In the third of our five-part blog series, Robin Wagner, TransUnion’s Senior Vice President for International Insurance, takes a closer look at how insurers can use data to improve pricing models and customer segmentation for a winning formula in an increasingly saturated market.
How firms benefit from using credit and alternative data for risk and price segmentation
So far, we’ve covered how data can help you go digital and direct to build competitive advantages in the new insurance landscape. But without the right pricing and segmentation, even the best digital platforms in the world won’t help you hit your numbers.
The benefits of improved data-driven pricing go beyond helping insurers achieve quarterly targets. They also help insurers serve previously uninsured market segments, ensuring more consumers than ever before benefit from having adequate insurance in place. In short, advances in pricing enable what we call insurance inclusion.
Insurance pricing: then and now
Taking their word for it
There’s a saying in insurance that there is no such thing as bad risk, only bad pricing. If you consider the journey insurance pricing has undergone over the past few decades, the expression is true. As pricing models have improved, insurers have been able to take on more consumers from more diverse risk segments and grow their books profitably.
Before alternative variables like credit data were introduced to pricing models, insurers had to rely on declared variables: the information consumers provided in their applications. Even in the late 90s, many US insurance companies still relied solely on this information to ascribe a risk tier and pricing structure to each customer — they simply had to take the consumer’s word for it and trust that the data submitted was accurate and true.
This type of data never went through third-party validation, which made the risk of non-disclosure of important information a very real threat. Consumers could and did either purposefully submit incorrect information or leave out information by accident, which led to incorrect pricing, skewed risk profiles and premiums that were either too high (pushing consumers to competitors) or too low (leaving the insurer at risk of paying out more claims than the premiums they collect).
The proof was in the numbers. Profitability in the insurance industry is expressed by the combined ratio, with ratios under 100 indicating a firm is making an underwriting profit, and ratios over 100 signaling the firm’s operations are operating at a loss. During this period, the combined ratio for the US insurance industry was consistently higher than 100, reaching approximately 116 in 19921 a clear sign that inaccurate pricing was heavily impacting the profitability of insurers nationwide
Moving to credit-based insurance scores
It was around this time that innovative firms in the US started looking at using other variables like credit to improve their pricing and risk segmentation strategies. There is a clear correlation between credit risk and insurance behavior. Credit data is regulated, controlled and ubiquitous and, since most consumers looking for insurance were likely to have credit profiles, it could be used as a strong, verifiable indicator of the risk.
To support these improvements, TransUnion introduced the first generic credit-based insurance score in 1999. Today almost 95% of insurers2 in the US use credit as one of their top three variables in underwriting, leading to improvements in the average combined ratio, which was approximately 99 in 20183.
The benefits of using validated credit data
Insurers are now benefiting from the use of verifiable, widely available third-party data in two ways: first, it helps create a better customer interaction; and second, it helps them drive financial inclusion and profits at the same time.
Unlocking underserved markets to drive inclusion and grow profits
Insurance forms part of the backbone of any market’s economy. It helps citizens protect their asset base, which in turns helps build a stronger economy. The future of insurance lies in making it relevant and affordable to broader populations.
Different and alternative data sets are integral to this process, as they let you dig deeper into segments that can’t be scored using traditional models. Using advanced segmentation tools, they can price better to give more people access to their products.
But the benefit of using this type of strategy gives insurers more than a social responsibility feather in their cap: it benefits their bottom lines. The more effective your pricing models are, the better you can determine risk. You can manage higher risk tiers by means of appropriate premiums and reward lower risks tiers with discounted rates.
By building insurance markets that use alternative data integrated with traditional credit scores, you have access to untapped and uninsured market segments. Advancements in credit scoring deepen this effect, enabling insurers to target two underserved market segments that hold great growth opportunities: thin-file customers who have limited credit records and wealthy customers who seldom use credit.
Benefits of a solution like CreditVision™, which uses trended credit data for scoring, have been exceptional. In South Africa for example, we found there were 2.5 million consumers who could not be scored using traditional credit data but were scoreable using trended data. That is a considerable, untapped market where insurers can safely expand their portfolios and deliver much needed products.
Using pricing models to improve customer experience
Using third-party and alternative data sources allows for smoother onboarding, which creates better customer experiences. The usual pen-and-paper underwriting process is a very onerous and complicated one that doesn’t resonate with today’s consumers. Even companies that use digital application forms still require consumers to fill in question after question, their frustration growing with every click of a mouse.
Thanks to advancements in pricing and segmentation, insurers can use real-time and validated data to confirm various pieces of information (to assist with ID verification and fraud prevention) and even automatically populate certain sections of an application form
When the unforeseen happens and customers submit their claims, the claims process itself can be sped up considerably by these advancements in automation. And, thanks to the data-driven pricing and premium structures, insurers can confidently pay out claims knowing they are in line with the firm’s risk appetite, making it a win for all parties involved.
Tech today: there’s still work to be done
Increasing the uptake of advanced models
While many major markets are surging ahead with data-driven pricing and segmentation, there is still a lot of work to do in other regions. Markets like India and Mexico still don’t use data and analytics to underwrite vehicle drivers. They simply look at the replacement value of the asset that is being insured, so even if you have a long list of prior insurance claims you can still get a pretty decent premium on a well-priced car. For these markets and others, we need to overcome strong legacy structures by empowering (and convincing) large insurers to adopt data and third-party validation. In line with this, we’re currently working with direct insurers in India to help boost the adoption of data-driven pricing and we’re consulting with insurers in Hong Kong where, despite other tech-savvy data tools like our eKYC ID verification solution, the market still has as a combined industry ratio of 104%.4
General insurance pricing using alternative data has progressed considerably over the years. Tools like CreditVision use trended and alternative data to provide as much as a 25% lift in consumer credit risk scoring over traditional credit risk scores5. The inclusion of additional third-party data such as claims and coverage information adds another layer of risk protection by helping insurers create a more 360-degree view of a consumer, their credit behavior, their insurance history and any geographical risks the insured asset presents.
This type of information is invaluable in markets like Brazil where firms rely solely on alternative data to complete risk segmentation and pricing for off-bureau customers. Using our models, industries like vehicle insurance can leverage scores that provide deep insights into unbanked and thin-file consumers so they can better price and segment customers that could not be scored before and create a holistic view of the risk that these consumers present.
Artificial intelligence and machine learning are also adding powerful predictability to the mix and they’ve helped us develop intricate, industry-specific models much faster by taking unstructured data and combining it with traditional scoring models.
But machine learning isn’t faultless. What you gain in speed, you lose in interpretability. In other words, you may not have a clear enough understanding of how machine learning used information to come up with certain rates, leaving you with no real explanation of why a model works…and why it doesn’t. We are, however, seeing some great advancements in this area and I believe in the next two years we will have AI and machine learning models that provide a compelling balance of speed and interpretability that can give insurers further competitive advantages
Problem solved: Alternative credit data helps insurers use advanced segmentation tools — with AI and machine learning solutions on the horizon — so they can better price and segment consumers to mitigate risk and grow profits while promoting financial and insurance inclusion.
Our next challenge: compliance and solvency management
We’ve addressed how insurers can better interact with customers, whether it be digitally or direct, and how they can find effective ways of pricing and segmenting those customers. But data challenges don’t just impact external operating factors. In our next blog, we see how data can improve compliance and solvency management for insurers