How Arc works.
Arc is based on deterministic calculations at its core, and uses probability to explain itself – to marketers, business people, and policymakers – and to help you make better decisions. It is built on samples and statistics: honed, not infallible.
Underneath sits a selection of published models – see the segments – tweaked with industry knowledge and a little smarts to bring them to life, much aligned to the marketing theory in Green But Not Heard by Mat Moakes.
Our evidence base combines primary research with UK households, run across three studies during 2026, with published third-party segmentation. Method notes, sample composition and field dates are available on request. It measures population shares, persuasion response and whether real people recognise themselves in the output – they do, at a mean accuracy of 4.0 out of 5, with one in five calling their result uncannily accurate.
Every claim the engine makes carries a grade: measured in our own studies, drawn from published evidence, editorial judgement awaiting citation, or predictive – stated as prediction, at stated odds, and validated continuously against live consumer confirmation. Findings only publish after pre-registered statistical gates and correction for multiple comparisons; this year that standard cost us several of our own favourite findings, which is exactly what it is for.
The detailed methodology – gates, thresholds, weight provenance and the full adjudication record – is available on request for serious evaluations.
Built from understanding, honed by evidence.
What we build comes from a deep understanding of decarbonisation – the technology, the properties, and above all the consumers. We combine the segmentation models in the public domain with our own research from the Green But Not Heard project, apply deterministic calculations, and use AI to make the insights richer. Honed with data and feedback – and honest about being built on samples and statistics.
The archetype population shares now come from our own representative quota pilot, run in 2026 with 226 people. That is a real measurement rather than an assumption, and it is still a small one – treat the shares as estimates good to roughly seven points at worst, and expect them to firm up as the sample grows.
Four appeal types – financial payback, independence, moral and planet, and social proof – were scored 0 to 10 by members of each archetype in our 2026 pilot. Where a measured score exists for an archetype and an appeal type, the band comes from the score under a mapping fixed in advance: 7.0 and above is receptive, 5.5 to 6.9 mixed, 4.0 to 5.4 wary, and below 4.0 resistant. Those cells are marked as measured, with the number of people behind them. Every other cell keeps the editorial matrix and its published citations, unchanged.
A finding rather than an accident: under this mapping no measured cell falls below 4.0. Nobody in the pilot was outright hostile to any of the four appeal types, so ‘Likely resistant’ is unreachable for measured cells in v2.0. Resistance in the model comes from how a message is framed, not from the underlying appeal.
v1.1 – 13 September 2026. Lens 2 archetype prior updated from uniform to measured population shares (representative quota pilot, 2026, n=226). All regression expectations recomputed. Independent cross-implementation verification scheduled.
Message layer v2.0 – 13 September 2026. Editorial receptivity matrix augmented with first-party measured appeal-type scores (GBNH pilot 2026) for four archetypes under a pre-registered score-to-band mapping. Editorial cells unchanged.
- More in Common – the seven British segments (2025)Source →
- Britain Talks Climate & Nature – Climate Outreach & More in Common (2025)Source →
- GBNH Archetypes – our own lens, sizes measured in a representative pilot (2026, n=226)
- Defra – Pro-Environmental Behaviour Framework (2008)Source →
- Rogers – Diffusion of Innovations (1962)Source →
- Yale – Global Warming’s Six AmericasSource →
- Nesta – heat pump consumer segmentation, eight audiences (2025)Source →
Arc is for strategy and communication design – a tool for being thoughtful, strategic, and measured across product, price, promotion, place, people, and process. Never pricing, credit scoring, or the exclusion of any group.
An auditable wrong number is still a wrong number – which is why the working is always on the page.