Insurance companies that make use of advanced predictive analytics—such as the Celina Insurance Croup—report being able to more efficiently assess and price risks, make better decisions and ultimately improve their business.
Headquartered in Celina, Ohio, the 100-year-old firm operates in six U.S. states through four entities: the Celina Mutual Insurance Company; the National Mutual Insurance Company; the Miami Mutual Insurance Company; and the West Virginia Farmers Mutual Insurance Association. The firm’s experience in investigating and implementing predictive analytics, and specifically analytics that leverage machine learning, is detailed in a new case study from Celent, “Celina Insurance's Predictive Analytics Initiative - The Machine Learning Factor,” authored by Nicolas Michellod, a senior analyst in Celent's insurance practice.
Predictive analytics synthesizes and analyzes large volumes of data to identify interrelationships between various risk attributes. These attributes can come from internal or external sources, can be current rating variables or non-rating variables and can help insurance companies with more accurate pricing, more efficient underwriting, higher retention of profitable risks and more prolific growth opportunities, according to the case study. Larger personal lines insurers are using predictive analytics extensively, while mid-size and smaller firms are starting to introduce it to their operations. Predictive analytics is commonly used for redesigning rating plans and for refining pricing, underwriting and risk selection.
Rather than rely on human intervention, machine learning leverages algorithms that learn how to improve a model as more data is collected. Interestingly, Celina was not familiar with machine learning or its benefits when it first began looking at predictive analytics, according to the Celent report.
"New techniques and emerging technologies are enabling business innovation or at least improvements in insurance," said Celent’s Michellod in a prepared statement. "Celina Insurance's curiosity and interest in what's new on the market allowed them to identify how machine learning could add value to their business."
In 2010, Celena’s chief actuary recommended that the company start using the technology, saying “it is probable that we are already losing profitable business to competitors who are better able to match premium with exposure through the use of predictive analytics.” Celina mapped out a four-phase strategy to investigate and deploy the technology and began by identifying the business applications that could benefit the most. A broad range of business and IT staff, including C-level executives, were involved in the process.
Next, Celina commenced the vendor selection process. After looking at a number of solutions, Celina asked for demonstrations and undertook proof-of-concepts with three different providers. Once the selection was made, the insurer began a phased implementation that began during the fourth quarter of 2010 and was completed during the first quarter of 2011. The system went live at the start of the second quarter, 2011.
It during the selection process, while testing EagleEye Analytics’ Talon software, that Celena first became acquainted with machine learning and the technology’s potential.
Machine learning can be used in a variety of ways, including for detecting spam and credit card fraud, and for speech recognition (Apple’s Siri uses machine learning to understand and attempt to answer voice requests). Insurance-based machine learning is designed to account fo the complexities related to insurance data, and Celent recommends that insurance companies consider vendors with experience and expertise with insurance-based analytics.
According to Celent’s Michellod, predictive analytics with machine learning can help insurers lower their total loss ratio, enhance risk adversity selection and prompt model adjustments. The Celent report notes that Celina Insurance actuaries and mathematicians are now using Talon and leveraging predictive analytics to refine the rating of their insurance products and improve their insurance fundamentals.
