In One Paragraph

Almost every "website carbon" tool on the market estimates a site's emissions from one number — how many bytes the page weighs — multiplied by a yearly national average for how dirty the electricity grid is. It prints a single figure with no margin of error and no way to check it. EcoPigs does something fundamentally different: where a site installs our badge, we measure the carbon of its real visitors — the actual energy their devices use, against the live electricity grid at the moment they visit — and we report it as an honest range, not a falsely precise number.

1. Why This Matters Now

Carbon reporting is becoming mandatory. CDP disclosures, SECR, B Corp, CSRD — more organisations every year must put a number against their digital footprint. The problem: the numbers the industry currently produces are estimates dressed up as facts.

A typical web-carbon estimate uses page weight as a stand-in for energy — so it cannot tell the difference between a 1 MB photo and a 1 MB piece of heavy, battery-draining code. It uses a year-old national average for grid carbon, ignoring that electricity is far cleaner at 3am on wind than at 6pm on gas. It assumes a cold first visit every time, ignoring that real visitors arrive with files already cached. And it reports one number with no error bars, implying a precision it does not have.

For a marketing badge, that is tolerable. For a figure going into a regulated disclosure, it is not. What you can measure, you can manage — and what you can measure, you can defend. That gap is what EcoPigs exists to close.

2. What We Built

EcoPigs v3.0 is built on the EcoPigs Measured-Carbon Model (EMCM), our published methodology. In plain terms, it does four things no estimate-only tool does:

It measures real visits, not a lab guess. When the EcoPigs badge is on a site, it records the energy real visitors' devices actually spend loading each page (based on the work their processor does), and the real mix of new versus returning visitors. The headline figure becomes the true average a real visitor experiences — not a one-off cold scan.

It uses the live electricity grid. Emissions are weighted by the carbon intensity of the grid at the time and place of the visit, drawn from live national grid data — not last year's average.

It calibrates against a real dataset. A brand-new site with little data is compared against the patterns we have measured across hundreds of similar sites, so its estimate is grounded rather than guessed — and as it gathers its own real visits, the figure shifts to rely on its data. The more we measure, the better every estimate gets.

It tells the truth about uncertainty. Every figure comes with a range, and every part of it is labelled by how we know it: measured, modelled, or estimated. We never show a bare number as if it were exact.

The shift in one line: from estimating a website's carbon to measuring it — and being honest about the difference.

3. The Estimate → Measurement Ladder

EcoPigs gives a result at two levels, and we are explicit about which you are looking at. A free scan (no badge) is a one-off estimate of a cold first visit — a modelled figure, with a range. With the badge installed, it becomes a measurement of your real visitors — your actual visitor-average, with a tighter range. Both use live grid data and the full lifecycle.

A real example. A cold scan of one of our own sites estimated 0.029 g of CO₂e per visit. Measured across its real visitors, the true figure was 0.012 g — about 2.5× lower, because real people arrive with files cached and the cold scan cannot see that. The letter grade barely changed; the reportable number changed a lot. That difference is exactly why a site doing a serious carbon disclosure needs measurement, not an estimate.

Comparisons like this are always a site against itself — its own scan versus its own real visits. We never imply the gap between two different sites is "what the badge did."

4. The Honesty Doctrine

This is the part designed to survive scrutiny from a sceptical expert, and it is the heart of the brand.

• No invented figures. If we cannot measure or credibly model something, we say so — "unavailable" beats a made-up number.
• Ranges, not false precision. Every figure carries a 90% confidence range.
• Labelled by evidence. Each number is tagged measured, modelled, or estimated.
• Reproducible on demand. For any score, we can show the full working — inputs, factors, sources, and the uncertainty behind it.
• Honest about our own weak spots. Where a coefficient is a conservative estimate rather than a measured value, we publish that openly rather than hide it.

We would rather report a wider, honest range than a narrow, flattering one. A wide range that is true is worth more than a precise number that is not.

5. What We Measure, What We Model

Total website carbon has two parts, and we treat them differently and transparently. The operational footprint — the energy of running the visit across device, network and data centre, against the live grid — is measured on badge sites, from real device energy and real visits. The embodied footprint — the manufacturing and lifecycle carbon of the devices and infrastructure involved — is modelled using an industry-standard method, and is typically the larger share of the total.

Because the embodied share is large and currently modelled rather than measured, its uncertainty dominates the overall range — and our figures show that honestly, rather than letting the measured part make the whole number look more precise than it is. This is deliberate, and it sets up our next release.

Two figures, one hierarchy. Because of that split, we show two things, never in competition. Your score and grade are what you can actually move — the operational footprint, the headline. Beneath it sits an Estimated System Footprint: the fuller picture, adding a fair time-share of your visitors' device-manufacturing carbon. It is a bigger number (roughly 10× the score), with a wide band and low confidence, and it is context, not a truer number. The device-manufacturing share is an allocation of hardware that exists regardless of the visit — informational, not additive to a corporate carbon inventory. We keep it out of the grade on purpose: a grade has to respond to action, and manufacturing-by-session-time is something a developer cannot move. In one line: your score is what you can move; your system footprint is the whole picture, including the device life this visit rents — shown honestly, with its uncertainty.

6. The Dataset Behind It

EcoPigs is backed by a growing body of real measurement, not just a calculator: around 1,130 websites with carbon grades, 16,000+ measurements across 870+ sites over 10 continuous months (August 2025 to June 2026), across 13 countries.

We are precise about what this is. Today it is a deep dataset — continuous, real measurement over many months — rather than a broad survey of the entire web; it leans toward the UK and certain platforms. So we publish findings scoped to what the data supports (for example, "measured across N sites on this platform"), always with the sample size shown, never as a claim about "the web" as a whole. As more sites install the badge, the picture broadens — and that is the work ahead.

Why this is a moat, not just a feature: a competitor can copy a formula in a week. They cannot copy ten months of real measurements. The methodology and the dataset reinforce each other — the more we measure, the better and more authoritative every figure becomes.

7. Where It's Going

Our roadmap follows one consistent move: replace an assumption with a measurement.

• bytes → measured device energy ✓
• yearly grid average → live grid ✓
• assumed cold visit → measured real-visitor mix ✓
• next (v3.1): assumed device lifespan → measured active time — so that even the embodied (manufacturing) portion is attributed by real usage, not an assumption. The headline: "even your embodied footprint is measured, not assumed."

Beyond that: a public Web Carbon Index (benchmarks by sector, platform and country, scoped honestly), and the same rigour extended to the carbon of AI/LLM usage — a fast-growing footprint almost nobody measures yet.

8. Claims You Can Cite

• EcoPigs measures the real carbon of real visitors on badge-installed sites, rather than estimating from page weight alone.
• Emissions are weighted by the live electricity grid, not an annual average.
• Every figure is reported as a confidence range, labelled measured / modelled / estimated, and is reproducible on request.
• The operational footprint is measured; the embodied (manufacturing) footprint is modelled using an industry-standard method, and its uncertainty is reflected honestly in the range.
• Backed by 10 months of continuous real measurement across hundreds of sites — a deep dataset that broadens as adoption grows.

The full technical methodology, including all coefficients and citations (IEA, Ember, Malmodin, Cloud Carbon Footprint, the Green Software Foundation SCI / ISO 21031), is published in the EMCM technical methodology and is machine-readable at api.ecopigs.co.uk/api/v4/methodology, with a reproducible calculation trace available for any measured site.