The uncertain element 193
information and struggled to make sense of the complex data relationships
that these policies create.
Bargaining data
The financial incentives and rewards incorporated into behaviour-based
life insurance were in principle attractive to the customers. They compared
the behaviour-based instrument to car insurance products that reward
accident-free policyholders with bonuses. The smart policy was seen as a
similar mechanism that compensates people for staying healthy. Most of
the customers that participated in the focus groups considered these reward
structures to be fair. This is in part because the companies do not, at least
openly, punish unhealthy or inactive policyholders. Instead, all customers
retain their basic level of coverage (or premiums) and can gain bonuses (or
discounts).
However, due to their experiences with the tracking devices, some customers doubted whether the self-tracked data was reliable enough for assessing
activity levels and determining rewards. The inaccuracies and deficiencies
of such data are widely known (Gorm & Shklovski 2019; Pink et al. 2018),
and our informants also reflected on the devices’ inability to measure their
activities correctly; the data did not resemble their ‘real selves’ (Lupton
2020). Thus, even though people did not oppose the policies’ rewarding
structures per se, they had concerns with the trustworthiness of the behavioural data. Two of Company Y’s customers, Teemu, an IT professional in
his late 30s and Anne, a sales manager in her 40s expressed their concerns
as follows:
TEEMU: But how they are going to measure it [health]; that is the tricky
question. What data is it based on?
ANNE: Yeah, that should truly be something trustworthy. It cannot be
merely the device: it’s not enough.
TEEMU: Yes, it can’t remain open to interpretation.
Unlike car insurance, where eligibility for bonuses is checked annually, in
smart insurance the idea is that policyholders’ risk scores could be assessed
and determined based on real-time data (Meyers & Van Hoyweghen 2020;
Zuboff 2019). However, at least in our case study, this idea appears to be
unfeasible in life insurance due to both consumer objections and technological and regulatory limitations (Tanninen 2020; Tanninen et al. 2020).
Many of our informants recognised that the usefulness of behavioural data
stems from longer time series such as monthly averages. This was also the
approach in Company Z’s policy, which rewarded its customers based on
their average score over a period of several months. As the final estimation
was based on this longer time frame, policyholders appeared more accepting of small inaccuracies in their data.
information and struggled to make sense of the complex data relationships
that these policies create.
Bargaining data
The financial incentives and rewards incorporated into behaviour-based
life insurance were in principle attractive to the customers. They compared
the behaviour-based instrument to car insurance products that reward
accident-free policyholders with bonuses. The smart policy was seen as a
similar mechanism that compensates people for staying healthy. Most of
the customers that participated in the focus groups considered these reward
structures to be fair. This is in part because the companies do not, at least
openly, punish unhealthy or inactive policyholders. Instead, all customers
retain their basic level of coverage (or premiums) and can gain bonuses (or
discounts).
However, due to their experiences with the tracking devices, some customers doubted whether the self-tracked data was reliable enough for assessing
activity levels and determining rewards. The inaccuracies and deficiencies
of such data are widely known (Gorm & Shklovski 2019; Pink et al. 2018),
and our informants also reflected on the devices’ inability to measure their
activities correctly; the data did not resemble their ‘real selves’ (Lupton
2020). Thus, even though people did not oppose the policies’ rewarding
structures per se, they had concerns with the trustworthiness of the behavioural data. Two of Company Y’s customers, Teemu, an IT professional in
his late 30s and Anne, a sales manager in her 40s expressed their concerns
as follows:
TEEMU: But how they are going to measure it [health]; that is the tricky
question. What data is it based on?
ANNE: Yeah, that should truly be something trustworthy. It cannot be
merely the device: it’s not enough.
TEEMU: Yes, it can’t remain open to interpretation.
Unlike car insurance, where eligibility for bonuses is checked annually, in
smart insurance the idea is that policyholders’ risk scores could be assessed
and determined based on real-time data (Meyers & Van Hoyweghen 2020;
Zuboff 2019). However, at least in our case study, this idea appears to be
unfeasible in life insurance due to both consumer objections and technological and regulatory limitations (Tanninen 2020; Tanninen et al. 2020).
Many of our informants recognised that the usefulness of behavioural data
stems from longer time series such as monthly averages. This was also the
approach in Company Z’s policy, which rewarded its customers based on
their average score over a period of several months. As the final estimation
was based on this longer time frame, policyholders appeared more accepting of small inaccuracies in their data.