Why this keeps happening

Someone checks whether Product A beats Product B on kg CO2e. Someone else checks whether Product A beats Product B on price. Nobody checks whether the carbon Product A removes, divided by what it costs to remove it, beats the next swap on the list.

The RIBA 2030 Climate Challenge gives a project an embodied carbon target to hit. Nothing in that process asks whether the money spent chasing it, swap by swap, was well spent. Two reductions can look identical on a carbon report and be worth entirely different amounts of scrutiny.

It is the construction equivalent of judging two diets purely on calories and never once asking what either one costs you in willpower. Both numbers matter. Only one of them usually makes it into the meeting.

The test: cost per tonne

Not producing a carbon number, but knowing whether paying for it was worth it. It borrows its benchmark from a Trellis framework built on Goldman Sachs figures, of all places, an outfit not usually consulted for moral clarity.

Run your own swap

Enter the cost premium and the carbon saved for a swap you’re weighing up, and see which band it actually falls into.

Three swaps, ranked

The three swaps below, run through the cost per tonne test. Swap 1 uses the sourced Harwell EPAC figures; Swaps 2 and 3 are illustrative figures only, not measured results. Replace them with your own product data.

Counting the reduction is the trickier half

Every band above assumes the tonne of CO2e on each side of the comparison is the same tonne. That assumption is where most cost per tonne exercises quietly go wrong.

The Construction Products Association’s review of EPD comparability put a name to this problem: two declarations can both be correctly produced and still not be comparable to each other. Nothing on the front page of an EPD flags that for you, so four checks have to happen before a cost per tonne figure is trusted. It is the least exciting paragraph on this page, and also the one most likely to save you from an argument you cannot win.

None of this means the EPD system does not work. It means a kg CO2e figure is a starting point, treated as a range with a margin of error, not an exact number to three significant figures. Tighter alignment is coming through EN 15941, the data quality standard published in 2024, but EPDs will take time to catch up to it. Until they do, working from data that has already been standardized and verified does more of that work for you than reconciling mismatched EPDs by hand.

That is the specific problem 2050 Materials’ database of environmental product declarations is built to solve, by holding products to a consistent methodology before a comparison ever reaches a cost report.

Cement replacement quietly reduced

This one is easy to miss because nothing on the drawings changes. The concrete grade stays the same, the structural design doesn’t move, but the GGBS content in the mix can creep down from what was modelled early in design back toward standard Portland cement.

It is the rare value engineering conversation that argues for itself. Enjoy it. They do not come along often.

Getting a reliable answer starts with a product level EPD rather than a category average, which is exactly the kind of lookup 2050 Materials’ platform is built to make quick instead of a research project in itself.

Timber frame traded for steel or concrete

Or the reverse: mass timber never seriously priced against the frame it could replace. Mass timber schemes look good on the carbon budget and the planning application at concept stage, and the moment a taller core, a longer span, or a fire strategy question shows up, the frame quietly reverts to steel or concrete, carbon consequence unpriced. Which is a polite way of saying nobody wanted to be the one to write “we swapped the timber for concrete” into the sustainability statement.

Run the comparison the other direction and the same gap applies. The embodied carbon case for mass timber is well evidenced, but it typically arrives with a cost premium, a fire strategy and acoustics conversation, and a residual value question at end of life. That puts it in the middle or top band of the test, not the bottom, and skipping the analysis because timber is assumed to be the sustainable choice misses exactly what the test is for.

Whichever direction the swap runs, a fair comparison needs a carbon figure and a cost figure sitting side by side on the same basis, which is easier when both numbers already live in one place, such as 2050 Materials’ platform, rather than two spreadsheets that rarely get compared directly.

Facade or brick buildup, value engineered on looks alone

Facades are visible, expensive, and full of components that all look interchangeable on a cost plan. A recycled content brick or block gets swapped for a cheaper standard equivalent, the elevation looks the same in the render, and the carbon difference, sometimes substantial, sometimes marginal, never gets priced against what else that budget could have done. A render has never once lost an argument on carbon grounds. That is exactly the problem with letting it settle one.

This is precisely the kind of comparison the Embodied Carbon Optimiser on 2050 Materials is built to make instant, swapping components and seeing the carbon consequence right away rather than reconstructing it after the fact.

The five minute check

Tick each box as you actually would in a value engineering meeting. Watch what happens to the answer.

Where 2050 Materials earns its keep here

All three swaps share the same gap: a comparable, verified number that never made it into the room before the instruction went out. Nobody enjoys chasing that number down at 4pm on a Friday before a cost report is due, which is a fairly specific scenario this section exists to prevent.

Useful whether you’re a specifier running the comparison or a manufacturer whose product needs to be found in time to make the case for itself.