Each month, customers choose from 3 'targets' - the better your driving score, the cheaper your price.
Quote
30 day customers can choose from 3 prices based on traditional premium factors PLUS a variance which correlates to different driving score targets
Educate on driving score range required to not be 'cut'
Educate customers on how their premium can potentially decrease overtime if they improve their driving score (potentially graph projection or adjustment scale)
Payment
Customers pay the full amount upfront
We store the card details
Weekly Recap
Show progress of target (or baseline) vs actuals on a graph
Positive reinforcement for good driving → small weekly rewards
Coaching for bad driving
30 day Cashback + Adjustment
Customers are told if they have met, exceeded or did not meet target
For exceeding target: customers get cashback on the difference and they can choose the challenge for the next month (maybe get reward too)
For meeting target: no cashback, get reward
For not meeting target: next month's premium is increased (maybe also give a small reward e.g. 'We're here to help you improve your driving! Have a coffee on us, we believe in you')
Customers who got lower than minimum score risk being 'cut' each quarter
Renewal
Customers can choose if they are getting the balance as cashback or use it against as discount for renewal premium. The renewal premium value would be higher (eg. the balance on the account is £100, the cashback value is £100 OR the renewal discount is £120)
Assumptions
We can relatively confidently estimate risk based on driver score (the risk model already exits traditional + score + contextual)
Cost of acquisition for monthly vs yearly is the same (to be confirmed)
Self-selection of the bad drivers (they are not going to choose telematics products) - in order for this mechanism to work we need either:
in the long run, all zego insurance is going to have telematics (mandatory)
non-telematics insurance to be more expensive
We are assuming that we're going to be able to easily explain how the pricing works - this is going to be very tricky, considering that the pricing model is a blackbox and there are going to be a lot of iterations and more data points added along the way - what factors are we going to present to the customer?