Predict the Outcome,
Boost the Bottom Line
ISO optimizes efficiency and profitability by being a leading source of information about risk.
As the global economy evolves, all companies are increasingly leveraging predictive analytics to make better business decisions. Karthik Balakrishnan, Vice President, Analytics, ISO Innovative Analytics; Beth Fitzgerald, Vice President, Commercial Lines and Modeling, ISO; and Glenn Meyers, Vice President and Chief Actuary of ISO Innovative Analytics, know what goes into successful partnerships.
BEST’S REVIEW: Why is ISO building predictive models?
BALAKRISHNAN: Being able to effectively identify and leverage opportunities, differentiate from competition, control costs, and allocate resources and capital appropriately, are some of the challenges facing insurance carriers. With profound implications on the top and bottom lines, these challenges are especially critical to address in soft market cycles. Predictive modeling can help with these challenges. If one can accurately and reliably predict future outcomes, then companies can proactively make sound business decisions. Predictive models are increasingly being used in the insurance industry in areas such as marketing, underwriting, pricing, claims and loss control, etc. With ISO’s predictive models, companies can make better decisions.
FITZGERALD: ISO has the infrastructure and years of experience in data collection, actuarial analysis and filing that analysis with each of the individual state regulators. By investing in predictive modeling, ISO has expanded its resources to include additional data sets and statistical tools to broaden its statistical capabilities.
BR: What are the critical factors for success?
MEYERS: What you need for any predictive-modeling project is a well-defined goal. Then you need access to relevant data. In all the predictive-modeling projects I’ve been involved with, cleansing and gathering the data takes over 85% of the time. You also need a solid modeling infrastructure, including a good statistical package that can manipulate large data sets and provide a variety of statistical models. Forming effective multidisciplinary teams with business, data manipulation, and statistical modeling experts is another critical factor for success.
BR: And what are the challenges in making sense of the data?
MEYERS: One cannot dump a lot of data into a statistical packaging and expect a good model to just pop out. You need human intervention at every step. You’ll find that there are several different models that may give good predictions, but the final model often is determined by the business objectives.
BR: How will the industry use your solution?
BALAKRISHNAN: Interested carriers can either use our complete model “as is” or license the refined and enhanced “data features” in the model, which they can use as inputs in their internal modeling efforts. For instance, ISO’s Risk Analyzer Personal Auto® is being filed as an optional rating plan and carriers can choose to adopt this plan as is, or simply license its components for use in their modeling initiatives. Another option is only available to carriers that have agreed to work with us to develop models. They receive a custom model that fits their business and data, in return for working with us to create these industry-leading solutions.
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35