Not a carbon footprint, an emissions inventory or a sustainability certification. An annual, frozen and signed report computed from real operator-reported activity sessions — never a live or on-demand number. Coverage is still nascent: most signals read as absent until operator adoption grows, and an early year may publish nothing at all. That is the intended behaviour, not a bug.
Before any year is written, the platform-wide evidence base must clear a coverage floor: at least 30 distinct contributing tenants AND at least 1,000 real reported sessions, evaluated over the same 90-day-lagged window as the report itself. Below either, nothing is published — the honest empty-state, not a flattering number. As the operator base grows, an official report will appear here; until then there is no live figure to show.
≥ 30 distinct contributing tenants AND ≥ 1,000 real reported sessions, or nothing is published for that year.
A zone is released only with ≥ 10 distinct contributing tenants; below that it is suppressed and only counted, never shown.
No emissions inventory, no LCA, no certification. A published score does not mean an operator is that percentage sustainable.
Programme lead: Marta Incarnato
Signals come from real reported outcome sessions — a session that actually happened — from operators whose tenant has given research consent. There is no on-demand proxy and no modelled session behind any number on this page.
Read a score honestly · At most this index composites a small set of declared operational signals for zones with enough real reported sessions. A score of 60/100 does not mean the operator is 60% sustainable. It is a weighted position on whichever signals a zone actually had data for, nothing more.
Up to six operational signals are composited per zone. Each is included only where the zone has data for it; the weights of the signals actually present are renormalised, so a zone scored on fewer signals is not silently zeroed — but it is flagged comparability: "partial" rather than compared directly to a fully-covered zone.
These are the live weights the engine composites. Access/travel is the highest-weighted signal and, deliberately, the least substantiable one — it is a participant-declared estimate, never presented as measured.
The declared low-impact share of how participants got to the activity: foot/bike/public transport count fully, mixed partially, car/boat partially, flight not at all.
This is a PARTICIPANT-DECLARED ESTIMATE of access mode, not a measured travel-emissions figure. It is the single highest-weighted signal and the least substantiable one — we say so on every published document, not just here.
The share of propulsion-classified sessions that ran without direct combustion during the activity itself (wind, human or gravity powered, for example).
An observable propulsion label, not net CO2e. It excludes support craft and equipment lifecycle, and it says nothing about travel to the site.
The operator-declared renewable share of the facility's energy source: renewable counts fully, mixed partially, grid not at all.
Operator-declared, not independently metered. Directional only — it is not an absolute energy-consumption disclosure.
Electric share of declared powered equipment against combustion equipment in the same session set.
Observed equipment category only. Grid mix, manufacturing and charging are not counted.
How evenly recorded sessions spread across each activity's own expected operable season, not a naive calendar year (so a genuinely short season, like skiing, isn't unfairly penalised for being short).
Needs at least 12 sessions in the zone to compute at all, and is dropped when the operable window is under about two months (too few monthly buckets for the number to mean anything). A concentration metric, not an environmental-impact one.
The declared share of equipment that is rental/shared rather than owned outright: rental/shared counts fully, mixed partially, owned partially.
A WEAK proxy for sharing / product-service circularity intensity, not a lifecycle or material-flow measurement. Deliberately differentiated so that declaring honestly can lower a zone's score, never merely inflate it.
The mean declared participant count per session, reported for transparency.
Disclosed only, never folded into the index. Intensity needs a carrying-capacity baseline we do not have yet, so scoring it would be arbitrary.
The ≥50% coverage gate · A signal is composited for a zone only when at least half of the zone's sessions carry it. Below that, the signal is dropped and the remaining weights renormalise. This is why the new signals will mostly read as absent for now: a single sparse declaration is not allowed to drive a heavily-weighted signal for a whole zone.
Missing signals are renormalised out rather than silently zeroed. That is honest, but it hurts comparability: a 75 built from two signals is not the same as a 75 built from all six. So every zone carries the signals it had, the effective weight applied (availableWeight) and a comparability class, and the document reports an overall dataCompleteness. A "full" zone is one where all six composited signals were present.
Known limitation · because a missing signal is renormalised away while a declared-but-poor signal counts against a zone, a zone that withholds a signal it would score badly on can read higher than one that declares honestly. We do not impute a default for a missing signal — that would penalise honest non-reporters — so instead we make the gap visible via comparability and availableWeight. A high score standing on thin coverage is disclosed, not hidden.
A session that actually happened, from a tenant with research consent. No on-demand proxy, no modelled session.
A zone is released only with ≥ 10 distinct contributing tenants; below that it is suppressed and only counted.
Before any year is written: ≥ 30 distinct contributing tenants AND ≥ 1,000 real sessions, or nothing is published — the honest empty-state, not a bug.
The reporting window is clamped 90 days behind now, so a just-completed period can't be published before its data settles.
A published year is immutable. Corrections create a new numbered revision that supersedes the prior one; the old revision is retained, never deleted.
The public endpoint always serves the most recently published, signed year (or a specific year on request). It never computes a fresh figure on demand. Every published document carries its year, revision, publication date, signing role and methodology version, so nothing public can be mistaken for a live or provisional figure.
The new signals depend on operators actually declaring access mode, facility energy, equipment provenance and group size per session. Until adoption grows, most zones will show most signals as absent, and an early year may clear the platform floor without clearing it for every signal — or may not publish at all. We would rather show that plainly than round up.
Protected-area proximity is a candidate future signal. It is not yet captured, not yet disclosed and not yet scored — nothing on this page or in the published document reflects it, and we say that explicitly rather than let the absence go unremarked. Wiring it would additionally require an approved capacity baseline before any score, not just a proximity flag, could be composited.
Average group size is captured and disclosed today, but deliberately held out of the composite: intensity needs a carrying-capacity baseline we do not have, so scoring it would be arbitrary.
Until that review and assurance conclude and can be named, we do not describe the Index as certified, assured, aligned or compliant with any external framework.
Every candidate zone fell below the k-anonymity floor, or the platform coverage floor itself was not met (in which case nothing is published at all). As the corpus grows, the released-zone list populates here, each row carrying its own signal coverage and comparability class — absent signals shown honestly as —, not as zero.
A made-up kitesurf operator in Tarifa, Spain, one zone in a fictional published year. Every number below is invented to show the shape of a real entry — including a signal that gets dropped for thin coverage, which is the typical near-term shape while operator adoption is still low. It is not observed platform data and not an environmental certification.
availableWeight: 0.70). The result is marked partial and should not be ranked against a full result. It does not say the operator is 71% sustainable.Three fictional archetypes showing how the signals move with the shape of an operator, and where a naive reading would go wrong. Illustrative only, not observed cohort data — most of these signals are still thinly declared across the real platform.
Wind-driven activity means a high non-combustion propulsion share. Because declaration rates start low, facility-energy and provenance signals are the ones most likely to be dropped for thin coverage in an early year, leaving access mode and non-combustion as the signals that actually move the score.
Non-combustion propulsion share
Electric-equipment share (rescue boats) · seasonal concentration
No combustion during the activity itself. Seasonal evenness is normalised to the operable window rather than a full year, so a short but evenly-run season doesn't score as artificially concentrated. Protected-area proximity — even though many touring lines cross sensitive terrain — is not a signal today; it is future roadmap only.
Non-combustion propulsion share · seasonal evenness (once ≥12 sessions/year)
Access/travel low-impact share, if declared honestly for long approach drives
A session-weighted view across mixed operators in a region. Useful as a coverage and composition tracker year over year, but only when the comparability class and dataCompleteness travel with the number — and only once the platform-wide coverage floor is cleared for that year at all.
Non-combustion share (modal mix) · facility renewable energy, as adoption grows
Equipment-provenance circularity (weak proxy, differentiated by design)
A published report may support sustainability reporting as supplementary evidence. It does not establish CSRD/ESRS, GRI, GSTC, Travelife or Blue Flag compliance or alignment. ESRS uses double materiality and firm-specific disclosures that a Goable signal does not replace. The crosswalk below shows the narrow slice of each framework the Index touches and the much larger remainder it does not — the full mapping is published separately (framework crosswalk document).
| Framework | Where a signal may be cited as evidence |
|---|---|
| CSRD / ESRS | Relates to E1 climate-change drivers and E5 resource-use/circularity signals; not an E1 GHG inventory and makes no double-materiality claim. |
| GRI | Directional inputs that correlate with GRI 302 (energy) and 305 (emissions) themes; not a 302/305 disclosure, and does not compute Scope 1/2/3. |
| GSTC | Touches Section D energy, transport and seasonality themes; protected-area/biodiversity themes are addressed only by the future, not-yet-scored signal. |
| Travelife / Blue Flag | Complementary operational signals; the certifications set their own audited criteria and are not conferred by this Index. |
Raw access is for tenants on their own data, or for authorised researchers under disclosure controls. Third-party data, cross-operator data, re-identification in rare zones and commercially sensitive demand signals are all out of scope.
A published year may be cited as a supplementary, methodologically documented operational indicator where appropriate. It is not a compliance statement and does not stand in for an audited disclosure.
Recorded from real operator-reported outcome sessions under per-tenant research consent; self-declared unless stated otherwise, not independently verified.
Given a published document, the methodology at its stamped version, and the consented sessions for the period, a third party can recompute every released zone index and the overall figure. Suppressed zones are, by construction, not reproducible from the public document — that is the point of k-anonymity.
When a year clears the coverage floor, the machine-readable, signed document is published under Creative Commons Attribution 4.0 International. It may be cited as a supplementary, methodologically documented operational indicator where appropriate; the only ask is the attribution string carried on the document itself.
Goable Sustainability Index. No official report published yet. CC BY 4.0. https://goable.io/sustainability-index
The methodology is public and unfinished by design. If you run outdoor operations, review sustainability data, or study protected-area governance, the fastest way to improve this index is to challenge it. Design partners help shape the signals and the carrying-capacity work; reviewers help pressure-test the wording and the limits stated on this page.