As insurers evolve to become more responsive and reflective of customer priorities, product development and distribution will fundamentally shift. Insurers will develop products tailored to individual needs, priced appropriately, and distributed effectively. And data, especially in the form of consumer and market data, is the key that allows them to do this. Organizing a team capable of extracting value and insights from insurers’ data, and using those insights to improve product development, enables insurers to improve competitiveness as the market changes.
Actuaries have been the key to using and making sense of the data that insurers collect since the 18th century, often creating the necessary tools and techniques along the way. The intense rise of data collection and technology solutions outside the insurance industry has seen a new profession emerge: the data scientist.
Many insurers see the additional value data scientists bring and are starting to build teams. Typically, they have been put to work on problems outside the traditional actuarial domain. This means data science and actuarial teams generally don’t work together when asking and answering the questions that ultimately help large insurers rise to meet industry shifts. Across the industry, there remains some confusion as to how the two different professions work together, or even if they can.
To maximize the value insurers can derive from their data and optimize the product development process, a blend of actuarial and data science capabilities is essential. One does not negate the other, and the combination of both affords more capabilities and opportunities than either can derive on its own.
Actuaries vs. Data Scientists
In order to understand how and why this is true, it’s first essential to understand the key differences and similarities between data scientists and actuaries.
- Responsibilities: to organize and analyze data in order to make informed predictions about the future as well as the corresponding recommendations
- Techniques: relying on a combination of statistics, data visualization, and pattern recognition in data
- Expertise in: statistics, model uncertainty, being business goals oriented
- Data scientists mostly focus on using computationally intensive processes to find hidden patterns in large datasets, on which they base their predictions. Actuaries use a combination of statistical techniques and their business knowledge and insights to form theirs. Some would say this difference is akin to science versus art.
- Domain expertise: actuaries, due to their rigorous formal training and historic leadership in the insurance industry, tend to be better versed in its business processes, regulatory and accounting requirements, claims management, underwriting and product design
- Coding expertise: actuaries have traditionally not been required to be proficient in coding or software development, which is a data scientist’s bread and butter
- Tools: actuarial tools typically involve a combination of custom actuarial system and Excel, while data scientists are more coding savvy and will likely have a background in a variety of languages like R, SQL, Python, and more
More insurers are prioritizing data scientist recruitment, and as PwC discussed in its report on the subject, this creates a valuable opportunity to encourage integration between the two. The fundamental gap for actuaries is the knowledge of the tools and techniques available for dealing with large datasets, and the coding skills needed to use them. As a result, they often miss the valuable insights lurking in the data. Data scientists lack the in-depth knowledge of how the insurance business model works. This means they often answer the wrong question or provide answers that are incoherent with the business context. There is much to gain when the two teams can work together and create a more nuanced approach to valuable data extraction.
While SOA has been incorporating certain data science skills, like predictive analytics, into its core education syllabus, finding real world crossover hasn’t been obvious. Encouraging collaboration and cross-training between actuarial and data science teams has been challenging in practice, due to common distrust between the two disciplines. Actuaries value business acumen and so sometimes have the tendency to find data scientists naive. Data scientists value a rigorous black and white approach, so could interpret actuaries’ analysis as unreliable, or even irresponsible. Both are unfair.
At Montoux we’ve seen this first hand; our decision science platform uniquely combines both actuarial and data science directly in its software. In order to do this, our strong team of both data scientists and actuaries are constantly collaborating. This ensures our life insurance customers get the maximum benefit from our software’s capabilities and helps them make better strategic decisions by integrating insights developed by data science teams into existing actuarial capabilities.
The importance of finding ways to integrate actuarial and data science, however it’s accomplished, comes from the availability of large, integrated datasets and the insights they hold. Leveraging these insights can improve insurers’ internal processes, making them more refined and agile. A combination of actuarial and data science is essential for this approach. When both capabilities collaborate in processes like product development, it means better products for customers, better portfolio performance, and a more customer centric approach.
Why Product Development is the Best Opportunity to Combine Data and Actuarial Science
Demands on insurance products are rapidly changing. Customers want a faster, more flexible, and personalized insurance product than the industry is historically used to delivering, especially in light of COVID-19. This means the product development process needs upgrading.
In order to really tune product development towards customer centricity, insurers must:
- Be able to analyze customer data understand their behavior and the reasons behind it
- Constantly track of competitor movements in the market, including what is and isn’t working
- Track individual products and prices in the market in order to understand why they perform the way they do individually and as pieces of the portfolio
- Understand the performance of the product from an underwriting, claims and operational perspective
- Know which product features customers value and which are a problem for the insurer
- Have a system that allows for constant modeling, experimentation, and ‘tweaking’ in order to test what does and doesn’t work in terms of price and distribution
This may seem like a tall order, but incumbents have massive data resources to begin making changes. The ability to explore the potential of different products, prices, and distribution techniques, gives insurers’ teams the ability to optimize their processes like product development. This leads to better data analysis and insights into an insurer’s portfolio and an understanding of how to adjust products and prices to hit targets. Ultimately, this is a repeatable, measurable approach that lends itself to success.
The reason product development is the ideal opportunity for actuarial and data science to merge is because all of the adjustments listed above require both sets of capabilities. The only way to achieve this optimization is to fully apply both sets of skills in a collaborative setting through a team of multi-skilled individuals. The product development process evolves, taking full advantage of both skill sets while ensuring the work is cohesive and supported. It also means the process has the potential to become fully optimized, turning every opportunity that presents itself into a lever for success.
An Opportunity to Innovate in a Meaningful Way
Many insurance customers are losing sight of why their insurance products are relevant to them, and it falls on the industry to remind them. Regardless of the approach, the importance of the opportunity that lies in innovating the product development process in insurance is significant.
While there are many differences and similarities between data scientists and actuaries, one certainly doesn’t render the other obsolete. Data science is crucial for gaining a clear and robust understanding of the massive amounts of data coming to insurers from the market and customers. Actuaries play an irreplaceable role in the product development and pricing process that’s not going away. Finding innovative ways to encourage meaningful collaboration between the two in product development means an opportunity to optimize the process to deliver better products to customers and better results for insurers.
One of the ways to encourage this is to have tools that simultaneously improve insurance processes and offer an opportunity for actuaries and data scientists to collaborate directly on the same platform. Montoux’s decision science platform is one example, being fully grounded in and reliant on both sciences to provide its full capabilities for life insurers. Integrating tools like this not only ensure that processes are enjoying the benefits of both actuarial and data science, but provide opportunities for these two skill sets to work directly within the same platform to the benefit of all.