Approach
Evidence first, then judgment.
Good decisions rarely need complicated models. They need the right question, evidence you can check, an honest statement of uncertainty and a way to see whether the decision worked.
Step 1
Frame the decision
We agree the decision to be made, the realistic options, the time available and what evidence would change your mind.
Step 2
Gather and grade
We combine your data with primary sources and market signals, and grade each piece of evidence for reliability.
Step 3
Analyse and test
We use methods that can be checked, separate correlation from cause, and run small tests where data is thin.
Step 4
Decide and measure
A clear recommendation with its confidence, then a measure of what changed after you acted.
Evidence grades
How sure are we?
Every finding in our work and research carries one of four grades. The grade describes the evidence, not how much we like the conclusion.
| Grade | Meaning | Typical source |
|---|---|---|
| A | Measured. Directly observed in reliable data, with the method stated. | Sales, cost and advertising data; controlled tests |
| B | Documented. Stated by a primary source. | Regulations, official statistics, platform documentation |
| C | Reasoned. Inferred from evidence, with assumptions stated. | Models, comparisons, analogies |
| D | Practitioner view. Experienced judgment, labelled as opinion. | Network specialists, expert interviews |
Principles
What we hold ourselves to.
- State how sure we are, and why.
- Prefer simple methods that someone else can check.
- Separate what moved together from what caused what.
- Recommend doing nothing when the evidence supports it.
- Measure after the decision, not only before it.
Use of AI
Faster analysis, checked by people.
We use AI tools to gather, structure and analyse information faster. An analyst checks every output, sources are cited, and client data is handled under confidentiality terms and kept out of model training.
Start a conversation