At its core, motor insurance pricing rests on three pillars:
- The person – who is driving, and their history.
- The vehicle – what they are driving, and the cost of repair or replacement.
- The place – where they live and drive, and the conditions they are exposed to.
The first two pillars are well understood and deeply embedded in insurer models. The third 'place' has always mattered too, but the data now available gives us the opportunity to understand it in far greater depth. Geography is more than just a location marker. It is a lens into behaviour, environment, and exposure.
From Postcode to Prospect
Postcodes have been a foundation of UK insurance pricing for decades. Every insurer already factors them into rating tables. But beyond their use in pricing, postcodes can also help answer a bigger strategic question: where else should we grow?
If you know which parts of your book perform strongly, the natural question is: where else looks like this? These are your lookalike geographies, postcodes that share the same environmental and socio-economic profile as your best customers, even if they’re outside your current quoting footprint.
By identifying and targeting these areas, insurers can expand their footprint with greater precision, backed by evidence rather than assumption.
Why Lookalike Geographies Matter
The UK motor market is fiercely competitive. Switching rates are high, margins are thin, and new business growth is harder than ever to achieve. Traditional levers, adjusting rating curves, tweaking acquisition spend, or opening new channels, only move the needle so far.
Lookalike geographies offer another lever. By focusing on postcodes that mirror the conditions of your strongest-performing segments, insurers can:
- Target market entry with more confidence.
- Guide broker development towards high-potential regions.
- Support pricing strategy with clear evidence of where risk is genuinely lower or higher.
And because geography is rooted in real-world factors, the logic is easier to explain across the organisation, from pricing to distribution, and from underwriting to compliance.
Apollo + Atlas: A Combined Lens
At Consumer Intelligence, we combine two complementary capabilities to unlock this opportunity:
- Apollo – our machine learning pricing engine, built with explainable AI tools to show not only how the market prices individual risk factors, but also how those factors interact. Apollo helps quantify the strength of relationships between features, for example, how geolocation signals combine with driver characteristics or vehicle attributes to shape competitive pricing.
- Atlas – our postcode enrichment model, which describes the environment around each postcode using over 200 engineered features.
Together, they provide insurers with a dual view:
- Apollo shows how the market prices geography and how it interacts with other factors.
- Atlas explains why geography matters by revealing the underlying signals.
By linking them, insurers can not only benchmark their own approach against the market but also identify new postcodes with the same profile as their best customers. That is the essence of lookalike geographies.
Building a Richer Picture of Risk
What does this look like in practice?
When building a richer picture of risk, Atlas doesn’t rely on single signals. It layers together multiple dimensions of the environment:
- Who lives in the area – age distribution, household composition, education, employment.
- Deprivation indices – not just an overall score, but specific components like long-term unemployment or transport availability.
- Accessibility of services – proximity to GP surgeries, hospitals, or key amenities, which influence commuting and driving patterns.
- Affluence and socio-economic status – indicators of financial stability and household resilience.
- Housing and property mix – household size, property type, council tax band, and property values.
- Transport and road environment – vehicle ownership, commuting modes, road network quality, and historic collision data.
Taken together, these features create a multi-dimensional fingerprint of place.
Because Atlas risk scores correlate strongly and linearly with observed loss experience, this enrichment isn’t just descriptive, it is predictive. That means lookalike geographies don’t just look like your best customers on paper; they behave like them in practice.
From Correlation to Confidence
The strength of Atlas lies in the fact that its scores are not abstract constructs. They have been shown to align closely with real-world loss outcomes. Higher Atlas risk scores translate into higher observed loss ratios, and lower scores map to lower losses.
This predictive power turns enrichment into action. It allows insurers to expand into new areas knowing that the profile they are targeting has already been demonstrated to align with better risk.
Rigour Behind the Model
Atlas wasn’t built by simply throwing data into a model. It was carefully engineered:
- Over 200 features from trusted, official sources (ONS, DfT, Met Office, Land Registry, DVLA, police data).
- Features grouped into domains like commuting, housing, crime, weather, and affluence.
- Translated into 50 calibrated risk buckets, capturing gradations of risk across the UK postcode system.
Crucially, we used explainable AI tools and cluster analysis to interrogate the relationships within the data. This helped us understand why certain features are predictive, not just that they are.
For example, when examining categories of crime, it became clear that some indicators of social environment are more predictive of risk than others. It’s not about simple correlations, it’s about building a transparent story of why geography matters. That transparency is essential for both pricing and governance.
A Framework for Growth
With Atlas and Apollo, insurers gain more than two models. They gain a framework for growth through geography:
- Define the profile of customers or areas that deliver stronger outcomes.
- Benchmark how the market prices those same features, including the way geography interacts with other risk factors.
- Map postcodes that share similar profiles, your lookalike geographies.
- Expand your footprint with confidence, supported by evidence that is both predictive and explainable.
This isn’t about replacing existing approaches. It’s about enriching them with new clarity, and extending them into a new dimension of growth.
From Footprint to Future
Motor pricing will always be about the person, the vehicle, and the place. With Atlas and Apollo, insurers now have the tools to understand “place” in greater depth than ever before and to use that understanding not just for pricing, but for growth.
Lookalike geographies turn postcodes into prospects. And with risk scores that correlate to loss, they turn prospects into profitable customers.
From footprint to future, this is the next step in making insurance pricing not only more predictive, but more commercially powerful.
Submit a comment