number set without a way of reaching it is not obviously more honest than a well-designed intensity commitment.
And a serious intensity target does something an absolute target does not: it forces a company to think about the relationship between growth and carbon. For a fast-growing business, holding emissions flat while doubling in size is a genuine achievement that an absolute target either understates or makes look like failure. Decoupling - growing the business without growing the footprint at the same rate - is a real strategic discipline, and designing for it usually requires changing the operating model rather than trimming it. An intensity target can be a sharper recognition of business context, and a more considered piece of strategy, than a headline absolute number nobody has costed.
The right answer is not one or the other. Absolute targets set the destination and keep the company honest about the atmosphere. Intensity ratios describe the machine, and tell you whether the underlying operating model is actually getting better. You need both, and the ratios are the half we routinely neglect.
It is worth remembering that the opposite of a good idea can also be a good idea. Absolute and intensity targets are exactly that kind of pair: both are right, for different reasons, and each fails in a way the other does not. The mistake is rarely choosing the wrong one. It is becoming so wedded to one that you stop hearing what the other was trying to tell you.
Why ratios are the better comparison instrument
There is a second reason to care, which is comparison.
Try to compare a diversified conglomerate with a specialist manufacturer on absolute emissions and the exercise collapses immediately: different sizes, different structures, different numbers of businesses inside the wrapper. Comparing rates of decline is not much better, because a company that has already done the cheap work will show a slower rate than one that has not started, and look worse for being ahead.
Intensity ratios cut through that. They describe how carbon-intensive the underlying operating model is - the machine rather than its scale - and they do it in two dimensions at once: the static position, which is how efficient the model is today, and the growth position, which is whether the model gets better or worse as it expands. That is the comparison a strategist and an investor both actually want.
The six
Revenue intensity - tCO2e per £m of revenue. The default, the most widely disclosed, and the weakest of the six as a comparator. It answers: how much carbon does each pound of sales carry? It has two genuine merits. It is near-universal, so it is almost always available. And it flows up through the value chain: one company's revenue is another company's spend, which makes revenue intensity the basis on which spend-based emissions factors are built, and therefore the workhorse of Scope 3 estimation. Its flaw is that revenue is a price, not a physical quantity, so the denominator moves with pricing, mix, currency and inflation - none of which have anything to do with carbon.
Operating intensity - tCO2e per £m of operating cost. The best general comparator, and the argument for it is below. It answers: how much carbon does the actual running of this business produce, per pound spent running it?
Asset intensity - tCO2e per £m of property, plant and equipment plus leased real estate, on Scope 1+2+3 for consistency with the other headline measures. It answers: how carbon-intensive is the physical infrastructure this company owns and occupies? The most revealing ratio for asset-heavy firms, because it surfaces stranded-asset risk - the plant, fleet or estate that a transition may render uneconomic long before it is written down.
Workforce intensity - tCO2e per FTE. It answers: how much carbon does each person in this business carry? Genuinely useful for people-based businesses - professional services, technology, financial services - where headcount, not physical output, is the real unit of activity. Weak in capital-intensive sectors, where automation can flatter it dramatically.
Economic intensity - tCO2e per £m of enterprise value including cash (EVIC). The investor's ratio, and the one embedded in EU regulation: under SFDR, financed emissions are attributed by dividing a company's emissions by its EVIC. It answers: how much carbon does each million of capital deployed in this business produce? Indispensable for portfolio work, but it moves with the market - a company can halve its economic intensity through a share price rally, having changed nothing at all.
Profit intensity - tCO2e per £m of EBITDA. The stress test. It answers: how much carbon does this business emit for each pound of operating profit it generates - and therefore how exposed are those earnings if that carbon is ever priced? This is the ratio that leads directly to carbon-adjusted EBITDA, and the one an investor should reach for when thinking about transition risk in a valuation.
Why I prefer cost over revenue
The single most consequential choice in that list is between the first two, and I come down firmly on the side of cost. It is worth being precise about why.
Revenue is operating cost plus margin. Carbon is generated by operations - by making, moving, heating, cooling, computing, delivering. It is not generated by margin. Margin is priced, not produced; it is the outcome of market position, brand and pricing power, none of which emit anything. So dividing carbon by revenue mixes a physical quantity with a pricing quantity, and the pricing quantity varies enormously between businesses that are, physically, doing exactly the same thing.
Set a high-margin and a low-margin business side by side. They might run identical operations, identical fleets, identical facilities, and produce identical emissions. On revenue intensity the high-margin business looks materially cleaner - not because it is, but because it charges more. The same distortion runs through product portfolios inside a single company: premium lines look green, commodity lines look dirty, and the difference is pricing.
Divide by operating cost instead and the pricing layer is stripped out. What remains is a comparison of the underlying operating model and the infrastructure behind it - how much carbon this machine produces per pound of running it. That is the like-for-like a conglomerate, a boutique and a fast-growing challenger can all be measured on.
There is a sharper version of the same point. On revenue intensity, a company can improve its carbon performance by putting its prices up. Nothing physical changes; the ratio simply gets better. Any metric that can be improved by an invoice is not measuring what we think it is measuring.
Which makes one detail in the regulation worth noticing. Under SFDR, the indicator actually named "GHG intensity of investee companies" is defined as emissions per million euros of revenue, while the EVIC-based measure is filed under "carbon footprint". The default is not just a habit; it is written into the rules. Being aware of that is part of reading the number rather than accepting it.
Read them as a set
None of the six is the answer, and choosing a favourite is the wrong instinct. Each denominator answers its own question honestly, and each has a condition under which it flatters or misleads. Revenue drifts with pricing. Economic intensity drifts with the market. Workforce intensity drifts with automation. Asset intensity is powerful for heavy industry and close to meaningless for a software firm.
So you read them together, the way you read a set of accounts rather than pulling a single line from one. Absolute tonnes for the atmosphere, and the gross position for what the company actually did. Operating intensity for the honest comparison of the machine. Asset intensity where the balance sheet is heavy and stranding is the risk. Profit intensity where the question is what happens to earnings when carbon is priced. Read one alone and you will be misled; read the set and the shape of the business appears.
That is the whole discipline, applied one level down from the statements: not a number, extracted and ranked, but a set of related measures read in context. Holding all six against a single, provenance-graded footprint - so the same tonnes can be looked at through every denominator that matters, and the differences between them can be seen and explained - is exactly the kind of thing Reverberate exists to make routine.
Sources
Final Report on SFDR Delegated Regulation amending RTS — ESMA / European Supervisory Authorities. The SFDR principal adverse impact indicators, including the definition of GHG intensity of investee companies as emissions per million euros of revenue, and the EVIC basis used for financed emissions.
Carbon Footprinting Demystified — MSCI ESG Research. How financed-emissions metrics are constructed across PCAF, TCFD, SFDR and others, including EVIC as the attribution factor and why absolute financed emissions are poorly suited to comparison.
SBTi updates rules to allow less ambitious near-term targets — Corporate Knights. The April 2026 change letting companies spread reductions across the full pathway to 2050 rather than concentrating them before 2030, with the net-zero endpoint unchanged.