The modelling behind every number

What we deliver

• Statistical analysis and survey design
• Econometric modelling and forecasting
• Cost-benefit and sensitivity analysis
• Financial modelling and scenario planning
• Applied machine learning where it genuinely adds insight
• Clear, plain-English reporting — we explain the method as well as the result

Methods and standards

• Regression modelling — linear, logistic and panel/fixed-effects specifications
• Quasi-experimental designs — difference-in-differences, propensity score matching, regression discontinuity
• Time series analysis and forecasting, with prediction intervals rather than single numbers
• Survey design, sampling and weighting, including non-response adjustment
• Cost-benefit, sensitivity and Monte Carlo analysis
• Financial and scenario modelling with a documented assumption register
• Index construction and segmentation for targeting and prioritisation
• Applied machine learning where it genuinely beats a simpler method — and we will tell you when it does not
• Official data sources — ONS, NOMIS, Census, Indices of Multiple Deprivation, BRES and sector equivalents
• Model quality assurance following Aqua Book principles: documented, version-controlled, reproducible, peer-reviewed

We also review and challenge models built by others. If a number will be defended in public, it is worth having someone try to break it first.

Who it is for

Public bodies and research partners · Consultancies needing an analytical partner · Businesses with data they are not yet using · Organisations that need a model reviewed or challenged.

How it works

1. Define the question and the decision it supports.
2. Audit the data you have and identify gaps.
3. Build, test and document the model.
4. Present results with the assumptions and their sensitivities made explicit.