Your ESG report is a scorecard. Your CFO needs a capital plan.

Every sustainability platform on the market can tell you what you emitted last year. Very few can tell you which cheque to write first.

That gap is not a reporting problem. It is the reason most corporate climate programmes stall after the first inventory.

What actually happens after the first report

A team spends four months building a greenhouse gas inventory. They chase utility bills, argue about boundaries, learn what a Scope 3 category is. At the end they produce a number — say 3,870 tonnes of CO₂ equivalent — and a report that satisfies the regulator.

Then someone in the room asks the only question that matters: so what do we do about it?

And the platform has nothing to say. It has measured the problem beautifully and offered no purchase on it. The sustainability lead goes away to build a business case in a spreadsheet, alone, using cost assumptions they are not qualified to make, for a capital committee that will ask questions the spreadsheet cannot answer.

Finance does not speak in tonnes

A capital committee does not evaluate proposals by how much carbon they remove. It evaluates them by return, payback and risk. Present a list of emission reduction measures ranked by tonnage and you have handed them a document in a language they do not use.

Rank the same measures by cost per tonne abated and something changes. Now every measure is an investment with a price, and the list sorts itself into three groups a CFO recognises immediately:

• Measures that pay for themselves. The energy they save exceeds what they cost over their life. These do not need a business case — they already have one. They need a slot in the schedule.
• Measures that cost money. Real abatement, real spend. Now it is a budget conversation, held on the same terms as every other budget conversation.
• Measures nobody has costed. Which is not a category to put on a chart. More on this below.

That ranking is called a marginal abatement cost curve, and it has been the standard tool in energy and climate economics for close to twenty years. It is not new. What is new is having one drawn from your own inventory rather than from a consultant's benchmark deck.

What the curve actually shows

Each measure is a block. Its width is the tonnes it removes each year. Its height is what removing them costs per tonne. Cheapest on the left, most expensive on the right.

The widths are the part a table of costs cannot give you. A measure that is very cheap per tonne but removes almost nothing is a thin sliver on the far left — technically attractive, strategically irrelevant. A measure that removes a quarter of your footprint at a modest cost is a wide block in the middle, and it is where the programme lives. You see that in a second on a chart and never from a list.

Below the zero line sit the measures that pay for themselves. In our demonstration inventory, three of them account for 353 tonnes — before anyone has argued for a budget.

The arithmetic has to be honest

This is where most carbon abatement tooling quietly falls apart, so it is worth being specific.

The cost of a measure is not its capital cost divided by its annual saving. A chiller optimisation that lasts fifteen years and an LED retrofit that lasts ten are not comparable investments at the same price, and treating them as such makes every capital-heavy measure look worse than it is.

The correct figure is a levelised cost, which spreads the capital across the asset's life at the organisation's own discount rate.

The capital recovery factor is what makes a twenty-five year rooftop solar installation comparable to a two-year compressed-air survey. The discount rate is the organisation's own hurdle rate, not ours, and it is shown on the chart — because changing it reorders the curve, and a reader is entitled to know which one produced the picture they are looking at.

We compute this in exact decimal arithmetic, not floating point. Financial figures that get added together and presented next to an inventory follow the same rules as the inventory.

What we refuse to put on the curve

A cost curve looks quantitative. That is the entire point of it — and it is also the reason a dishonest one is worse than none at all.

Measures with no costs entered stay off the chart. Not defaulted to zero. A measure recorded at zero cost would be the widest, cheapest thing on the curve and would sort itself to the far left, recommending itself to a capital committee on the strength of nobody having priced it. Ours are listed beside the chart under "not costed", with a count, so their absence is visible rather than convenient.

Contractual measures are counted separately. A green tariff or a renewable energy certificate changes your market-based Scope 2 figure. It does not change what your sites physically emit. Adding a green tariff to a lighting retrofit and presenting one combined total is the most common overstatement in corporate climate reporting, and our chart will not do it — the contractual saving appears in its own line with an explanation of why it is not in the total.

Measures that shift emissions between scopes are flagged. Electrifying a vehicle fleet moves emissions from Scope 1 to Scope 2. On a coal-heavy grid the net reduction is small. That belongs on the label, not in the footnotes.

Indicative costs are drawn hatched. Where a figure came from a rule of thumb — derived from the organisation's own energy price and a published payback range — rather than from a supplier quote, the bar is visibly different. A capital committee needs to know which bars came from a vendor and which came from an estimate.

The measures come from your data, not from a template

The suggestions themselves are generated against your actual emission lines. Not "consider energy efficiency" — which is what an unconstrained model produces and what nobody can cost, schedule or verify. Instead: this site's purchased electricity is the largest single line in your inventory, chiller optimisation typically removes 5–15% of it in a tropical climate, and here is the metering you would need to confirm that before you rely on the number.

The assistant chooses the measure and proposes a percentage within a published planning range. It does not calculate the tonnage. The system does that, from your approved inventory, in the same deterministic engine that produced the inventory in the first place. So every figure on the curve traces back to data somebody approved.

The honest limit

We should be clear about what this is not, yet.

This does not connect to your finance system. It does not read your capital budget, post to your ledger, or replace your investment appraisal process. What it does is produce the analysis in the form that process expects — levelised cost, payback, capital intensity, a ranked programme with a net annual figure — so the conversation can start from something other than a blank spreadsheet.

That integration is a reasonable thing to want and it is on the roadmap. Claiming it today would be exactly the kind of overstatement this whole feature is designed to avoid.

Why this matters more than another dashboard

The sustainability software market has converged on a shape: capture data, produce a disclosure, show a dashboard. That shape assumes the hard part is measurement.

It isn't. Measurement is laborious but tractable. The hard part is that the person responsible for reducing emissions usually has no budget, no capital authority, and no way to make the case in terms the people with both will accept.

A cost curve does not solve that by itself. But it changes what the sustainability lead walks into the room holding. Not a report about last year. A ranked investment programme, priced in the terms finance already uses, at their own discount rate, with the self-funding measures separated from the ones that need a decision.

That is a different conversation. And it is the one that actually moves tonnes.