Top Trends in Insurance Analytics
Developing capacities to deal with big data, predictive analytics, and predictive modeling is the singular current focus for strategic insurance companies today.
By Étienne Castonguay, Partner and Co-founder, InEdge
Few industries are as accustomed to understanding that information is a significant competitive advantage as insurance companies. After all, only accurate information can specify the monetary value of real estate, automobiles, and lives.
The insurance industry is challenged by big data, a collection of data sets so large and complex that it has become difficult to process using on-hand database management tools. Decreasing costs of storage have created warehouses of data. Like all industries, the challenges include the capture, curation, storage, search, sharing, analysis, and visualization of data. For the Insurance, industry, data is a strategic resource. The traditional data sets for insurance companies include huge data sets from both the actuarial and underwriting lines of business.
In addition to warehouses of data, the evolution of mobile and geospatial devices has amplified the number of data sources to which insurers have access. Together, these influences have created a significant challenge with the potential of real increases to revenue streams.
The current challenge for insurers is deciding what to do with this data. Do they store it, mine it, structure it by its source or stream, or leave it unstructured? The alternatives are perplexing, but one thing is certain: data alone is worthless. It is the insights derived from the data that matter. With the emergence of big data, the possibility of deriving insights has increased dramatically.
Predictive Modeling in Insurance
The insurance industry has become more complex as it has grown. Increased commercial options and growing appetites for properties all foster this growth. Predictive analytics is used effectively to manage increased variables. However, experienced decision makers are required to assess the value of options. Humans are subjective when it comes to assigning the correct weight to the judgment factors. Advances in predictive modeling are paving the way to clarifying the choices and assigning a reliable value to judgment factors.
Predictive models capture relationships among many factors to help people assess the assessment of risk (or potential risk) associated with a particular set of conditions, thereby guiding decision-making for successful transactions.
Over the past few years, predictive modeling has become pervasive in our lives, evidenced by the one in five couples that now meet through a dating website and book title recommendations on websites such as Amazon or contact recommendations on LinkedIn that are used easily by most people. Predictive modeling anticipates future outcomes based on historical trends.
The ability to accumulate massive amounts of data adds further challenges to the development of predictive models. In farm insurance predictive modeling, for example, it will soon be necessary to make use of information relating to local weather trends and events rather than relying on macroscopic models.
Both local weather and large-scale trends are now readily available through geo-localized mobile devices. Predictive modeling provides carriers with the opportunity to create usage-based insurance offerings with premiums based on integrated data from these devices. For instance, the OnStar system, built into many vehicles, could be fed into premium calculations; premiums would vary as a function of how a vehicle is used. Simply put, if you can build a risk level and can run the model online, the system can suggest an increase.
Predictive Analytics in Insurance
Predictive analytics encompasses a variety of statistical techniques that analyze current and historical facts (data mining) to make predictions about future events. In the insurance sector, predictive models use patterns found in historical and transactional data to identify risks and opportunities.
The world of big data constitutes a paradigm shift for insurance companies. Many insurers are still working through issues in their transactional data. Many insurers are not used to using data for operational status and decision support because of both skepticism and inconsistencies. Business line-specific data has resulted in inconsistencies in reports, leaving executives asking, "Which version of the truth should I believe?"
In essence, insurance companies are data rich but information poor. The value of the information can be unreliable and it can be very difficult to determine a single version of the truth. Even large companies have difficulty trusting the certainty of their information. Of course, they have trained people who are experts at retrieving information from their own area of responsibility, but the views of the information across areas of responsibility often clash. Obtaining the "right" information to interpret the discrepancies is difficult. The ability to pull trustworthy information out of the data is complex.
Despite these challenges, commercial insurance is becoming more efficient by creating increased automation for decision-making. For example, automobile insurance providers need to accurately determine the amount of premium to charge to cover each automobile and driver. Predictive analytics can analyze a few years of past automobile claims data, as well as other records to predict how expensive a client is likely to be in the future. Predictive analytics can help underwrite these quantities by predicting the chances of illness, bankruptcy, default, and so on. Predictive analytics can also reduce the amount of time it takes for the approval of coverage. Well-designed predictive analytics can lead to proper pricing decisions, which can then help mitigate future risks. Keep in mind that risks are the focus of insurance companies.
A Final Word
Big data, predictive modeling, and predictive analytics pose numerous challenges to many industries, but insurance companies face exciting and profitable outcomes in creatively addressing them and refining these capacities.
Étienne Castonguay is partner and co-founder at InEdge, an insurance BI and analytics consultancy, where he is responsible for sales and marketing. He has 25 years of sales management experience in the distribution of information technology solutions. He held various sales positions with Sybase, Sun Microsystems and Hewlett Packard. You can contact the author at Etienne.firstname.lastname@example.org.