Visualize: Insights that power innovation

Visualize: Insights that power innovation

5 questions to ask before building a claims fraud analytics system

By Jessica Turner  |  July 30, 2018

Claims fraud analyticsWith the industry focus and media attention on data analytics in insurance, many SIU leaders are considering enhancing their team’s investigative capabilities with a fraud analytics solution.

But there’s a dilemma: Do you build a system internally or partner with a service provider?

It may seem practical to create an in-house system. After all, you know and trust the capabilities of your company’s IT department, and you can customize the solution to your SIU team’s needs. But building your own fraud analytics system is more involved than it seems, and the time and costs add up quickly—especially for small or medium-size carriers.

Before you green-light building an analytic solution, here are five questions to consider:

1. Do you have the time and resources?

Building a data analytics system requires a significant amount of design, coding, and testing that can overburden an IT department. Developers will often have to balance their regular tasks with supporting the creation of a new system.

Unexpected issues often arise that prolong the process and add to the costs. I worked in an organization where it took us almost a year to create an in-house analytics system, and then we ended up abandoning the project altogether because it was draining too much time and resources.

2. Are you prepared to hire new personnel?

If you commit to building your own solution, you’ll need to add analysts to your team to evaluate the data, review claims for fraud scoring, and monitor the system. A standard analytics team includes a data scientist, data engineer, and business intelligence analyst.

Not only do these positions add a significant expense to your budget—data scientists earn an average of $121,000. annually and data engineers earn $138,000—finding the right talent to fill these roles is challenging. According to an MIT study, 40 percent of companies struggle to find and retain data analytics professionals. Are you prepared for a prolonged talent search?

3. How will the data analytics staff integrate into the organization?

If your company hires a data analytics professional, chances are he or she won’t work exclusively with the SIU team. In my experience, analysts become part of the IT department and serve many business units in the organization. So, you may have this dynamic new fraud detection solution, but reviewing claims data may not be the priority for the new data scientist.

The benefits of a fraud detection system are limited without a dedicated data scientist.

4. How will learning a new system affect your team’s workflow?

Whenever you implement a new technology system, a learning curve is involved. Not only will your team have to be trained on a new system, they will have to adjust to a new workflow. As your team tries to get up to speed, the process potentially distracts from the cases that need to be investigated. Questionable claims may potentially go undetected, or the claim cycle may be extended as you try to get up to speed on your new system.

5. How effective is the system?

Any analytics program is only as good as the data available to analyze. You’ll likely be able to track and analyze only your own claims data, which limits the system’s effectiveness. Access to claims data from other carriers would create a more comprehensive solution better at detecting fraudulent activity. For example, if a customer has already filed three claims this year with another carrier, your in-house system wouldn’t be able to detect that information.

A simpler way to detect fraud

Creating your own SIU analytics system has several moving parts and may be more complex than initially anticipated. Whether its hiring an analytics staff, implementing the solution, or managing the change, taking your fraud detection to the next level requires plenty of planning, research, and resources.

Many carriers choose to partner with an established fraud solution vendor instead. Service providers offer ready-to-use solutions that can be implemented quickly and easily and require less time from your staff.

Verisk has an automated fraud detection solution, ClaimDirectorSM, that compares your claims to more than a billion claim records in ISO ClaimSearch®, using proven fraud indicators. The solution can be customized to meet your company’s needs and triage claims based on your specific rules.


Jessica Turner, MBA, CFE, CIFI, is fraud analytics and services manager at ISO ClaimSearch. You can contact her at jessica.turner@verisk.com.