Measured progress, checked outside
Measured change, read by somebody with no fee riding on the reading.
An independent marketplace being built for nature-based improvements to water, air and the places they sit in: quantified carbon-dioxide equivalents, measured improvements in water quality, and the outcome record behind each. Providers bring the change, verifiers assess it without seeing who pays, and sponsors buy it. The architecture underneath is what anything priced against an outcome has to settle first.
State of play
What exists, and what is still a design.
Credible Exchange is a design. No market is running: there are no live listings, nothing has been issued to anybody, no standard-setting body has been appointed and no independent auditor has been engaged. The word verified stays off this platform until the last two exist.
What exists: the rules a provider must satisfy before its work can be listed. Three listing types (a quantified carbon-dioxide equivalent, a measured improvement in water quality, the funded units inside one program zone), all held to the same conditions. The machinery that decides which verifier assesses which program, the most finished piece of the design. And one provider whose work can be described in specifics, because a regulator has already examined it.
Three gates, in this order
Name a standard. Appoint a body to hold it. Engage an auditor who answers to that body, and whose finding that body can withdraw. A signature is worth nothing without a body behind it, and a body is worth nothing without a standard to apply.
No launch date is attached. Meanwhile a prospective sponsor or provider can name the coastline they care about, so the first zones defined are ones somebody is waiting for.
Why the architecture is shaped this way
Anything that pays out against a number has to settle first who reads the number.
No instrument is priced here and no market is running. The design still answers the questions one would put to it, because those questions are what separate a measurement from a claim.
A payment made against an outcome needs four things fixed before the work starts: where the outcome is measured, over what period, against which reference, and by whom. This platform fixes all four early, for reasons of its own — a market in this category comes apart at the joint between the assessed and the assessor — and those same four fixtures are what a loss calculation has to run on.
The zone edge is agreed in writing before a program opens, which turns the perimeter of an exposure into a document. The period is fixed, which makes an event datable. The reference reading is taken before the intervention it is a reference for, which is what stops a comparison being assembled afterward. And the fee is standardized by program type and paid by a party the assessor does not see, which is what takes the incentive out of the reading.
Where a structure runs off an index, those four decide whether it can settle at all. Agreed layers and bands, a floor, a ceiling, a stated trigger and adjustment on a published cycle each rest on readings a third party can stand behind. The perils in question already get written somewhere: bloom events, oxygen deficit, damage to natural resources, escaping oil, and what remains in a system after a remediation has closed. Each of those is measured at a place, over a period, against a reference, by a party somebody can name.
This platform runs none of that today. What it has is the part that has to be built first.
Four fixtures, and what each one closes
- The edge. A fixed square on a global grid, carrying the water column above it and the ground beneath. Two programs over the same water stay checkable against each other.
- The period. One square, one program and one period together make the unit. An outcome with no period attached cannot be set beside another one.
- The reference. Baseline before the intervention it is a baseline for, on a plan settled before any reading is taken against it. A reference picked afterward is a result picked afterward.
- The reader. A fee set by program type, a payor the assessor does not see, and assignment by randomized round robin to the next qualified assessor for that program type.
No standard has been named and no auditor engaged, so nothing here is called verified. Both gates sit ahead of a first listing.
The three sides
Three sides, and the one in the middle cannot see who is paying.
Each side gets its own tools. The tools keep the sides apart.
Providers bring the change
Operators who can put equipment and people on a specific stretch of water or land, hold a permit from the regulator of that jurisdiction, and produce a measurement trail somebody else can audit.
Two tools: a listing, and a payout against it. The bar for the first is deliberately awkward: environmental marketplaces most often fail on supply, listing intentions instead of operators.
What a provider has to show
Verifiers assess the change
Qualified assessment bodies and laboratories, one program type at a time. They define the method, integrate the metrics that method calls for, and are paid out on the same rails everyone else is.
They are never given the name of whoever covers their fee. Work reaches them in turn, at random, and they take it or leave it.
How a verifier is assigned
Sponsors buy the change
Corporates with an environmental commitment, funds with a stewardship mandate, philanthropies, resort and tourism groups, port and utility operators, and public bodies with a coastline to answer for.
They see what happened inside a program zone, buy the sponsorship, then certify, retire or transfer it to whoever should hold it. Picking units and sending support for them happens on a separate platform built for that job. This market is where the assessed result comes to rest.
What a sponsorship buys, exactly
A program zone is the bounded area a provider has committed to improving. A unit is narrower: one square of ground or water, under one program, for one period, and all three are required to name it. The same square under a different program is a different unit, and so is the same square next period. A supporter takes units one at a time or in a block, each carrying exactly one sponsor.
Listings are not restricted to one operator, one family of systems or one programming platform. Other vetted nature-based outcomes belong here on the same terms. A market selling only its parent's output would be a catalog with a fee attached.
There is no fixed order of events. Equipment is sometimes on the water long before a sponsor appears. Sometimes the work is finished and the outcome already logged, and a sponsor arrives afterward to retire it, move it or keep it.
And sometimes a sponsorship starts a program: systems spun up and sent to a zone against chartered outcomes, with nothing in the water on the day the money is committed.
Independence
The link between the developer and the auditor, taken apart on purpose.
This category of market keeps failing at the same joint: the party being assessed picks the assessor, and pays the assessor. Four design choices break that, and they hold without relying on anybody's good character.
- The verification fee is standardized The platform sets the amount for a program type, and neither the party being assessed nor the party assessing has a say. With no price to move, nobody shops for a cheaper audit.
- The payor is anonymous to the verifier The name of whoever covers the fee stops at the platform's ledger. Often that payor is the provider itself, running the work at its own cost and risk before any sponsor has bought anything, and the finding has to read the same whichever it was.
- Assignment is a randomized round robin A program goes to the next qualified verifier for that program type, drawn in turn from the pool holding that qualification. That verifier accepts it or passes. There is no list to work down and no relationship to cultivate.
- Nothing proceeds until the commitments are in place A program sits in a holding pattern until at least the provider and the verifier have both committed to it. An assessor engaged once the readings are in has already missed the baseline, the reading every later number is measured against.
Together they remove the direct financial link and the selection link between developer and auditor, the two levers that make a flattering result easy to arrange.
That settles who assesses. What they assess against is still open. Machinery is not a standard.
Verification
Six conditions a listing has to meet.
All six, or there is no listing. A market in this category has already failed on each one.
- A method published before the work starts What will be measured, with which instruments, at what depth or distance, on what schedule. Public before anyone touches the water, so it cannot be rewritten to fit the result.
- A baseline measured on the water, before anything starts Real readings from the actual site, against its own oceanography. A baseline back-calculated afterward encodes what its author already knew about the season.
- Measurement by a party with nothing riding on the answer Not the provider. Not the sponsor. Not a consultancy whose next contract depends on a good number. The assignment machinery exists to satisfy this condition, and everything else rests on it.
- A named standard, applied by a named body Neither has been settled, and the assignment machinery does not settle them. This is the largest open item on the platform, and the first of the three gates.
- Results published whether or not they flatter anybody Including seasons where the measurement came back flat, or worse than the year before, on the same schedule and in the same place as the good ones. A file with the poor seasons lifted out shows only which seasons the provider liked.
- A stated boundary on what the number means Every measurement has a place, a period and an edge. A listing describes what happened inside that edge and makes no claim about anything outside it.
Outcome data is anchored to a fixed spatial grid, and every reading carries the numbered square it was taken over. A zone can be read for what happened in it or, where a program is sponsored ahead of the work, for what has been committed to happen there.
Boundaries
The edges of what a sponsorship is.
A program can keep greenhouse gas the water would otherwise release, such as methane from a lake bed or nitrous oxide off a nutrient-loaded coast, out of the air, and Alarivean is eligible for climate finance where a case supports it. That is a separate instrument, kept out of every sponsorship.
- A sponsorship pays no return and appreciates in no market. A sponsor receives a perk redeemed in person: no transfer, no resale, no cash value. Standing is recomputed from the contribution record each time it is asked for, so there is no position to hand on, and a program is only activated here once it is configured that way.
- Whether a program carries an offset is a setting fixed when it is configured and declared in its listing. Some let supporters reach unit-tied carbon-dioxide equivalents through the exchange; many do not, and no program is assumed to do either. A funded unit inside a program zone is not an offset and entitles nobody to emit anything anywhere. A quantified equivalent is an offset, with its own standard, its own quantifier and its own counterparty. Both can sit on this platform. Neither may be dressed as the other.
- Not self-certified. A platform owned by the same group as one of its providers cannot verify that provider's results. The conflict is structural, so the auditor has to be somebody else, appointed under gate three.
- Not a source of impressive global totals. Published estimates of what the world's reefs are worth span nearly two orders of magnitude by method, useless for pricing anything. Only place-specific numbers with a named source get printed.
- Not a service subscription. An owner who wants their own bay held in better condition buys a year-round program on it directly from a licensed service operator, ending only if performance goes out of spec or if they are not satisfied once the minimum period has been served. A sponsor funds work on somebody else's water, over a stated stretch of it, and gets back attribution and the measurement file. Same crews, same instruments, two different things on paper.
Why bother building it
The stake is a working coastline, and it is measurable.
A reef is the fishery, the coastal defense and the tourism revenue in one structure.
In 2022, for the first time, farms produced more aquatic animals than boats caught. Food security is moving into penned water, and penned water fails first when oxygen drops or a bloom drifts in.
The fourth global coral bleaching event put 84 percent of the world's reef area, in 82 countries and territories, under heat stress, against 68 percent in the third event and 21 percent in the first.
A sponsor cannot fix that. A sponsor can pay for a bounded, measured repair on one named place, before it is carried out or once it has been, and prove it either way.
A provider on record
If your coastline is the problem, start with the water company.
Most of what fails on a populated coast fails through water: oxygen falling out of a bay, a bloom closing a beach or an intake, a heat event bleaching the reef a diving economy depends on. None arrives without warning to somebody who has read that water all year. A client subscribes to a standing program on a named body of water; the condition it must hold, and the safety limits bounding it, are settled on paper by the client and the authority with jurisdiction before any deployment.
Alarivean, Inc. deploys Sophia Delta water systems: SeaBreather as the platform, with ReefHealer, HABslayer and ChemSlayer as duties on it. By the operator's own account those carry most of the service work, and the proprietary development behind them runs continuously; nothing outside the company stands behind either statement. Alarivean is staffed from cities on several continents. It holds Florida Department of Environmental Protection permit FLOA00062 for a red tide mitigation field trial, documented independently by the Sarasota Bay Estuary Program and the nonprofit START.
One permitted trial is a field record of one. Florida's permit register confirms it.
The oxidation duty runs inside the vessel. What leaves it returns at flow to the water it was drawn from, with nothing tanked or carried off, and is measured against a number at the outlet, so a bottle can be filled where the work is happening, alongside a hull still on station. The number is the limit governing that receiving water, settled by the regulator that consents the work and stated in the service agreement before a hull moves. It is never lifted off a drinking-water code, because a bay is not a mains supply. Every program is designed to seat local scientific institutions on the crew's live water-quality feeds for as long as it runs.
Institutions on those feeds are oversight: somebody outside the company reading the instruments while the work happens, which exists today because the operator builds it into the program. Verification is an auditor applying a published standard for a body that holds it, and gate three is why nobody claims that word yet. Either can fill a bottle at the sampling point and send it to a laboratory of their own choosing. Whoever holds that water, or a qualified agent acting for them, can put a live deployment on hold the moment they suspect the agreed limit has been crossed, until the doubt is settled. A listing issued here would have to survive exactly that.
The operator contracts to the final state of the water and to the safety parameters bounding it. Which technique reaches which water stays the company's choice. Treatment and separation capability sits either side of the oxidation stage, chosen against what the water is doing that week. A listing written against a final state (the number, the authority that agreed it, the party holding the bottle) survives that; one written against a technique goes stale the first time the operator finds a better one.
Nobody has published an open-water result for this class of treatment yet, Alarivean included. The estuary program has asked for the job of measuring it. Technical help went into that permit application; funding deliberately did not, which leaves the program free to publish whatever the readings show. It wants any claim reaching beyond Sarasota Bay to stand on numbers its own scientists collected, and so does the operator. Producing that evidence is the next piece of work on this coastline.
Related in the group
Immediate, Significant, Scaled
Tell an operator about the place.
Whether you would sponsor the work or need it done, the first message is the same: which water, what goes wrong, in which months, and what it costs when it does. Alarivean answers with a read on that specific site.