Big data is now ubiquitous in the insurance industry, but most insurers are merely scratching the surface when it comes to effectively harnessing its value. In fact, a recent study revealed that 60% of claims data is still unstructured, and that 16% of high-risk claims are still neglected. In addition, about a third of those polled are developing a strategy for dealing with their data, but only 9% have actually launched a formal program. It’s clear that not only effective metrics, but also effective analytics are essential for identifying opportunities and threats — and for driving the right performance improvements in the organization. In other words, insurers must exploit data in order to maximize performance.

Predictive analytics involves looking for patterns and trends in claims data that can foreshadow issues and outcomes — so leaders can take action with the appropriate people, processes, and technology. Many insurers are already using analytics to help with fraud prevention. But in our experience, fraud is merely the tip of the iceberg. Predictive analytics has the potential to affect many other areas within the claims organization, including (but not limited to) loss triage, recovery, specialized interventions throughout the claim life cycle, case reserving, and claim evaluation. In fact, predictive analytics can do as much or even more in these areas as it can for fraud prevention. And while the focus of this article is on analytics within the claims function, Nolan believes predictive analytics can also drive significant business improvement in such areas as underwriting and actuarial interventions. In other words, there’s still a lot of value to be realized in the proper application of big data.

Predictable Impact

Here are some of the most compelling ways that predictive analytics can transform your claims organization:

1. Litigation Avoidance: Certain characteristics of past claims are powerful indicators in predicting whether a current claimant will retain an attorney or pursue litigation. When analytics identifies this subset of claims, specialists can flag them early and apply advanced rapport techniques or other interventions to help minimize severity.
2. Case Reserving: By examining the obvious and not-so-obvious data values and key correlations, claims experts can spot case reserve issues that are likely to develop. This facilitates more targeted reviewing, and can even help organizations develop procedures for automated escalations.  
3. Severity Interventions: By recognizing claims that exhibit signs of severity inflation (legitimate or otherwise), you can respond sooner with the appropriate interventions — whether it’s a medical review, expert review, enhanced case management, internal review, fraud referral, or something else depending on the situation.
4. FNOL Triage: Getting the right claims into the right hands — without hand-off or hesitation — can have a profoundly positive impact on severity control, efficiency, and customer satisfaction. Predictive analytics helps identify patterns and characteristics that can facilitate more accurate FNOL triage. And that process becomes even more powerful when it’s coupled with an automated assignment tool that also flags any subset of claims that might require extra supervision.
5. Claim Evaluation: Given a validated set of data, the range of past outcomes can, in fact, predict the trajectory and outcomes of future claims. And while relying solely on such values is risky business, organizations that use analytics appropriately as part of a good-faith evaluation can dramatically improve consistency and performance in evaluating claims.
6. Recovery: Data analysis can reveal recovery candidates and can also inform recovery success rates. By redesigning your recovery operations to take advantage of these new insights, your business can increase recoveries, speed up cycle times, and avoid unnecessary recovery expenses.
7. Supervisory Oversight: Analytics can help identify claims that are more likely to benefit from supervisory input. Rather than requiring supervisors to review every file at rote intervals, organizations can improve efficiency and effectiveness by focusing more on claims that demonstrate a specific need for special oversight. 
8. Fraud Detection and Response: Recognizing patterns of fraud pre-dates big data — going back to the manual “red flags” and “yellow flags” that prompted SIU interventions for decades. So it’s no wonder predictive analytics first made its mark in the claims world as a powerful fraud-prevention tool. Now, carriers are realizing the full value of their data and redesigning their fraud response capabilities — both at the strategic level (e.g. program design, insourcing vs. outsourcing, vendor selection, etc.), and at the tactical level (to determine which cases get what kind of response).

From Data to Decisions (and Action)

Data by itself is important. And the modeling and data mining techniques employed by predictive analytics can turn the right data into powerful tools for positive change. The challenge is that the best data in the world is meaningless without the proper analytics. So the key is to develop a strategic vision — and then use data and analytics to focus and refine the execution.  

Here’s what you can do today to either harness the value of your data or get more from your current program:

1. Ask the big questions: Don’t design an analytics program without considering the underlying business issues. At the same time, don’t try to solve everything at once. One or more well-defined business questions or problems should provide enough focus to get started with analytics or advance your current practice.
2. Know where to look: In the beginning, focus on the business questions whose answers are most likely found in the data you already have. At the same time, take care not to overlook some less conventional sources of data. For example, unstructured data from adjuster notes, claims correspondence, or freeform fields (whether in core or ancillary processing systems) can be incredibly valuable.
3. Keep it simple: Fraud detection and response is one of the most mature areas of analytics for Claims. Consider starting there if analytics are new to your organization. If you’re already experienced in analytics, consider other business issues such as litigation avoidance, severity recognition, or case reserving. 
4. Shop around, and ask for help: Depending on the complexity of the business issues and the sophistication of the required analytics platform, you may need to go to market for software or services — or you may be able to work faster with more rudimentary methods. An independent advisor can help identify business drivers and opportunities based on your individual markets and performance. The right advisor can also help cut through the clutter of sales pitches from vendors with only one solution set to sell. Beware of wizards with only one potion.

Once the business issues are defined and the analytics platform is established, the goal is to get practical, actionable information from the newly-acquired business intelligence. From there, leadership must put that information to work through new and updated business practices and processes. Only then can the organization derive the real value from an analytics program.

The process can be daunting, but help is readily available. Skilled advisors with deep industry experience can guide businesses through planning and implementation, measure for impact, and fine-tune the system going forward. When properly designed and built, a good analytics program can create a continuous, self-sustaining cycle of improvements.

Garbage Out, Garbage Out

It goes without saying that the intelligence derived from analytics is only as good as the underlying data that informs it. Still, organizations risk acting on skewed analytics mined from flawed data — and the consequences can be catastrophic. The good news is that there are practical ways to create an organizational culture that’s committed to data quality.

Prediction is Power

With the proliferation of big data and the emerging potential in predictive analytics, insurers are realizing it’s more important than ever to evaluate the data being mined in their claims organizations. It’s not just about the quantity — or even the quality — of the data. By identifying important trends and patterns, predictive analytics can make the same data work harder — so leadership can anticipate issues and respond accordingly.

Bottom line: By tapping into the potential in your data, your claims organization can be far more effective — and create significant efficiencies along the way. You’ll pay the right amount for more claims. And your customers will be more satisfied with their claims experience. It just takes a commitment to understanding and effectively applying analytics.

How is your business responding to the influx of big data — and how have you applied analytics to move your organization forward? As always, we welcome you to share your experiences.