How do consumers balance the data tracking inherent in behavioural life and health policies, with concerns about autonomy? Recent research has looked at what consumers say and do in relation to devices such as smart watches. I explore what this tells insurers about the design of such policies.
We know that autonomy is important to consumers, even if they rarely use the word itself (more here). And we know that behaviour based policies are being launched now on a regular basis. This is mainly in markets such as life, health and motor, and in both group and personal lines portfolios.
A recent survey by Policygenius in the US found that consumers had strong feelings about data tracking apps (more here). Over two thirds of Americans would not install an app that collects driving behaviour or location data for any insurance discount amount, up from 58% the previous year. Yet this also means that one third are willing to engage with them.
If insurers are to move opinion on such apps and devices, understanding what consumers using them like and dislike about the experience is pretty key. They could of course just rely on sentiments like ‘that is the way things are going’ and ‘we might as well give it a try’. However, that is unlikely to deliver the long term traction that insurers need in order for them to grow their portfolios of behaviour based policies. Better to design a policy to deliver sustained engagement.
Recent research in Finland draws upon detailed focus group work with a cross section of policyholders holding behavioural life policies with two Finnish insurers. I met the authors at an EU workshop in Bologna last year and had the opportunity to discuss their (at that time preliminary) findings with them. Their final paper is worth reading if you work in policy design, product governance, data ethics or underwriting. What I’ll do here is highlight their main findings.
Consumers Voice Concerns
An important general finding is that “the new data relations created by behavioural policies raise doubts and uncertainties that pushes people to engage with questions related to autonomy.” Note that autonomy was not something introduced into the focus group discussions by the researchers. They found that complex deliberations around free will "...surfaced organically in the conversations, emphasising that autonomy was viewed as under threat from the intrusive nature of new insurance forms.”
They go on: “…people mobilise notions of individual choice and self-reliance to set boundaries for the insurance policies’ reach.” And people do this by framing themselves as "freely choosing individuals, capable of managing their lives and determining the limits of their data relationships."
That element of choice is key to the acceptability of being monitored. If choice was not built into the policy, then they would opt for a different policy. It stops the tracking features from becoming overly controlling and gives users a sense of self-determination.
Participants in the focus groups wanted these behavioural policies to support only positive rewarding. While the possibility to opt into a rewarding scheme is appealing and justifiable, punishing people is seen as unattractive and borderline immoral.
In terms of the co-existence of the user and the device which behavioural policies encourage, what was found to be important for this kind of co-existence to work is the costumer’s feeling that she is benefitting from the policy and the device’s output – the technologies should support her in achieving her goals.
In ethical terms, the technologies actively support the user’s sense of self-determination. The key word here is ‘support’. The device has to help the user to do what they already want to do, complementing and, within limits, reinforcing their existing intentions in relation to health.
Sometimes, when the device has been regularly supporting what the user was already experiencing, they may pay less attention to it, perhaps even use it less often. So there’s a policy design question here. How does the insurer differentiate between good ‘lack of use’ (because the device is no longer adding any value) and bad ‘lack of use’ (because the user can’t be bothered). The two need to be differentiated.
Here we see another key word coming out of this research: alignment. When the policy’s nudging elements align with the user’s interests, the value both sides gain from this is positive. Alignment drives engagement.
Generally, the interviewees in the focus groups “make no connection between their everyday tracking activities and the insurance infrastructure, even in case where the tools are part of the insurance policy and provided by the company.”
This could be seen as a positive – users just get on with tracking and forget about the insurance element. However, it is probably more of a negative, indicative of a lack of engagement. Some interviewees reported that...
"...they had stopped tracking because of lacklustre experiences with behavioural policies. They were not sure whether their tracking practices really mattered, as they did not receive substantial feedback from the insurance company; what little communication there was felt deeply impersonal. Because they did not receive reminders when they stopped using the devices, some began to think that the tracking features were merely a marketing gimmick without any real effect."
As the researchers comment, indifference to the behavioural policy may be just as damaging to the insurer’s interests as any irritation and frustration that might arise with the app or device.
Not surprisingly, interviewees at the focus groups were vocal in their feelings when apps and devices failed to align suggestions with their needs:
"It tries to force its internal logic on people's lives and thus fails to recognize the complexity and contingency of human experience and what really matters at a particular point in time."
The nudges can sometimes feel intrusive and out of sync. And this in turn undermines the user’s sense of self-determination. The device seeks to exert an arbitrary control, by for example, registering knitting and crocheting, but not cycling or walking with a buggy. Their sense of autonomy is undermined by a sense of intrusive and annoying control. The device nudges them to be active at the end of a day taken up with looking after toddlers. The alignment is broken.
For the researchers, their key finding was that autonomy needed to be seen not only through an individualistic lens, but through a relational one as well:
"...rather than a struggle between pure individual autonomy and total submission to the machine, what is at stake is a more contextual zone of contrasting desires and situational negotiations and practices."
The narrative of individualism behind much of the promotion of behavioural insurance, and that of personalisation more broadly, is too simplistic. It needs to be broadened to recognise the different aspects of autonomy. There is certainly an individual aspect to autonomy, but there is also a relational aspect too. For people in everyday life, this relational aspect is experienced in two ways. Firstly, there is relational autonomy in terms of others, such as communities, institutions and traditions (more here). And secondly, there is relational autonomy in terms of the technologies and infrastructures that deliver behavioural insurance itself.
One way of looking at this is to think of behavioural insurance as something ‘done with’ the consumer, rather than something ‘done by’ the insurer to the consumer. The challenge for the sector is that it has been built predominantly upon the latter. A significant shift in culture is needed to ensure a ‘done with’ mindset permeates behavioural insurance. In that sense, autonomy on both sides of that relationship needs to be weighed up so that the alignment behind a ‘done with’ approach is secured.
Time for insurers to put on their learning caps.
The researchers were Maiju Tanninen, Turo-Kimmo Lehtonen and Minna Ruckenstein, and their paper is "Trouble with autonomy in behavioural insurance" published in June 2022 in the British Journal of Sociology.