Our method

Translate the decision into answerable questions

A broad request to investigate a business rarely provides a useful starting point. The requirement should identify the decision, the people authorized to instruct research, the entities involved and the uncertainties that could change the outcome. An investment team may need to test an ownership claim; counsel may need a chronology; a board may need to understand how a policy development affects a commercial relationship. These are different questions and should not produce interchangeable research plans.

The initial brief also records what the organization already knows, which documents have been reviewed and which assumptions remain untested. A materiality threshold helps distinguish issues that merit further enquiry from background information. Reporting dates, geographical scope, exclusions and changes to the mandate should be agreed before collection. Where authority or purpose is unclear, the question must be resolved before the research proceeds.

Build a source plan that can withstand challenge

Source selection follows the question. Corporate registries may support an ownership chronology, published court material may establish a litigation event, and an appropriately conducted expert interview may explain a market practice. A source plan should identify what each category can establish and what it cannot. Public availability, apparently official presentation or a prominent interviewee does not by itself establish reliability.

Collection should retain enough context to understand a record later: issuing organization, relevant date, entity identifiers and the distinction between an original publication and a reproduction. Human-source accounts require a separate assessment of first-hand knowledge, recency, incentives and possible bias. If several reports repeat the same original allegation, they are not independent confirmations. The plan should leave room for contradictory evidence and information that undermines an initial hypothesis.

Separate facts, interpretations and unresolved leads

Analysis is the process of explaining why the collected information matters. It may involve a chronology, an ownership map, comparison of competing accounts or structured scenarios. Entity matching is particularly important: similar names, transliterations and shared addresses can produce false connections. A relationship diagram should show whether a link is documented, reported or inferred rather than making every connection appear equally established.

A useful assessment considers alternative explanations and states what evidence could change the conclusion. An absent record may reflect limited access, a different filing system or an incomplete search rather than misconduct. Confidence describes the strength of the evidence and reasoning; it is not the same as the probability of a future event. Material gaps should remain visible in the report, not disappear in the executive summary.

Brief the decision-maker and define the next enquiry

The report should start with the question, the material findings and their implications. A supporting account then explains sources, confidence, contradictory information and the limits of the research. Different readers need different levels of detail: a board may need an executive assessment, while counsel or a diligence team may require documented leads and a source trail. Access and handling arrangements should reflect the mandate and sensitivity of the material.

Intelligence does not remove the need for professional judgment. The final discussion should identify what can reasonably be decided now, what requires further research and what belongs with legal, regulatory, financial or other qualified advisers. A refresh or monitoring brief is useful only where new information could affect the decision. Its triggers, reporting format and scope should be agreed rather than assumed to continue indefinitely.

What clients should expect at each stage

At the requirement stage, a client should expect to be asked pointed questions rather than to dictate a technique: what decision is pending, by when, who is authorized to instruct the work, and what would make the findings actionable. We also ask what has already been tried, since duplicating existing diligence wastes the client's budget and our time. A requirement that cannot be tied to a specific decision is usually not ready for research and benefits from further internal scoping first.

During collection, clients are entitled to know, at a summary level, what categories of source are being used and why a particular one was excluded — for instance, a restricted database the firm does not access, or a jurisdiction where a record is simply not public. Source validation is not a single checkbox: each item is assessed for who produced it, when, with what apparent interest, and whether other independent sources support it. A claim repeated across several outlets that trace to one original report is treated as one data point, not several.

At the analysis stage, the test we apply to our own work is whether a sceptical reader could trace each conclusion back to its evidence and judge the confidence attached to it. We actively look for information that would undermine a preliminary view, because a report that only confirms the client's existing assumption is of limited value. The final brief states clearly what remains unresolved, what would change the assessment, and whether continued monitoring or a defined follow-up enquiry is proportionate — reassessment is treated as a normal part of the method, not a sign the first report was deficient.