I think there will be two primary models.
1. Self Directed x Commodity Product
> Better version of today’s aggregation services
> Illustration system to show how simple products can meet insurance needs
> Potential buyer given the option to share more information to get a discount (e.g. share health records stored on a cloud to get underwriting discount)
> Insurers compete on pricing (e.g. identify potential buyers with lower lapse risk and offer lower premium)
> Insurers that are currently investing in bancassurance models are likely to pursue this (i.e. whatever algorithm they use to identify good risks in the bancassurance channel can be applied to the digital channel)
2. Assisted Advice x Customized Product
> Customized product makes sense because different people have somewhat different combinations of needs and combo products can provide better value (e.g. buying Manulife Synergy can result in 30% savings relative to separate term, disability and critical illness products)
> More importantly, different people have different preferences (e.g. some people would not buy insurance without return of premium rider because they want to get something back even if the insurable event does not occur)
> A good adviser does not show the client twenty different combinations and cause confusion; they can quickly figure out what the client is likely to be interested in; a good assisted advice digital channel should be able to do the same thing for the middle income market (the affluent market will probably always want multi channel service)
> It’s not clear to me who will pursue this (e.g. insurers playing in the affluent market can do product customization but may not understand the needs of the middle income market)
> This is what I’m really interested in
Stephen took the machine learning course on Coursera and really enjoyed it, so I signed up for the 10 week session starting on April 22. Looking forward to doing a bit of programming again!