One of the more intriguing aspects of Big Data for the industry is that a good deal of data that seems unrelated to insurance is actually proving useful in predictive analyses and other areas. Insurance companies have always been data-driven, but Big Data has opened up many new possibilities.
To put things in perspective, Big Data is not a concept limited only to the insurance industry. Big Data has become a driving force in fields as varied as politics, social media and science. The good news for insurers is that however much data they may be juggling and trying to coax into a useful format, it is only a fraction of the data regularly gathered by an astronomy project, like the Sloan Digital Sky Survey. With the reassurance at least that they don’t have to cope with, literally, an entire universe of data, how can insurers move forward and make the best use of their own Big Data?
Underwriting and predicting customer behaviour
Insurance companies now have the ability to correlate an enormous amount of information about potential and current customers, and to make predictions based on that information. For example, a study by direct reinsurer Gen Re revealed that with a life insurance product that was sold through a bank, there were links between customer banking behaviour and persistency.
Combining wellness and loyalty programmes can give insurers access to an enormous amount of data that can be used to make predictions about customers, as well as encouraging healthy lifestyles that reduce the incidence of claims. Insurers have found that they are also able to assess how factors like socioeconomics, based on the neighbourhoods that customers reside in, can predict mortality and persistency. Specific professions can also be better linked to disability claims and mortality. These are just a few examples of the many ways in which insurers have access to more data than ever before, and data that will allow them to shape their products and target their desired customers more effectively.
Coordinating customer behaviour on a website with call centre logs can provide valuable information about what online interaction may have prompted the call. Furthermore, rich data regarding call centre interaction, customers’ positive or negative experiences, and re-enrolment numbers linked to these interactions, can be examined. As with the underwriting information obtained above, one advantage of this data is that gathering it requires no action on the part of the customer.
Big Data permits fraud investigation to switch from a focus on claims to a focus on individuals. Cohort analysis and behaviour among networks of people, as well as the types of claims submitted by individuals and behaviour of beneficiaries, can all be examined and analysed now. Predictive analysis can draw on more tools including modelling, database searching and test mining to catch fraudulent claims.
More information can be captured and linked to claims including images, video and third-party notations. For fast track claims, Big Data can produce analytics that help to mitigate a tendency towards overpaying. Even a tendency towards litigation can be assessed. Data from social media can also be harvested to support claims adjustors. In the summer of 2013, a company called Bright Planet analysed all the tweets about flooding in Toronto and mapped their locations for an insurance company. The adjustors could therefore better assess whether a claim might need further investigation and could also predict where the largest numbers of claims would come from.
Big Data and technology
In the past, a great deal of the information obtained by underwriters or gathered about customers might have been spread across a variety of databases and spreadsheets. The technology associated with Big Data means that information can now be collated in one place. The process of switching over to newer systems will be a slow one in most cases, and for now, traditional systems for managing data still have a place in the industry. In fact, these relational databases still handle certain types of data better than the newer systems and insurance companies should look to take the best from both approaches for now.