RM: The topic of our roundtable is using big data to improve risk management and determine a company's total cost of risk. The challenge faced by the insurance industry today is, now that we have access to an unprecedented amount of data, how do we put it to practical use? How do we use it to influence risk management decisions? But before we get to the big data element, we first need to understand what we mean when we refer to "total" cost of risk. What combinations of insurable and non-insurable factors make up that framework?
Kobyra: From our perspective, for example in workers compensation, the clients have to deal with both their retained losses within their deductible as well as the fixed cost premium. I would define total cost of risk as a function of the retained losses, the insurance premiums associated with that, plus frictional costs such as taxes and claims management.
Flatt: When we talk about the total cost of risk, it is the retained losses for a lot of our clients that represent the biggest portion of the spend for their risk management programs. As much as 60% to 80% of the overall cost is within the retained losses. Just to clarify some of the other components, you have claims administration costs, you have collateral, you have the value of volatility and then, of course, you have the insurance transaction, so all those things together are how we define total cost of risk.
Hoffman: There are a couple of other components from a risk manager's standpoint that I would add. My entire staff, whether it is claims management, loss prevention, insurance procurement or analytics, is all part of our total cost of risk and is all included in risk management allocations.
Fick: From the insurance carrier perspective, the biggest component that we focus on is the insurable losses. A lot of our business is corporate risk management business, which retains a good percentage of that risk, so the retained losses are the primary focus when we think of total cost of risk.
Flatt: One other thing I would add is that it is important when you're looking at a strategy for how you're going to address, reduce and manage these costs, you're looking at everything in a more holistic way because there is a lot of interrelation. The things that you can do to address one aspect or element of total cost of risk may impact another.
RM: What is the biggest change in how total cost of risk has entered into the risk management framework today now that you have more data to feed your calculations?
Zeiner: That's the key point: The information in many cases is readily available, the universe is much bigger, there are better and stronger benchmarks, there's more actionable data that we can use, there's more credible data that we can use and, for our customers, it allows them to plan and execute-not only in the near-term, but over the long-term horizon as well.
Kobyra: There are quicker remediation solutions that come right to the table with the data that we have today. You can quickly identify where there are problems brewing. From that perspective, I think that's where it has changed dramatically. In the old days, you used to have to have to wait maybe a year to see how things developed and then develop actionable items to try to change the trend. With the data we have today, you can try to replicate positive trends in the areas where it is not working.
Litterer: It's the ability to put an action item in place as a result of the data. It's not just a collection, it's the usage and the coordination of risk-engineering claims and other alternate services to bring solutions to the table faster.
Zeiner: We're able to do a much better job forecasting. In the past, as Paul mentioned, we were looking back and using the data. Now, we can use the data to look forward and really focus on clarifying our view of where problems are developing so that we will have a much better sense of where we are going to end up.
Fick: The whole concept around predictive modeling is big on both sides of the equation-the post-loss as well as the pre-loss side. We now have the ability to look at the characteristics of certain claims to learn what we can do better the next time we see the same traits show up. We are also doing this on the pre-loss side where we're looking at the characteristics of the customers' risk profile to avoid the loss altogether by getting out in front of any problems.
Hoffman: From a data standpoint, it's an interesting transformation. The insurance companies I worked with 10 or 15 years ago were still on old mainframe systems. Now the insurance companies and the TPAs all invested in technology and understand the importance of having the data coded correctly because "garbage in, garbage out " doesn't do me any good as a risk manager. With the new technology and capabilities, if I have great data, I can work with my safety professionals and say, "These are the problems we have in this part of the store," or "these are the problems we have in this warehouse." That's what really gives us the biggest bang for our buck.
Flatt: From a broker's perspective, we have a tremendous amount of data gathered from the business that we're placing and the programs we are helping our clients manage, and technology allows us to leverage that data. It really helps us customize and specialize programs for our clients. It's not a one-size-fits-all kind of an approach. Let's take this data and let's make sure we're making relevant recommendations on potential solutions for the client or recommending actions they might want to invest in or take to reduce their costs. Executing on what we learn from the data is critical.
Zeiner: The medical component of our data is a bigger part of the challenge today. It is also the area in which we at Zurich, as well as the rest of the industry, are investing a lot of time and energy to really understand what is behind an injury, what the best treatment protocols are, and how you bring a person back to work effectively and keep them back at work. Five or 10 years ago, that data just didn't exist, so there were a lot of assumptions made about practices that we can now narrow down to very specific opportunities.
RM: What kinds of customers make particularly good candidates for a tailored "total cost of risk" mitigation approach? What mindset and commitment are required?
Zeiner: For someone who is focused on continual improvement, this is an effort that extends over one year, three years or five years. Sometimes you don't reap the full benefits in the first year, so you need to have a partner who is thinking long-term, who is able to invest, who is able to see it through, and someone who can change course midstream. The data allows us to structure some quick-hit improvements, some midterm improvements, and then longer-term sustainable improvement. It's those longer-term grabs where we need a partner who is really focused on continuous improvement and who embraces the data, which not all customers will do.
Flatt: In terms of a profile of the type of buyer, if they have that commitment, we have tools we're able to leverage that can help everyone-not just a particular customer. Some things will resonate more with certain buyers, like the retained loss component, but there are guaranteed cost buyers who don't have retained losses that clearly will benefit from what we're learning from the data.
Kobyra: I agree. Underlying losses are the driver for the underwriters to quote the program, so if that's under control and in focus, then your total cost of risk will be lower from account to account.
I think all customers, regardless of size, should focus on actionable items to contain loss costs.
Litterer: The perfect profile for a client, whether it's guaranteed costs or loss-sensitive, is one who also has the resources and ability to implement the tools that we have to offer. So they should have their own claims professionals and their own safety professionals to mirror and implement what we do. If it's a professional risk manager versus a buyer of insurance, that is a big difference. A buyer of insurance is probably not the best profile for a successful total cost of risk approach.
Fick: Part of Larry's point about the commitment of the buyer is that, while they have the commitment, they may not have the resources, and that really is what we're here for. The broker and the insurance company need to partner up to provide those resources, as well as to become an extension of the risk management department for the customer. I think that's really what you see when you talk about the commitment. Most customers buy into the concept. It's just a matter of having the resources available.
RM: Craig, as the risk manager at the table, is this sort of approach a tough sell for your organization?
Hoffman: I've been at my company for three and a half years. When I took over, my division was an insurance-buying organization called the insurance division. With the right resources, I have transformed it to a risk management division. We didn't have a workers comp claims team. We looked at comp as a cost of doing business, and just relied on the carrier to tell us when the person was ready to come back to work. We have totally transformed that relationship. We are now a customer that is looking at total cost of risk and we understand we have opportunities to reduce that risk and are looking forward to doing that with the right partners.
RM: Did you need to have any of those quick hits to sell it?
Hoffman: One of the quick hits that we did get was a reduction in premium up front, which was definitely helpful for me to sell it. When you save money up front, it allows you to invest money in the staff that you can then build that program around.
RM: We're talking a lot about how this sort of analysis works in general, but can you think of a specific scenario where this has led to measurable improvement?
Zeiner: When we look down through our cost-driving lens at Zurich, one of the things we know, and it is common knowledge, is that certain states cost more than others. So with a national customer who has national exposure, those states aren't all treated equally. We try to get the customer to understand that those higher-cost states probably need some immediate attention. Typically, we'll look at those two or three high-cost states and put a plan in place. You begin to see the benefits accrue in the first quarter and the second quarter, and that gives the risk manager the confidence to go a little further in year two or year three. Once we've built up credibility with results, we want to work closely with the customer's business managers to help them think about how they do business and the things they do that can change outcomes.
Hoffman: In January of this year, we had a third party warehouse operator hiring and managing part of our workforce, and they had an experience mod factor of well over two and a half due, in large part, to a failure to focus on safety and manage these claims. So one of the things we did is convince executive management to eliminate the third party, hire those associates and change that culture. We still have a way to go, but I see this as a three-year turnaround until we can say we are a safe culture at that facility. We want every associate to be safe.
Flatt: These types of success stories are key to client differentiation with the carrier. As a broker, we want to make sure that we're using what we know about our clients from data and analytics, and actions they have taken to position them best within the market and with their counterparties. Having the investment and the commitment to doing the right things to control their losses be credited on the front end before actually seeing the results is important and clearly an indication of a solid trading relationship with the carrier.
Zeiner: What we often find is that companies don't always understand their own costs, so one of the first efforts you go through is to really try and educate them on what generates their costs, why their costs are the way they are, and what are some of the things that we can collectively do right. Big data allows us to make many comparisons to frame the opportunity for the customer.
RM: What about third-party administrators? Where do they fit in from an underwriting perspective?
Fick: If you look at total cost of risk and outcomes, they're a big component. A customer like Craig is going to consider all his alternatives and, given the size of his program, TPAs play a big role here. The key when using a TPA, from our perspective, is to make sure we're still a part of that process. We want to be a part of the relationship. We want to use our past experiences to deliver insight to the customer and work closely with the TPA to make sure we're achieving the optimal outcomes. Our goal is to work even closer with the TPA to achieve this.
Litterer: It's a fruitful partnership. The larger the scope and size of the company, a lot of times, a TPA is the way they want to go for more portability and potentially increased claims service, so it's something we have to work within and just use our tools in a different way to help augment what they do. There's no less or more carrier involvement, it's just a different path.
Zeiner: We all still have the same ability to look into the data, to collect the data and to have those discussions, so the customer relationship doesn't change. It's just a matter of who is going out and implementing the changes at the end of the day.
Hoffman: From a risk manager's standpoint, the TPAs give you options. You may or may not be excited about your carrier's claims handling services and you may want a TPA to provide consistency to your program. I have a TPA on my liability program that we've been with for 10-plus years, which has allowed us to switch carriers three times during that period. The TPA allows us to maintain consistency with a team of adjusters. I don't have three different data sources. In addition, when you talk about the big data coming in, it's consistent. The same TPA has had my data for 10 years, which has allowed us to customize the data fields for our business. As a retailer, there are different things that we want to look at. We think about the different aisles in our stores-the grocery aisle, the dairy aisle-and the TPA has the ability to have those custom codes. Carriers do too now, but that gives us flexibility.
Flatt: When we're talking with our clients about the use of a TPA in an unbundled program or the use of carrier claims handling in a bundled program, ultimately what we try to help them with, in terms of their decision-making, is understanding who they are partnered up with that's going to achieve the best claim cost outcomes, who has the right networks and so on. For example, after a claims handling issue is identified, the ability of the TPA or carrier to change providers in real time to make sure clients benefit from better outcomes on the claims is important.
RM: Ultimately, how will the influence of big data and an understanding of total cost of risk change how insurance companies underwrite?
Litterer: There is no endpoint. It is a constant evolution and process of improvement. Once you get to that 360-degree review and implement your plan and procedures, it's about constant improvement. Your tasks may change because of different factors, different hot points and different pressure points, but as the data changes, so will your actionable items.
Flatt: I think that we are able to have a very customer-specific solution set. It's becoming less about the insurance transaction and more about looking at all the other aspects of total cost of risk in concert, so I see being able to deliver more customized and specialized solutions and the most efficient programs to our clients as one of the major influences of leveraging big data as a broker.
Fick: Everyone's talking about big data. Really, it's what do you do with it, what insight do you give the customer, and, when you have insight that may need a solution, can you provide the solution? Then you take that one step further and use that information to better underwrite the risk to create a tailored program that meets their needs.
Kobyra: I think Craig said it best earlier: it's not an insurance department anymore, it's a risk management program. You don't just worry about the insurance once a year and transact the renewal. You have to use
the data to make effective changes to actually lower your total cost of risk. It all circles back to a 365-day cycle, not just a once-a-year renewal with the insurance company.
Hoffman: You've seen risk management departments grow. Some companies have one risk manager; I've got a team of 13 individuals and am looking forward to growing that, hopefully, as I continue to reduce that total cost of risk. But from an underwriting standpoint, big data also allows me to understand my loss history and analyze the programs that the brokers and carriers are presenting to me. Should we be taking on that variability, should we be increasing retentions using our captive and really taking on the volatility of the insurance marketplace? Every risk manager loves to be in a soft market, but when those hard markets come, you have issues if you haven't done all these appropriate things to reduce your total cost of risk in the first place.
Kobyra: The fundamental change from 15, 20 years ago is that there was only the one essential buyer in any organization and that was the risk manager. They made the decisions. Now there are many within an organization who are working each and every day to keep the cost down.
Litterer: It's a bigger bet on our part from the insurance carrier believing in what we're going to do and how it's going to pay off. We're taking our credits upfront and giving that to the client when we know that it may take 12, 18, 24 months, whatever it may be, to see results, but we're taking an upfront bet on our abilities.
Fick: I think the biggest change is going to be if you look at traditional underwriting, underwriters like to see it on paper, actuaries like to see it on paper. They like to see a five-year history, take a look at the losses and use historical data that may not be reflective of the risk today to predict what is going to happen. If we're asking customers to make this investment, use analytics, improve the risk profile, put programs in place to improve on the pre-loss side, we have to make a certain bet that improvements are going to take place, and you're not going to see that show up on paper right away. Insurance companies are going to have to tap into their past experiences, take a little bit of a rearview mirror approach and use the knowledge they have to get a leg up on the competition and more appropriately price the business for the customer. That's the change I see coming in underwriting.
Flatt: One of the bigger changes is along this evolution. Clients are looking to brokers for the same level of sophistication as carriers when it comes to analytics and data and analysis and tools, so we put a lot of resources and investment into making sure we have those capabilities. Using what we know about our clients in parallel to what the carriers are doing helps the dialogue when we're negotiating on any element of total cost of risk, whether it is the risk transfer piece or some action on a safety protocol where you're not going to see the benefit for some time. We're all in this together and clients are demanding a higher level of sophistication.
RM: Overall, what will be the impact of big data on the insurance industry?
Fick: I think when you look at big data, it emphasizes the need to get even closer to your customer than ever before, so that you can understand their risk and what they have been doing to manage it. Big data is nice, but you have to be careful you don't just take data on the surface and apply it. You have to understand your customer and what they've been doing to drive the outcomes you see in the data.
Zeiner: We know a lot about what good cost performance looks like. In the past, we didn't necessarily share that information and perspective. It wasn't that we didn't want to do it but, in most cases, we didn't have the tools to do it effectively. Now we have the tools and experience to help customers understand and act to aim for top performance. Looking forward, we feel it is one of our primary obligations to our customers to help them understand what truly good performance looks like.
Kobyra: I think that those insurers or brokers who don't use analytics will be left behind. The way the world is going, if you don't have actionable data, you're not going to have a solution for the clients that you're trying to gain or keep.
The information in this article was compiled for informational purposes only. Zurich neither endorses nor rejects the recommendations of their discussion presented. Further, the comments contained in this article are for general distribution and cannot apply to any single set of specific circumstances. If you have a legal issue to which you believe this article relates, we urge you to consult your own legal counsel.