These two words are used interchangeably in conversation and mean entirely different things in practice. Validation asks whether the plan is sound before the project runs. Verification asks whether the reported results actually happened. Both are performed by accredited independent bodies, and projects fail them for different reasons.
What Each One Tests
| Validation | Verification | |
|---|---|---|
| Timing | Before implementation | After each monitoring period |
| Subject | The project design document | The reported results |
| Core question | Is this plan capable of delivering credible credits? | Did the reported outcome actually occur, measured as planned? |
| Typical failure | Weak additionality argument, undefined sampling design | Broken data traceability, deviation from the plan |
Validation examines whether the sampling design is capable of supporting the precision claimed. Verification examines whether that design was actually followed and whether the numbers trace back to field records.
The Deviation Problem
The most common verification finding is not that the numbers are wrong. Rather, it is that the project did something different from what was validated, such as relocating plots for access, applying a different allometric equation, or missing monitoring intervals, all without documenting the deviation or its justification.
Deviations are not automatically fatal. Undocumented deviations usually are, because the assessor has no basis to judge whether the change preserved the design's statistical properties.
What Makes Both Easier
Three habits, all cheap at design time and expensive retroactively.
Write the sampling design down in enough detail that a third party could execute it identically. Record deviations as they happen, with reasoning, rather than reconstructing them afterwards. And keep the data chain queryable so verification becomes a matter of retrieval rather than reassembly.
Projects that do these three things spend materially less on every audit cycle for the life of the crediting period.
Move beyond estimates
Verifiers test sampling design, uncertainty and whether a number traces back to the field. TREEO dMRV captures that evidence in real time, in one auditable chain.
