Invenio Protocols / Research / Note 01
Methodology · Research note 01
When a source supports a fact, but not the decision.
A documented fact is the beginning of a commercial judgement. The reasoning, the boundary of the evidence and the questions left unresolved need to remain visible too.
Note 01 Evidence & judgement
A research file can be accurate sentence by sentence and still leave its reader unable to decide. The missing step is often the connection between what a source establishes and what the commercial question requires. A citation should make that connection inspectable; it cannot supply the reasoning on its own.
Begin with the decision.
“Research this company” is a subject, not yet a useful boundary. Choosing a supplier, screening an acquisition and assessing a route into a new market require different evidence. The question determines the relevant geography, time horizon, comparison and exclusions. It also identifies what the research will not establish. Without that boundary, a detailed report may answer a question the client never needed to ask.
Separate the record from the interpretation.
Consider a hypothetical supplier whose filing reports revenue growth. That filing may support a statement about reported revenue in a defined period. It does not, by itself, establish that demand is durable, that customers are well diversified or that the supplier can meet a particular contract. Those are separate propositions. Each needs its own evidence or a clearly labelled analytical judgement.
Invenio’s Research Protocol distinguishes verified facts, sourced market estimates, directional benchmarks, inference and speculative hypotheses. The distinction helps a reader see where the record ends and reasoning begins. An inference can be useful, but its assumptions should remain available for challenge. A confidence label does not turn an unsupported proposition into a fact.
Make an evidence gap useful.
Missing information is most useful when paired with an action. If customer concentration is not disclosed, a brief should say so, explain why it matters to the decision and identify what would resolve it: a customer schedule, a management answer or another relevant record. “Unknown” describes the position. A precise follow-up question gives the reader a way forward.
The same discipline applies to conflicting sources. Record their dates, scope and basis before choosing between them. An older primary filing and a newer third-party estimate may describe different periods or definitions. Their disagreement should not be concealed inside a single precise-looking number. Where it remains unresolved, the uncertainty belongs in the conclusion.
Keep the date attached to the judgement.
The published Iomart Group pre-diligence specimen illustrates the presentation of evidence grades, source trails and questions for management. It is an analytical showcase, not a client engagement. Its stated position is as at June 2026, and the page has not been updated since production. The specimen can demonstrate a method without being a current assessment of the company.
That boundary matters whenever research is reused. A reader should be able to distinguish the date of a source, the period it describes and the date of the analysis. A later decision may require fresh evidence. Retaining the original date makes that need visible.
Review the argument, not only the citations.
Invenio’s published protocol includes human review before delivery. A useful review asks whether the evidence supports the wording, whether the assumptions are explicit and whether an alternative explanation could change the conclusion. It also asks whether each material unresolved question has a practical next step. A well-sourced file should make its limits easier to find.
The result is a judgement the reader can interrogate: what is known, what is inferred, what could change the view and what needs to happen next. That is the standard a commercial research brief should be built to meet.