Every number should
lead back to its evidence.

Three lenses. No single score.

Potential is a scenario based on explicit inputs. Discovery simulators are illustrations, not participant outcomes. Commitments alone receive no promised environmental savings.

Self-reported describes what participants tell us. Some activity supports a modelled estimate using their measurements. It is still not independently verified.

Verified requires an approved evidence and review process. This release credits no independently verified environmental outcomes. Email verification confirms email access only.

Water: measure before multiplying.

flow-time-v1: litres per minute × genuinely avoided running minutes per occasion × reported occasions = estimated litres avoided.

leak-duration-v1: prevented leak flow (L/min) × actual leak-free minutes within this reporting week. The number of repair occasions does not multiply the same repair again.

No universal default flow is applied. Enter your measured flow and genuinely avoided time. Fittings, pressure and habits vary. Do not report the same event under multiple water actions or the same leak-free period in two weeks. Confidence is limited because inputs are participant-reported.

EPA WaterSense guidance ↗ provides background, not a universal Indian household factor.

Waste: count what was avoided.

bags-avoided-v1: one eligible shopping occasion = one disposable bag avoided, only where a bag would otherwise have been supplied. No lifecycle carbon saving is claimed.

Other waste actions record occasions. Food-waste kilograms require a defensible baseline and are not estimated here. EPA reducing and reusing basics ↗.

Trees: planting is the beginning.

trees-survival-v1: newly planted trees, trees checked and trees observed alive are separate self-reported fields. Repeated survival checks are observations, not additional unique trees. No lifetime carbon storage or guaranteed maturity is credited.

Plant with permission, suitable local species and a maintenance plan. UN Decade on Ecosystem Restoration ↗.

Air & energy: no unsupported precision.

Travel and appliance actions record activity. Emissions estimates need the baseline vehicle, distance, replacement mode, factor and boundary. We do not invent them. Department of Energy guidance ↗ is background, not a universal local emissions factor.

Counting fairly.

Each verified account counts once. Each selected action counts once as an active commitment. A participant, action and Monday-start week has one report; editing replaces it. Participants must exclude overlapping activities. The system prevents duplicate records but cannot independently prove self-reports or detect every overlapping real-world event.

A positive weekly report is not a seven-day streak. Consecutive reporting weeks are labelled as such. Referrals count newly verified accounts, not clicks. City labels are optional and unverified. Period filters apply to reports; participant and commitment counters show current state.

Versioned and reviewable.

Each estimate stores its formula identifier, version, unit, inputs, assumptions and reporting time. Historical estimates retain their stored version unless an approved correction changes them. Litres, bags and trees are never combined into an arbitrary score.

Version challenge-1.0 uses approved formulas in code. CMS assumption edits do not silently change executable formulas. A new calculation requires review, tests and a version change. Milestones recognise reporting, not certified environmental performance.

Questions can go through AWE’s correction process.