By Alan Osborne, Head of Product, Insurance, for LexisNexis Risk Solutions UK
So as the volume of consumer data gathered by the motor insurance sector grows, so does the challenge to achieve ‘one view’ of the customer when changes or variations in their profile history occur. This volume increase has led to disparate information and potentially multiple records held for one individual. Only once we bring together the myriad of data sources and cross-check, to provide a single reference point, can we begin to derive a full picture and therefore real value. Just in the past few weeks one insurer has launched a combined motor and household policy, underlining the value of creating a total customer view, bringing data from across all lines of business.
Consider that an insurer needs to make sure individual entities or individuals can be resolved to a single data key. That sounds like a simple thing. But can you confidentially or confidently identify individuals across your different databases? If the data is really old and addresses haven’t been updated for many years, it’s sometimes very difficult to match a John Smith from 2016 to a record from John Smith 2010. John Smith (or Jane Smith) may have moved addresses three times in that period. Incomplete records can also be caused by customers mis-keying or mis-stating information in their application.
Can you identify addresses correctly? Can you identity households? These elements will all result in better insights if insurers can correctly match and align the data points together across all their datasets. The insurer also needs to link the data sources together. How can they pull it together efficiently and accurately? We call this precision and recall: in other words, how precise are the matches, and how much can you bring back? Linking technology can resolve, match and manage this information. The ability to link and cluster data from multiple sources is key to leveraging authentication, verification and other identity management capabilities.
Contributory databases have been a driving force behind insurance innovation for a variety of markets since the 1980’s. However, in the UK, when you start applying linking technologies to a contributory database of Motor Policy History including NCD, fascinating risk insights emerge. For example, 13% of motorists on average over-state their NCD , generating a focus for investigation or risk-priced pricing adjustment. A similar number under-state their NCD, so we know there is a big role for cross-checking and cleaning data in pricing/underwriting but also in fraud and other areas of insurance.
So how exactly does linking technology work? It starts with assigning a unique identifier to a new piece of customer data that is not derived from personally identifiable information, like date of birth. Then records with intersecting data points are linked together and assigned the same unique identifier. The identity profiles should then be continuously updated to ingest new records. This dynamic process paints an extremely accurate picture that accounts for identity changes over time.
By linking customer, public and proprietary information in this way motor insurers are able to
shed light on suspicious entities or transactions, quickly hone in on irregular data points or outliers, identify collusion or undisclosed relationships, eliminate false positives, protect customers’ personally identifiable information and ultimately create a comprehensive view of an individual.
Coming back to the NCD database used by motor insurers, when you apply linking technology up to 6% more individuals are correctly matched to existing data within the database providing a full view on that consumer’s NCD history. There is also a 78% match rate between the existing NCD database and new customer data. The benefits are obvious - insurance providers get a better match rate that will help improve the experience for consumers, because they can quickly be tied to their own No Claims Discount history at point-of-quote and won’t be confused with another person or be provided an inaccurate quote due to accidental NCD misstates.
Linking technology can also support insurers in the fight against fraud by matching and linking together more customer records than ever. This could be due to people having used different names or addresses in order to avoid declaring their actual NCD entitlement. As we know, most insurance providers offer large discounts to new customers with a long NCD history, therefore it is really important that they can verify honest customers’ declarations at point-of-quote, and help insurers and brokers weed out those who could be fraudulent.
|
