Auto insurers can leverage innovative pricing models, proprietary data and machine learning to drive the future of auto insurance 🚗
Auto insurance is a hyper-competitive industry. The primary basis of competition is price, and popular “comparison” websites let customers easily compare options from multiple insurers. Underwriting is critical for all types of insurers and this is particularly true within auto given the head-to-head nature of competition.
- Players with stronger underwriting experience positive selection bias, where they offer the lowest prices to the lowest-risk applicants and “win” good business.
- Players with weaker underwriting experience adverse selection, where they offer the lowest prices to the riskiest applicants and “win” bad business.
With that in mind, it’s no surprise that auto insurers are constantly looking for ways to innovate within the underwriting function. This article discusses a few exciting options.
1. Automated Machine Learning
Most auto insurers have highly-trained statisticians and data science teams that do terrific work. It’s their job to gather data that’s relevant to underwriting, prepare the data for use, identify appropriate modeling techniques, build predictive models and validate performance. This typically takes a few months and the models are then used as a core component of the overall underwriting process, generating risk scores that help determine eligibility and set rates.
Exponentially increasing data sets and the constant addition of new variables place a heavy strain on data science teams – by the time a model is built, it’s often time to start over! As a result, when organizations identify additional opportunities for predictive analytics, there’s often a shortage of data scientists to solve the problem.
Insurers are increasingly embracing next-generation tools to address this challenge and Automated Machine Learning (or AutoML) is leading the charge. As we discussed recently in our post How Automated Machine Learning (AutoML) is Leveling the Playing Field, AutoML builds upon the amazing work done by world-class data scientists and makes machine learning available as a service that companies can use to easily train accurate predictive models. In short, it drastically shortens the time required to build predictive models, empowering data scientists to focus only on the highest-value work while AutoML handles the previously time-consuming processes of data transformations, modeling and evaluation.
2. Proprietary Underwriting Data
As more insurers embrace next-generation analytical tools, proprietary data is becoming critical to insurers’ underwriting processes. Better data leads to better predictive models and proprietary data provides an “edge” over the competition.
Auto insurers have traditionally used basic data about the customer and automobile as part of the underwriting process. This information is predictive but very common, and therefore doesn’t create meaningful differentiation in underwriting vs competitors. In recent years, insurers are using new technologies – often placed directly within the car – to collect information about driving habits, including:
- What is the customer’s average driving speed?
- Does the customer stay within X% of the speed limit?
- How often does the customer swerve?
- Where are they typically driving (cities, highways, etc.)
- When are they driving?
This is proprietary information that can’t be collected on application forms and, since it directly related to driving safety, can be very predictive.
The primary challenges are cost and adoption. With regards to cost, the technology (hardware and software) used to collect and track all this information is typically expensive, so insurers need to determine whether the underwriting improvement justifies the cost (for example, if the technology will cost $100 per customer, will it lead to $100 in reduced losses?). With regards to adoption, the insurer needs to convince the beneficiary to put this technology in their car – this is both intrusive and can lead to selection bias that skews the results.
3. Innovative Pricing Options
Another way that auto insurers can innovate is with pricing. Traditional auto insurance pricing is on a “per year” basis, regardless of how frequently the customer drives. However, insurers are beginning to use other pricing methods, such as “per mile” or “per hour”, to better align the price charge with the risk taken by the insurer.
Similar to the discussion above, the challenges are typically around cost and adoption. These pricing methods require detailed tracking, which can be expensive and intrusive.
Auto insurers are always searching for ways to create competitive advantages within the underwriting function. AutoML, proprietary underwriting data and innovative pricing options are increasingly being used to differentiate the offerings of auto insurance companies. However, they often come at a cost – and it’s up to each insurer to decide whether the trade-off is worth it or risk being left behind.
DigiFi is a technology company that helps businesses make better automated decisions.
Our platform lets businesses easily use automated machine learning and rules management to optimize critical decisions with no coding or technical expertise required. Repetitive work that used to take hours can now be completed in minutes, letting your team focus on what matters most.
Learn more at digifi.io