<img src="https://ws.zoominfo.com/pixel/kZxG1sNctrruFoZSPoVD" width="1" height="1" style="display: none;">
Contact Us
Book A Demo
Book A Demo
Contact Us

How Target uses risk data to help manage claim and safety outcomes

Experience 2013: A preview of Target's session

Using predictive analytics and risk data to help manage claim and safety outcomes

logo: Aon eSolutions Experience Conference 2013I recently spoke with Eric Oldroyd, Group Manager, Risk Finance, with retail giant Target about his presentation at the upcoming Aon eSolutions 2013 Experience Conference in Atlanta. This will be the second time Eric has presented on the topic, but as you'll read in the Q&A below, Target has made major strides in its predictive analytics program, and Eric also plans to share some highly transferrable, practical advice about how risk managers can use RiskConsole to perform some basic, but still very useful, predictive analytics of their own.

What is your planned topic for this session?


What will RiskConsole users learn from this session?

  • How to use RiskConsole to perform "light" predictive analytics
  • How predictive analytics can be used to manage future outcomes

What's new from Eric Oldroyd's session at the 2011 Experience in Miami?

  • Target has moved from proof of concept to production, including using predictive analytics in its TPA's "toolkits"

How is predictive analytics performing for Target?

  • The model is working similarly to the proof of concept, and early indications are that it is driving improved costs

Eric Oldroyd: My topic is on how predictive analytics and data can be used to help manage claim and safety outcomes and how risk and safety managers can gain insight into what'’s going on in their risk management business through the use of data and predictive analytics.

What do you think RiskConsole clients will take away from your session?  

Eric Oldroyd: In addition to talking about Target’'s predictive analytics project, I plan to talk about some of the simple things risk and safety managers can do that are a “light” kind of predictive analytics. Specifically, using the data in RiskConsole to get better outcomes and understand where they can focus and what they should be focusing on as they look to improve future outcomes.

The term “predictive analytics” can sound intimidating, but it can be performed in a multitude of ways. It can be as simple as looking at one factor and seeing what the trends are for just one simple variable. Or it can be a multivariant-based, complex model that takes numerous factors into account. Either way, predictive analytics is, essentially, a way to really gain the power of the data that’'s housed in RiskConsole and understand ways that you can use it to your advantage.

For example, the age of the claimant is a common factor that’'s known for most claims. Using a business intelligence tool, it’'s easy to look at all of your claims and their outcomes; from there, you create a chart that shows the age of the claimant versus the outcome or cost of the claim, and you can begin to understand how correlated those might be.

To continue the example, let’'s assume that outcome and age of the claimant are correlated to some degree. And if we determine that as the age of the claimant increases the cost of the claim also tends to increase, then you can start to have different measures in place with your claims management in the future so that you perhaps approach the management of the claim differently if you do have a claimant who'’s older than average.

Please tell us a little bit about how RiskConsole is involved in this project?

Eric Oldroyd: RiskConsole is our hub for performing predictive analytics. It'’s our intake system for claims and also houses information about our store locations and our team members.

From RiskConsole we extract information about the claim, about the team member involved and about the location; we then export that data to our third party who actually will perform the calculation based on the models that we'’ve developed.

Once that calculation is done, that information is sent back to RiskConsole so that we can analyze it and communicate it to our third-party administrator so they can also use the predictive analytics results to manage claims appropriately and follow the various business rules that we’'ve set up to follow based on the results.

You presented on the topic of predictive analytics at the Miami Experience Conference in 2011. Where does the project stand now compared with two years ago?

Eric Oldroyd: Two years ago, we had done a proof of concept to understand if a model for the kind of predictive analytics we wanted to perform could actually be built. And if so, we wanted to know if it could differentiate between low- and high-risk claims and between low- and high-risk team members of actually having an incident to begin with.

Since then we'’ve actually productionized our claims predictive model so that once a claim comes in we now real-time score it and then use the information from the model to actually direct our TPA to go down various different operational paths depending on whether it'’s a low-, medium- or high-risk type claim.

The TPA will do different things depending on that outcome. The TPA doesn’'t rely solely on predictive analytics; it’'s a tool in their toolkit that helps them with things like deciding whom to assign the claim to and what actions should be taken with that claim depending on the score.

Also, in the last six months, we’'ve productionized a safely model that tries to measure the risk of locations or team members having an incident. From there we'’re prioritizing our safety projects and resources and procedures.

How is the predictive analytics project performing?

Eric Oldroyd: We believe the model is working similarly to the proof of concept. The first thing we'’ve done is try to make sure that the model is driving the same results. As we look at key metrics, we'’re confident that it is. For example, claims that the model deems lower risk are actually closing more quickly than ones that are deemed to be high risk.

What we don’'t know is: is it driving overall claim costs lower? We don’'t yet have enough data or the ability to concretely make that determination. But, we'’ve seen a few trends that would suggest that it is working well and driving improved costs. We’'re just not able to measure that yet.

Kathi Knight is Manager, Enterprise Account team, with Aon eSolutions, based in Dallas. Contact Kathi at 214-989-2575 or by email at kathi.knight@aon.com.

Registration Open

Jun 26, 2013

 | Originally posted on 

Subscribe by Email