Revenue Operations Strategy
Operating model, funnel definitions, forecasting cadence.
CRM & Sales Technology
Selection, implementation, routing, dialers.
Marketing Technology
Lifecycle automation, campaign and ad integration.
Data Analytics
Warehouse, pipelines, executive reporting.
Systems Integration
LOS, CRM, telephony, marketing and finance connected.
Software Development
BI platforms, routing engines and middleware built to fit.
.png)
May 1, 2022
Launching Australia’s First Pay-Per-Kilometre Car Insurer
KOBA set out to bring pay-per-kilometre car insurance to Australia for the first time — pricing every policy on how far people actually drive. That only works if the data behind it is right from day one. Illumination Labs joined before launch as KOBA’s data and analytics partner: implementing and maintaining the pricing model alongside their actuarial partner, building the analytics infrastructure from scratch, and putting monitoring in place to catch problems before they reached customers or reports.
KOBA Insurance
Project length
Pre-launch to scale
(ongoing partnership)
Technologies used
AWS, Databricks, Python, Fivetran, Plotly
Roles we played
Data Architect, Data & Analytics Engineer, Pricing Model Engineer, Analyst
Result
Launched as Australia’s first pay-per-kilometre car insurer, with pricing and data in place from day one.
01
Project overview
A pay-per-kilometre product lives or dies on its data. Every quote depends on the rating model, and every premium depends on trip data flowing in from customers’ connected cars. As a pre-launch startup, KOBA had none of that infrastructure yet — and there was no room for error in a product priced per kilometre.
We were brought in to own the data side of the launch: turning the actuarial pricing into production logic, building the systems that collect and model driving and customer data, and making sure the business would know immediately if anything went wrong.
02
Our execution
We owned the data side of the launch, from pricing to production monitoring:
- Pricing model. Implemented KOBA’s insurance rating model in partnership with their actuarial firm, turning actuarial pricing into production logic that prices every quote, and maintained it as the model was refined after launch.
- Analytics infrastructure. Designed and built the data platform from the ground up on AWS and Databricks, with Fivetran and Python pipelines bringing 10 data sources into one warehouse through 20+ automated workflows, and Plotly dashboards on top.
- Connected-car and app data. Set up collection and modelling of trip data from customers’ connected cars, alongside app engagement and core business data such as payments and email.
- Monitoring and alerting. Built frameworks that check the data continuously and flag anomalies in trip data, quote volume and pricing outputs before they affect customers or reporting, with alerts sent straight to Slack.
03
Key results
15,000+
quotes priced every day
800+
Australians insured in the first year
Also delivered
- A single source of truth for pricing, driving, customer and app data
- Problems caught by automated alerts, not discovered by customers
- A pricing model maintained in step with the actuarial team as the product evolved
Need help with a project like this?
We’d love to chat and learn more about what you’re hoping to accomplish. Fill out our form and we’ll get back to you as soon as we can.