The proliferation of mobile devices tracking how people move has underwriters eager to tap this data for risk assessment.
Wearables data has the potential to accelerate applications and inspire new products, but a prudent approach combined with the right partners and robust testing is key to making the most of alternative data.
We can all envision how a person’s exercise routine that they track in real time on their phone, watch or wristband could offer important insights into understanding their overall health status.
It may stand to reason that somebody who works up a regular sweat is likely to be on the healthier end of the spectrum and may represent a sound insurance risk.
But there’s a catch: physical activity data collected on wearable devices may not align squarely with the information we’ve traditionally collected to help our industry make good life and health insurance underwriting decisions.
Consequently, we need to understand the significance of this new data, interpret it with care, and deploy it wisely to maintain the integrity of our underwriting.
Ultimately, when integrating alternative data into our underwriting in pursuit of goals like reducing sales friction or trimming onboarding costs, we should simultaneously be aiming to preserve or improve the quality of our risk assessments, while avoiding costly errors.
A key message of our report “Getting practical with wearables” is that when physical activity data is used to augment an existing underwriting journey or to underpin ongoing dynamic underwriting, it can boost accuracy of forecasting mortality rates.
Conversely, outcomes may turn out to be less predictable should alternative data be used to replace or substitute existing underwriting information without fully understanding the extent of what is being omitted.
Know the limits
For a real-world example, let’s examine one kind of physical activity data – scientists call it “Metabolic Equivalent of Task,” or METs – collected in studies assessing the health and nutritional status of US residents.
When you examine the numbers, one key observation is the low correlation between measures of physical activity and whether those studied also suffer from a disease.
If an insurer were to base underwriting too narrowly on such METs data collected via alternative sources or use it to simply replace information from more traditional underwriting sources, it could result in people with diseases being added into the standard underwriting pool without being priced correctly.
This illustrates why reliance on alternative data without carefully considering its limitations can lead to adverse underwriting consequences.
Some InsurTech companies promote a view that using METs alone as a replacement may be as predictive as much of the data we’ve been collecting for years via traditional underwriting.
They maintain they can create a portfolio that’s 95% as predictive as existing underwriting using only physical activity data, age, gender, and just a few other risk factors.
While this sounds impressive, caution is warranted: Even if the frequency of inaccuracy introduced into the risk assessment process seems low, the severity of incorrectly priced risks can be significant.
On the other hand, METs have been shown to be an excellent fit for dynamic underwriting.
The data is widely available and easy to access.
Over the medium to long-term, the predictive power of tracking METs makes it easier to identify, price for, and drive healthy behaviour change which can help to reduce a customer’s risk or prevent their risk worsening.
Power of partnerships
Clearly, much depends on how alternative data is deployed.
The long tradition of insurers collecting information from people seeking to protect their families from financial hardship has served consumers well for more than two centuries.
It’s ensured protection products remain sustainable and appropriately priced for the risks we’re taking on.
Of course, the sophisticated data we collect today, often via medical examinations, can add time and cost, and can lead to some applicants abandoning the process prematurely.
The hundreds of millions of mobile devices now telling us how far we walk, ride our bicycles, swim, or run – and what our hearts are doing along the way – may help improve this experience for prospective customers, but we need to do this correctly.
Pairing a lucid understanding of new data with a sharp focus on an insurer’s individual goals is essential.
Whether the overall target is to improve risk stratification, expand dynamic underwriting, or personalise the underwriting process, alternative data can be helpful when deployed thoughtfully, with a firm grasp of its limits and possibilities.
We also recommend creating a robust test environment to pilot or experiment with different designs and different types of alternative data.
Test settings also make it possible to develop parameters tailored to clients’ individual risk appetites and the need to minimize anti-selection.