How we work — Quintavera

The workflow

01 Explore the problem
We define the central question. It is often not the one the organisation started with, and getting it wrong makes everything downstream worthless.
02 Scope the data
We establish what your data can and cannot support before anyone relies on it. Knowing the limits is part of the answer.
03 Quintaveran analysis
Our own analytical method, and the stage where the real work happens. Set out in full below.
04 Synthesise solutions
We resolve the modelling into a small number of real options, and identify the determinants driving each one.
05 Validate and activate
We test the recommended decision against challenge, then set out what putting it into effect actually requires.

Stage 03 of the workflow, in detail

Quintaveran analysis^™

Quintaveran analysis applies advanced mathematics to find and model the causal and interdependent relationships between otherwise disparate data points.

It goes further than describing those relationships. It predicts the geometric patterns underlying business information, the shape a situation is taking, to inform a decision rather than merely illustrate one.

Structural equation modelling, network analysis and system dynamics provide complementary methods for understanding what is happening within a complex system, investigating why it is happening, and modelling the potential impact of change.

Three analyses then ensure the validity, robustness and interpretability of the evidence, providing a sound basis for explaining current outcomes and evaluating the likely consequences of strategic interventions.

Causal inference modelling

Predicting business outcomes

Correlation is easy to find and rarely useful on its own. Causal inference separates what is genuinely driving an outcome from what merely moves alongside it. That is the difference between predicting what a change will cause and observing what it has accompanied.
Imputation analyses

Finding missing data

Real business data has holes in it. Records are incomplete, systems disagree, and the most important variable is often the one nobody captured. Imputation reconstructs what is missing from the structure of what remains, so that a gap in the record does not become a gap in the decision.
Determining relational significance

Revealing hidden truths

Not every relationship in a data set matters, and the ones that matter most are frequently not the obvious ones. We establish which connections carry real weight, including those crossing the boundaries between functions, which is precisely where nobody thinks to look.