When people think insurance, they think probabilities, statistics, and efficiency. That’s why customers have even less patience with their insurance companies when waiting for a response on a claim. While insurance organizations are some of the most adept companies at harnessing the power of data into developing products and risk profiles, there are often opportunities to apply the same discipline to the operations of their own organization, particularly their customer interactions. Minitab can help insurance companies address claims faster and improve their relationship with customers.
Measure Speed of Claim Response Time
The time taken to resolve a claim should be a key measure for insurance companies. With the output measure being one data point, it should be fairly simple to collect and measure. Using a sample data set and some simple descriptive statistics, the example below shows that average (or mean) for the time to resolution of claim is between 54 and 55 days. The data also indicates that the minimum time is 40 days and the maximum time is 75 days, so it provides a range of what the fastest and slowest times are which helps with goal setting.
Set a Goal and Brainstorm Possible Factors That Impact Performance
Slow claims cost the organization not only in terms of the customer experience, but the longer a claim goes unresolved, the more uncertainty for the organization regarding liabilities. Set a strategic business goal to settle claims within a certain timeframe. In this example, let’s set a realistic target of 50, which represents ~10% improvement in time to resolution.
Next, brainstorm the possible variables that could be impacting the timeframe of resolving claims. This could be anything from claim amount, type of claim, type of customer or even the agent that is handling the claims. The diagram below shows an example of a popular fishbone diagram, one of many popular brainstorming and structured problem solving tools.
Ready to Brainstorm Problems?
Use Predictive Modeling to Quantify Impacts…
In general, predictive modeling is helpful in assisting in making predictions as well as understanding the factors that are influencing the response. By using Minitab’s Automated Machine Learning Tool, not only do we get to see the best model (in this case Random Forests®), but we also get to see how other models performed.
In this case, the popular and traditional regression method not only performs the poorest, but also isn’t very accurate. Positively, the CART® model, ideal for visualizing relationships, performs relatively well.
By looking at the CART® decision tree below, it becomes clear that Auto claims take the shortest amount of time to be resolved, while Burglary & Theft claims take the longest. This provides the first area to address for improvement. Looking one step further, it is clear that case agent Rebecca is struggling with these cases, in particular. Training Rebecca on these particular cases could lead to some immediate improvement.
…And Operationalize the Model to Communicate Better with Customers!
Not only can this analysis help identify areas of improvement, but it can also help communicate with customers. By taking the factors at hand, and leveraging the most accurate Random Forests® model (as determined by Automated Machine Learning), we can operationalize the model to automatically communicate with customers. Using solutions like Minitab Model Ops, as these data points are collected, the model can calculate the estimated time to resolve the case and automatically communicate the timing to customers. This will ensure that customers’ expectations are set properly so they will not be disappointed. As you improve your performance not only can you overdeliver on your customers’ expectations, but you can continue to refine your predictive model to provide more accurate timing to customers in the future.
Looking to Solve Other Insurance-Related Challenges?