Tell us the decision you face and the time you have. We can then discuss the research, the sources and the format that would help. Scope, timing and fees are agreed before work begins.
An hour on the decision you face, what you already know and where the evidence runs out. We use it to establish whether the question fits what we collect, what it would take to answer and what a useful output would contain. Most engagements start here, and some stop here, because the honest answer is sometimes that the record cannot carry the question.
A bounded piece of work on one policy, actor, counterparty or asset question. It returns the assessment, the evidence behind it, the alternatives we considered and the observation that would make us revise it. The question is written down before the work starts and the wording does not move once it is registered.
The record kept running against a small set of written questions, each with a date and a rule for settling it. You get developments as they clear the corroboration rule, the probability on each question with the reason it moved, and a score against what happened when the date arrives.
These are the shapes an engagement takes. What you actually receive is set out on what we deliver.
A useful brief starts with a shared definition of what the work can answer. Scope should reflect source availability and the time you have.
Identify the research user, the subject, and the consequence of getting the question wrong.
Confirm geography, sources, exclusions, delivery date, format, use rights and commercial terms.
Assess the source record, challenge the interpretation, and record any gaps affecting the answer.
Explain the conclusion, alternatives, limitations, and the observations that would trigger another review.
An asset, contract, counterparty, institution or place at the right level of specificity.
The action or internal discussion that the research is intended to inform.
The deadline, and the policy, operational or procedural milestone that creates it.
What you already know, which assumption is unresolved, and the evidence that would help.