Data Driven Price Optimization
Context & Objectives
An international insurer needs to transform its pricing practices in order to make them more data-driven.
Deploy algorithmic pricing practices to dynamically set product offer, pricing, and promotions.
Outcome
In average across the group, improvement of both the top-line by +10% and of the bottom line by +2pp
Our approach
Step 1 – Identify and collect
- Identify and collect all data: internal data, customer data, car data, geocoded data, other external data, etc.
- Split projects for contracts renewals and for new business acquisition
Step 2 – Build
- Build machine learning scores to predict: customer risk, customer elasticities, customer behavior, market competitiveness, lifetime value
- Define the business targets and the price optimization framework and solver
- In parallel, design the IT architecture to host the solution (dynamic optimization of prices in production). Arbitrate the make-or-buy at each step of the pricing process
Step 3 – Pilot
- Pilot first renewal and acquisition cases with a given entity, then roll-out to 4 entities across 4 geographies