How Ground-Truth Assessments Reduce Market-Entry Risk
Verification does not remove risk. It converts unknown risk into known risk, which is the only kind you can manage.
Verification pays where an assumption is load-bearing, wrong often enough to matter, cheap to check, and expensive to discover late. Most plans have three to six such items. Checking those returns most of the available risk reduction; checking twenty returns marginally more at several times the cost.
"Ground truth" is easy to sell and easy to over-sell. It does not remove risk, it does not guarantee an outcome, and it is not always worth commissioning. This article sets out where it actually changes results and how to decide what to check.
What verification actually does
It converts an unknown into a known. That sounds modest, and it is the entire mechanism. An organisation cannot manage a risk it has not identified, cannot price a contingency it has not sized, and cannot decide against something it believes to be settled.
Verification produces value in four distinct ways, and they are worth separating because they have different magnitudes:
| Mechanism | How value arises | Typical magnitude |
|---|---|---|
| Avoided commitment | The plan is materially wrong and the company does not proceed, or restructures before committing | Largest — the entire committed cost |
| Avoided rework | A wrong assumption is corrected while the plan is still cheap to change | Large — rework late costs several times rework early |
| Right-sized contingency | Known conditions replace guesses, so contingency is set deliberately rather than defensively | Moderate — frequently reduces cost |
| Confirmed assumption | The assumption was right; the plan proceeds with justified confidence | Small directly, but it is what makes the process credible |
The fourth row matters. A verification programme that only ever produces bad news is being aimed at the wrong things, or is being run by someone with an incentive to find problems.
A test for what is worth checking
An assumption is worth verifying when all four of these hold:
- Load-bearing. If it is wrong, the decision or the plan changes. If nothing changes, do not check it.
- Plausibly wrong. There is a real base rate of this being wrong in this market. Not everything is.
- Cheaply checkable. It can be established by observation rather than by a long study.
- Expensive to discover late. Finding out during mobilization costs materially more than finding out now.
In our assessment, the items that most reliably satisfy all four conditions in Venezuela are: power availability at the working location, the corridor's controlling constraint, the usable medical facility and its real travel time, counterparty capability, and accommodation suitability. They are load-bearing, frequently wrong, observable in a day or two, and painful to discover after people and equipment are committed.
Where verification does not help
Being clear about this is part of being useful:
- Where the answer would not change anything. Confirming a fact that is already priced in and non-binding is spending without effect.
- Where the question is not observable. Intent, future behaviour, and political direction are not field questions.
- Where a qualified determination is needed. Load ratings, sanctions positions, and clinical judgements require engineers, counsel, and clinicians, not observers.
- Where the finding will be ignored. If the decision is already made politically, verification produces a document, not a decision.
- Where it is done too early. Condition-sensitive findings verified a year out will have to be redone.
How to make findings actually change decisions
The most common waste is not unverified assumptions — it is verified findings that did not travel. Three practices prevent it:
- Tie every finding to the assumption it tests. A finding filed under "site report" gets read once. A finding filed against "assumption 4: road access supports the mobilization schedule" gets acted on.
- Give unverified items the same visibility as verified ones. The list of what could not be established belongs in the executive summary, not an appendix.
- Name what would change the decision. Agree in advance what finding would cause the company to stop, delay, or restructure. It is much harder to rationalise away a finding you pre-committed to act on.
Sequencing that gets the most out of the spend
- Build the assumption register from the plan, including the undocumented assumptions.
- Rank by consequence, not by ease.
- Apply the four-part test and select a small number of items.
- Verify those, with the method and limits recorded.
- Feed findings back and let the plan change shape.
- Re-verify condition-sensitive items close to commitment.
- Record everything still unverified as accepted risk, explicitly.
What remains uncertain
Verification is a snapshot taken under whatever access was available, and its useful life varies by what was checked. It also has a selection limit: it can only test the assumptions someone thought to write down. The assumption register is therefore the highest-leverage document in the process, and the one most worth having an outsider stress-test — because the assumptions a team has not noticed it is making are, by construction, the ones it will not commission anyone to check.
Related capabilities
Which of your assumptions are load-bearing?
Request an independent assessment. We will build the register and tell you which items are worth checking.