Admiral’s plan to use information about customers harvested from their Facebook posts in order to price car insurance seems to hit a road block almost immediately when Facebook announced that it would ban the insurer from using its data.
Clearly, Facebook has not suddenly turned into a champion of privacy. Information about its users is Facebook’s greatest asset. Maybe Facebook expects a cut where others exploit it? Moreover, Facebook may be concerned about changes in the behaviour of its users. In the longer run, anything that might reduce the willingness of customers to disclose information about themselves could be a threat to the platform. Would social media be as attractive if there is self-censorship of youngsters’ thrilling exploits to get cheaper insurance?
Reports suggest that Admiral pitched the scheme as voluntary, only leading to discounts and not higher prices. No doubt schemes such as this (and car telematics) may expand the market, offering young drivers without a driving history insurance at more reasonable rates. However, the overall picture is mixed. Systematically picking good risks out of the general pool of drivers means that the average cost of insuring the rest will be higher. Premiums will be, too. Big data may produce efficiency gains through better risk segmentation, but inevitably this creates both winners and losers.
Micro-segmentation and the potential that some high risk customers may be left out has been identified as a potential concern with the use of big data by the FCA. As research we have undertaken for the CMA suggests, these concerns are well justified. Moreover, it’s wrong to pretend there are only winners. Drivers who don’t opt in, or those who don’t have a Facebook account could end up paying more – not because they are higher risk drivers, but because the insurer has less information about them. A premiums gap might create strong pressure to bare all and handover one’s data to save money. Such pressure to disclose information is potentially worrying, in particular when there are good reasons to believe that customers may not really be able to put the right value on their privacy.