Blog: Advanced analytics - music to insurers' ears

The Association of British Insurers estimates undetected fraud stands at £2.1bn a year. That adds about £50 annually to my - and your - insurance premiums.

However, as Bob Dylan once crooned: "The Times They Are A-Changin"; and we live in an increasingly digital age where that pile of data is getting beyond just ‘big'. According to Accenture, 44% of insurers estimate that the volume of data they are managing has grown by 50% or more over the last year. Insurers must find ways to analyse this data efficiently and accurately as part of everyday working practices. In fact, it's essential if they want to stay ahead of the fraudsters.

Nevertheless, our research has found that time-consuming, outmoded fraud detection techniques still plague the industry. Almost 20% of insurers state they do not use any technology to assist with fraud detection, relying instead on manual reviews of thousands of claims.

This leads to a detection problem. Companies that do not use automated fraud detection see significantly lower levels of detected fraud than their peers using advanced analytics.

With advanced data analytics, insurers now have the ability to more accurately assess risk, right down to the micro-level.

For example, insurance companies are now increasingly using telematics data. According to Berg Insight, there were almost five million such policies running in Europe at the end of 2014, and this is expected to reach 28 million by 2019.

Car insurers are using telematics data to assess each driver based on their driving behaviour. This includes distance travelled, but also hard braking, acceleration, and speed.

Some health insurers are using the data from wearable devices as an extra rating variable. In the US, Humana has partnered with Apple to get the health data collected by the tech firm's devices - which can lead to lower premiums for some of its policyholders.

Property insurance is turning to sensors to recognise potential claims and hopefully prevent losses. In Italy, BNP Paribas Cardif offers an insurance package that uses sensors to detect possible risks of fire, smoke and water, to better protect customers' homes.

It's not just accuracy of risk analysis that can be improved - there's also the matter of speed. Insurers need quick and accurate conclusions before they can draw the line under certain claims. Currently, they're getting answers in three to four months in some cases, rather than straight away.

With real-time fraud analytics, insurers can widen the scope of their investigations and automatically identify potential fraudsters at various points in the insurance chain. From the quote stage through to the application process, right down to the claims point, analytics can deliver immediate answers. As a result, insurers have more opportunities to shut down fraudulent activity earlier and reduce the cost burden for both themselves and their customers.

Insurance companies need to start investing more in analytics to crack down on fraud losses, make efficiency savings and, ultimately, be able to offer customers a lower premium. Some insurers have started digital transformation programmes that are fundamentally changing their business. Others, who haven't embraced analytics, will quickly find themselves falling behind the competition.

Since those insurers that have adopted analytics can now quickly identify potential fraud cases, they also know who the ‘good' customers are, so they can treat them with the sensitivity and care they deserve. As Hot Chocolate sang: "Every 1's a winner".