Bread production
ADEME Base Empreinte — official LCA database
Impact Case Study
It's the question that breaks most impact numbers. We build the ones that answer it — starting with a circular food startup in Amsterdam.
Daniela Colagrande had already built circularity into almost every part of her business. What she didn't yet have was a way to prove it.
Polpette di Pane is a circular food startup where founder Daniela builds circularity into nearly every operation — from cargobike deliveries to turning used kitchen oil into soap through a circular partner. A low footprint isn't her marketing; it's how she works.
Her pilot proposal promised partners "measurable ESG impact." But like many early founders, that promise ran ahead of the proof behind it — a phrase that felt right, without a defensible number underneath. And the moment it starts to matter — a pitch, a grant application, a partner's due diligence — "measurable" has to mean something. That's exactly the gap: the distance between claiming impact and being able to prove it.
Food rescue makes this harder than it looks. Avoided-emissions figures are easy to inflate — and an inflated number collapses the first time a funder asks how you got it. Polpette didn't need a big number. They needed one that would survive scrutiny.
We built Polpette a screening-level avoided-emissions model — a structured way to estimate the CO₂e saved by keeping surplus bread in the human food chain instead of letting it become waste.
Nothing in it is assumed. Bread-production emissions come from ADEME's official database; the substitution and end-of-life assumptions are grounded in peer-reviewed literature and recognised LCA data. Nothing is a guess we couldn't defend.
At its core is a discipline we apply to every assessment: the counterfactual. We only claim a saving where the alternative — the bread genuinely going to waste — was real. Where it wasn't, we don't count it. The result reflects Polpette's true impact, built to hold up to a funder, a reviewer, or an auditor.
Based on this assessment's per-kg factor. A planning estimate — the exact figure is confirmed with real pilot data.
Every figure in the model traces back to a source. Bread-production emissions come from ADEME's official database; substitution and end-of-life assumptions are grounded in peer-reviewed academic literature and recognised LCA data — not estimates we invented.
We held a sharp line on counterfactual discipline: a stream only earns a number if the alternative to Polpette's action was real. For the bread, it was — surplus genuinely headed for waste. Streams that didn't meet that test stayed out.
And we closed the assessment the way any defensible model should: by cross-checking the result against published literature. Our figure of 0.59 kg CO₂e/kg sits comfortably inside the established range — independent confirmation that the number is neither inflated nor invented.
This is where most impact numbers quietly inflate — and where we drew a hard line. Polpette runs two genuinely circular activities beyond the bread. Both are real. Neither is in the CO₂e total.
All collection runs on an electric cargobike — Daniela's first business purchase, chosen from day one. A van was never an alternative she was going to use, so we can't claim we "avoided" van emissions that were never going to happen.
Used kitchen oil becomes soap through a circular partner. But that oil would have been recycled anyway — a separate collection company already existed. Claiming a saving there would mean taking credit for something that wasn't ours to claim.
We left them out of the numbers — and kept them in the story, as qualitative strengths. Counting them would have produced a bigger figure. It would also have produced a weaker one. Our goal was never the biggest number. It was the one that holds.
"Impact without assessment and real data to share is not impactful as it can be." I always said this to myself and yet always postponed my measurements. And then I found CarbonSquare and I realized that with real numbers and data my startup can spark real interest to investors and clients due to the capacity of visualising what I am actually achieving with my impact actions. And I would never go back.
Because a number you can defend is worth more than a number that impresses — the moment someone asks "how did you get that?", only one of them is still standing.
This matters most at scale. Accelerators, funds, and impact investors increasingly need the startups in their portfolios and programmes to show environmental impact that holds up — to funders, to auditors, to due diligence. A claim that collapses under questioning isn't just a weak number; it's a reputational risk for everyone who backed it. That's the gap we close: credible, defensible impact data a whole portfolio can stand behind.
CarbonSquare is built for people in exactly that position: