How we
work.

Our method

Open a step to see its purpose and what to document.

1. 01Frame the question+

   Define the problem, operational context, scope and decision the investigation is intended to inform.

   What to document

   Record the research question, the unit of analysis, the intended decision and the criteria for a useful answer.

2. 02Represent the system+

   Examine activities, resources, information flows, dependencies, constraints and feedback relationships.

   What to document

   Describe how the work actually happens. Keep observations, stakeholder accounts and modelling assumptions distinguishable.

3. 03Specify the evidence+

   Translate research questions and competing hypotheses into measurable variables, data requirements and quality criteria.

   What to document

   Define required fields, observation periods, comparisons and measurement gaps before selecting an analytical method.

4. 04Prepare and govern the data+

   Establish provenance, access authority, reproducible transformations and appropriate safeguards.

   What to document

   Preserve original records. Document permitted use, transformations, quality failures and access restrictions. Use synthetic or explicitly authorised material for public experiments.

5. 05Select and test the method+

   Apply suitable analytical, statistical, machine learning, simulation or optimisation techniques. Evaluate assumptions, uncertainty and validity.

   What to document

   Begin with a deterministic baseline where appropriate. Check method requirements, prevent evaluation leakage and record parameters, versions and diagnostics.

6. 06Evaluate alternatives+

   Compare findings and possible interventions against operational constraints, risks and objectives.

   What to document

   State what each alternative requires, which assumptions drive the comparison and where evidence is insufficient. Keep recommendations subject to human judgement.

7. 07Engineer and observe+

   Translate justified findings into practical systems or operational changes, then measure outcomes and revise the model as new evidence emerges.

   What to document

   Version the implementation, define monitoring and outcome measures, keep failures visible and revisit the question when the operating context changes.

Research principles

Traceability

Connect findings to the source, method, version and parameters.

Reproducibility

Record enough context to inspect or repeat the work.

Validity

Check that the method and measurements answer the question.

Uncertainty

State assumptions, missing evidence and limits of interpretation.

Accountability

Identify maintainers, reviewers and decision-makers.