Customer engagement and underwriting through predictive analytics
Article information and share options
EVO is a data science company working with life insurers, pharmaceutical providers and senior care facilities. At the highest level, the company takes data off a variety of wearables.
The data is brought into a cloud-based environment and analysed for three principal purposes: scoring (ie risk); health guidance; and critical health issues needing to be urgently addressed. In terms of insurance, the use of data provides two key returns on investment: customer engagement and creating sticky policies; and risk assessment. These two parts are interlinked. If you do not get the customer engagement part right, you will not have the data to work on for risk analysis.
The right mix of these two pieces is like bringing a cocktail together. The demographic information behind behavioural data is an important information overlay. Wearables with geolocation data provides time and space dimensions. Health metrics can be brought together with psychographic information. Questionnaires, as currently deployed in underwriting are an equally important ingredient.
In terms of customer engagement, the best way to achieve that has proven to be mobile. That is the touch point to most consumers in the modern age. From mobile, data can be brought into the cloud; where consumers expect something tangible in return, and in real or at least near-real time. In order to promote long term engagement, consumers have to be motivated over the longer term, and interested in their own personal wellness journey. How we relate to our health, of course, differs across individuals, and any platform must reflect those differences. EVO has identified around eight or nine clusters of core personality types and how they relate to their health. This helps provide the relevant information for an individual on the home page, to nudge them towards further engagement. It is not an easy trick to perform, as abandonment rates on devices and mobile apps is exceptionally high. Typically within 14-16 weeks consumers may abandon a wearable. EVO has maintained a relationship so that 75% of its user group is still engaged after 180 days.
In the early days of wearables, risk data was expected to be extremely rewarding. It has yet to quite develop that way. Actuaries are not certain of long term relationships between activity data and mortality. This is particularly because wearables are not always consistent in their output on items such as heart rate data. The data will become stronger in that respect; and it may be that even if devices record different data, the rate of change will become more important rather than the rate itself. Moreover, the use of longitudinal data will identify users at a population level. Data suggests smokers walk at a slower pace than non-smokers, for example. It may not be entirely accurate at an individual level, but has some validity over a larger population. Smokers also have a heightened resting heart rate while asleep. Cross those two points of data, and the analysis gains in accuracy.
Data showed that just looking at step data alone over a long enough period did point to being able to see some major health anomalies, including heart anomalies and sleep anomalies. However, the efficacy of these markers tended to fade over 12,000 steps in a day. Effectively you are getting a window into someone's health, which may point to anomalies.
Although we are still some way off replacing existing underwriting processes with data, it is a trend that will come. However, wearables are both significant in consumer engagement and identifying risks you may not see in your current processes. It may highlight to you times you want to apply the brakes in the underwriting process, rather than step on the gas. The increasing use of data will be an evolutionary process that will initially complement existing practices rather than replacing them. In terms of consumer engagement, making a small percentage point gain in retentions will offset the cost of investing in such data programmes
Summary of Patrick Bewley's presentation at the Centre's Health monitoring event in December 2017. Patrick is CEO of Big Cloud Analytics. Summary by Simon Woodward.