Which decision needs evidence?
For example, understanding a change, comparing periods or investigating factors related to an outcome.
Data analysis · Modelling
Find what the data says and what to investigate next.
We analyse and model data to uncover patterns, relationships and correlations that help explain a question and inform better decisions.
01Clearer questions
02Relationships examined
03Better-informed decisions
The opportunity
Understand before assuming.
We start with the business question and the available data. We explore its quality, structure and context to distinguish useful signals, limitations and new questions.
01 Findings in context
Exploratory analysis
We review distributions, trends, segments, outliers and data quality to understand what the data contains and which questions it can answer.
02 Findings in context
Relationships and correlations
We examine how variables move together and look for relevant associations. A correlation can inform a hypothesis, but does not by itself demonstrate causation.
03 Findings in context
Data modelling
We define how to represent variables, metrics and relationships for analysis; where it makes sense, we add statistical models and evaluate their limits.
How we work with data
From scattered sources to findings in context.
We review sources, explore patterns and model relationships to explain what the data shows and what to investigate next.
Sources
Quality, structure and available data
Exploration
Distributions, relationships and outliers
Modelling
Hypotheses, assumptions and limits
Findings
Evidence to inform a decision
Conceptual schematic with simulated data. Associations are interpreted in context and do not demonstrate causation.
How we move forward
From a question to evidence for the next step.
Analysis depends on the data, its context and the question. We agree on those limits and explain clearly what the results do and do not show.
- 01
Frame
We define what needs to be understood, for whom and which decision the answer could change.
- 02
Explore
We review sources, quality, variables and the relationships that appear in the data.
- 03
Model and explain
We examine hypotheses, evaluate models when appropriate and present findings, assumptions and next steps.
A practical question
Do your reports mean the same thing by an indicator?
Two figures can differ because of their periods, sources or definitions. Before modelling, we connect each metric to its meaning so comparisons make sense.
What your team can review
Deliverables are defined by the agreed scope.
- 01
An understanding of the sources
Relevant data, definitions, quality issues and limitations that may affect the answer.
- 02
Analysis and models for review
Variables, transformations, assumptions and results documented to the agreed scope, so the team can follow the reasoning.
- 03
Findings that guide the next step
An explanation of patterns, uncertainties and further questions connected to the decision behind the work.
When it makes sense
When data is available but has more to teach.
This work can help teams understand their results before automating, scaling or investing in another solution.
- 01
There is plenty of data, but it is hard to turn it into a useful answer
- 02
Different indicators seem to tell different stories
- 03
The team wants to identify factors associated with an outcome
- 04
A hypothesis needs to be examined before making a decision
Before we talk
A useful starting point.
We can start with a question and a map of available information. It does not all need to be organised beforehand.
Where is the information?
List systems, spreadsheets or reports and who can explain their fields. No need to send them in the first conversation.
Which comparisons raise questions?
Bring indicator definitions and examples of differences across teams, periods or sources.
Let's talk
Let's explore what your data can answer.
Tell us about the question, the data available and what would be useful to learn before choosing the next step.
Book a diagnostic session directly with the partners