SAP-Sponsored Article: How to Get to Rapid Cognition
Contributed Article from SAP (Sponsored)
In the last issue of this newsletter, we described the need for rapid cognition in insurance and the ability to achieve faster decision making, and we made the case for using Master Data Management as a way to assimilate, harmonize and distribute both structured and unstructured data across the enterprise. While achieving rapid cognition may seem an overwhelming task, there are incremental steps insurers can take to start moving down the path to the intelligent insurance enterprise and achieving rapid cognition.
1: Make the Cultural Changes Necessary
The path to rapid cognition requires a cultural change, says Barry Rabkin, senior analyst in the Insurance Services practice of Financial Insights (Framingham, MA), an IDC Company. The first step is to determine whether or not the company even wants to achieve rapid cognition.
If the answer is "yes," the insurer needs to make cultural changes that will enable it to sense and respond quickly to internal as well as external data. For example, a P&C insurer will need to continuously analyze weather patterns for hurricanes. An auto insurer needs to continuously monitor the safety records of automobiles to assess risk and level of exposure.
Very importantly, insurers need a culture that facilitates and encourages the sharing of data. "We throw the term 'cultural' around a great deal," says Rabkin, "but what that means is that an insurer makes changes in compensation, communications, and policies to effect cultural change. If the consequences of not sharing data do not affect employees directly, this cultural change will not occur."
2: Identify the Data You Need
In addition to identifying both internal and external data, insurers also need to understand the context of the data. Where does the data come from? Is the data source reliable? How often is the data refreshed? Is the data updated often enough that it is useful? How granular is the data? In geographic and spatial applications, for example, Rabkin believes that ZIP code data is too volatile to provide the data detail insurers need. Instead, insurers should look at geocoding data to get down to the census track and block group detail.
3: Align Staff Competencies to the Task at Hand
Whether using embedded analytics or standalone business intelligence tools, insurance employees may need a higher level of skill than ever before. The level of skill required depends on the type of activity. For example, an employee responsible for running reports against data that is refreshed needs less skill than an employee responsible for creating new or enhanced reports. An employee whose job entails developing new hypothesis or identifying new sources of data needs an even higher skill level.
Having employees with the ability to ask questions such as, "Is profitability linked to the number or type of products a customer has?" or "What's the mix and profitability of business from our brokers?" is critical to insurers getting on the path to rapid cognition, says Rabkin.
The level of staff competencies is also tied to the types of tools - both processes and technology - used. Will you use tools that you can explain easily such as GLM (generalized linear models) or decision trees? Or will you use tools that can't be easily explained, if at all, such as neural networks or genetic algorithms? The more complex the tool, the more employees using the tools need to understand statistics and data underpinnings to make rapid decisions.
Rapid cognition is achievable for insurers, but it won't happen without cultural change, a deep understanding of the internal and external data that drives decision making and the ability to hire, train and retain employees with the appropriate skill levels.
On The Net
Tweet This