Pricing motor insurance in the age of data

Written by Benedict Harrison & Joe Hartshorn

Published on

How a data-driven pricing model can enable fairer insurance premiums for policyholders.

Image of chart showing data

Over the last 20-30 years, there has been little change to the way that motor insurers price their products. Industry standard pricing methods rely on grouping drivers based on a limited number of available factors, which include a person’s demographic profile, their vehicle information and driving history. If two drivers share similar information, they are judged to have a similar risk profile, which is reflected in their total insurance premium and so they will pay a similar amount for their cover.

Yet, despite how prevalent such practices are in insurance pricing, they are not without their shortcomings. Firstly, and perhaps most significantly, these factors do not actually measure a person’s driving skill and are only distant proxies for a person’s risk. The pricing groups that insurers assign drivers to may be suboptimal for actually predicting the risk of a particular driver. As a result, we are left with an outcome that seems inherently unfair, where drivers pay for someone else’s risk.

In addition to inaccurate risk profiling, some pricing factors, such as someone’s history of previous claims and convictions, may over-penalise the unfortunate few. A single claim is more likely an indication of an unfortunate set of circumstances than a reflection of a person’s driving ability. As a result, it may not be the fairest measure for insurers to use when calculating premiums.

At Zego, fairness is a cornerstone of how we build and price our insurance products. Our customers should be able to access cover that suits their particular needs, for which they pay a fair price, which accurately reflects their risk. So how can we improve upon these traditional pricing methods to profile a driver’s risk more accurately?

The solution for the insurance pricing problem lies in the ability to go beyond the standard insurance factors and measure traits and behaviours that are more meaningful for how they inform us about a driver’s risk. This is where the data becomes the centerpiece.

How can data help us do things differently?

Image of a mobile phone

Driving data can be collected from many sources, from vehicles and smartphones to roadside traffic, weather and pollution sensors, and made available through a variety of means, such as proprietary software solutions and data feeds, public-private partnerships as well as open data initiatives. Though, the data collected today is created in a menagerie of forms and more importantly, it is available almost instantly and in volumes that were previously unimaginable.

At Zego, our team is dedicated to leveraging this data to get a better understanding of the inherent risk profile and driving characteristics of our customers. We don’t have to settle for distant proxies and vague correlations.

Using behavioural, contextual and environmental information and data-driven models we aim to get closer to the real causes of risk. Moreover, such data-driven approaches minimise the reliance on outdated assumptions and ad hoc adjustments, all to make sure that we recognise the good drivers and offer the fairest insurance to all customers.

But with massive quantities of data comes the necessity for scalable systems that can help us unlock the most value. At Zego we use the latest and greatest technology, such as Python, Docker and Kubenetes, to create a stack that enables us to move from ideation and research to deployment (making a real impact!) in lightning speed. Moreover, we believe in diversity and in the richness of ideas it brings, hence our analysts and data scientists have the freedom and flexibility to explore and propose new technical solutions as well as product features.

Not only is our aim for smarter, fairer pricing about getting the most from the innovative and novel data that we’re using, it’s also about making sure that the decisions we make are fair and right. That’s why it’s important that our technical systems and pricing models are built with transparency, interpretability and robustness in mind,

Why does this matter?

Once an insurer is able to implement pricing that more accurately reflects a driver’s risk, they can provide customers with a fairer price for their cover. When we talk about a fair price, we do not necessarily mean the cheapest, but a premium which accurately reflects a driver’s actual risk. When a driver’s premium is calculated based on their profile and is tailored to their information as an individual, they are no longer paying for someone else’s risk.

But the benefits of this data driven approach don’t stop at fairer pricing. At Zego, we’re already starting to share actionable insights with our drivers to help them understand how they can improve their risk and save on the cost of their premium. By sharing this information with our customers, we can empower them with control of their insurance, helping them to improve their safety when they’re out on the road, creating a safer environment for all road users.

At Zego, we will continue to implement scalable pricing solutions, achieved through a combination of prior research with a novel data driven approach, all with the goal of fairer insurance in mind.