Ten years, four business models, and why we repositioned around independent verification rather than adding another product line.

Sustainability moved into the finance function. The risk is that it comes to mean whatever is measurable rather than whatever is material.

Verification spending is the installation cost of pricing externalised damage. What would make it cheaper, and who controls the data layer underneath.

EU sustainability rules apply flat compliance costs to suppliers of very unequal capacity. The distributional effect, and three things buyers can do about it.

Farmers generate the primary data behind corporate disclosure, carry the verification cost, and capture a minority of the value. Three counter-arguments.

Harmonisation fails on purpose, not on format. A two-layer architecture that lets one farm event mean different things to five institutions.

Audit programmes visit a sample of farms; claims cover all of them. What continuous monitoring closes, and the three things it genuinely cannot do.

CSRD scope narrowed in 2026, but LSR arrives in January 2027. What separates agricultural Scope 3 data an assurer accepts from data they cannot test.

The reporting obligation moved in 2026 and most agricultural Scope 3 guidance has not caught up. This piece sets out what applies after CSRD Omnibus I, what GHG Protocol LSR requires from January 2027, and what separates data an assurance provider will accept from data they will not.

Optical satellites go blind under cloud. How Sentinel-1 radar fusion restores a continuous NDVI series — and why fused values are estimates, not measurements.

German farmers expected the Düngeverordnung to cost yield. Two decades of nitrogen data suggest the regulation rewarded regenerative practice instead.

A 0.1% soil carbon gain does not hold 200,000 litres per hectare. The soil-mass arithmetic, the measured figure, and what soil carbon really buffers.

Sustainability resists one number. A narrow, testable satellite proxy for German nitrogen surplus, what it showed, and where it fell short.

Sample count follows from your confidence level and minimum detectable difference - not from a fixed ratio. What stratification changes, and what it cannot.

Verification usually takes 25–50% of a soil carbon project's budget. Across 8,845 ha of ESA pilots, stratified sampling brought it to 4.1% of credit revenue.

Practice is spreading faster than proof. Sequestration uncertainty bands, reversibility, and why permanence provisions matter more than a single measurement.

Trillion-dollar carbon figures describe allowance trading, not the voluntary market where soil carbon sits. What the July 2026 CRCF methodologies changed.

VT0014 changes the soil core's job from measuring an average to validating a model. What that does to cost, and the three risks that move with it.

A field map is a decision instrument, not a picture. Why perceptually ordered colour scales change which zones an agronomist actually acts on.