Analytics for Insurers: Hindsight, Insight and Foresight

Data analytics can turn insurance companies into fact-based cultures, allowing them to see business as usual in a new way. While the potential benefits seem unlimited — from business intelligence to predictive modeling — applying analytics is a complex process that requires support from the top down.

Tune into this episode of Deloitte Insights to learn more about applying analytics in the insurance industry.

Guests:

• Linda Pawczuk, Deloitte Consulting LLP and Insurance Technology practice Leader
• John Lucker, Principal, Deloitte Consulting LLP and Advanced Analytics & Modeling National Practice Leader

Transcript: Analytics for Insurers: Hindsight, Insight and Foresight

Sean O’Grady: Hello and welcome to insights. On this episode, we will discuss the benefits and best practices associated with applying analytics to the insurance industry. Joining us today in the studio is Linda Pawczuk, a principal in Deloitte Consulting LLP and leader of the Insurance Technology practice and we have John Lucker, also a principal in Deloitte Consulting LLP and the leader of the Advanced Analytics & Modeling National Practice. Folks, my first question is for the unacquainted and that is what are analytics and how are they currently being used in this industry?

John Lucker: That is a good question Sean and it is a broad question and it means a lot of things to different people. Think of analytics as a process, a way of taking data and using it to create a fact-based culture within an organization, within an insurance company, and that fact-based process allows a company to see things in new ways using the foundation of their business, which of course is data. We think of analytics as a continuum, going from the base of the continuum, as the data itself, the management of the data, the creation of intelligence from the data, business intelligence, and ultimately leading to the top of the food chain, so to speak, using advanced analytics or predictive modeling to actually predict the future and that continuum, we label as hindsight, insight, and foresight.

Sean O’ Grady: Linda?

Linda Pawczuk: Yeah, specific to John’s comments with regard to hindsight, insight, and foresight, I think it is important for us to point out too as related to those who are not acquainted with this, is there are some very important characteristics of analytics, not just what the information becomes, but more importantly too the information characteristics around the strategy. We know that organizations today that succeed with analytics are founded with good principles in support at the very top of the organization. Analytics is not an IT-led capability. It in fact is a business-led capability and so that support becomes very important. Part of the importance of that support around a solid information strategy and the right stakeholders and sponsorship of the process is also the understanding of the complexity around lots and lots and lots of data that have been aggregated over the years that John and I certainly know through our years in the industry. For instance, a company that is a Fortune 100 company has over 20 different chart of accounts, has 17 different source systems that they apply and get information from; in this particular case to reconcile the information, there are over 300,000 reports that they are required to look into to get to the analytics information. That is very complex and that is because of redundant systems, legacy system that have existed in the industry for years, so that data conversion aspect of getting to analytics is a very, very complicated issue.

Sean O’ Grady: I had like to go back to something you mentioned earlier in your remark Linda and that is that there are challenges when applying analytics; what are these challenges?

Linda Pawczuk: Well, as I mentioned before, with the company example, is when you have source data that resides across multiple systems, the process of converting data and producing high-quality information to get to good analytics is very complicated. You must put in place good governance structure. You must have agreement between the business and IT what data definitions in fact are around the data that you have, so that process of understanding the data, the quality of the information, and getting to a process of conversion can be very, very complex.

Sean O’ Grady: John, you agree?

John Lucker: I do but I would like to add another thought to that and that is around the data quality issue. I think a lot of companies get too hung up on data quality when it comes to the advanced analytic side of the equation. I think it is very important for data to be clean and as clean as possible foundationally and insurers need to remember that they are running a company with the data. They are doing their financials, they are reporting to regulators, and they are sending data to bureaus, so data generally is good enough for the process of advanced analytics. What advanced analytics tends to do in a lot of the processes and statistical procedures that are used, what it does is that it washes out the dirtiness. So when there is something wrong in data, it is very often randomly distributed and that randomness allows you to statistically wash the dirt away. So it is an important thing to keep in mind. A lot of companies will use data hygiene as a reason not to do something when in fact there is a lot of business value that they could realize.
Sean O’ Grady: Keeping the data clean! So we talked a little bit about these challenges, but really how does an insurance carrier get around these challenges and ultimately get value out of the analytics, John?

John Lucker: Well, first of all, I think understanding what that value is most important and it is very important for insurers to undertake a process of creating their benefit analysis for the docket of things that they want to do. Everybody at a company has a lot of ideas and those ideas end up making their way on to a list. What often does not happen is that people do not quantify the value of those ideas, so assuming those ideas have now been valued, now it is a matter of working their way through, understanding what they need to achieve, and then most importantly driving the organization to it. Analytics can do a lot of things in a lot of ways, but drifting away from the value that the benefit is supposed to give them is something that often happens. Change management is an enormous problem, Linda and I talk about that a lot.

Linda Pawczuk: John, spot on. Get clear about the improvements, get clear about the purpose of the analytics, that is very important and to that end too, the change management component of it is the organization. As I said before, this is a business-led IT-type initiative. It is about the organization enabling itself around the operational changes that need to happen too as new information comes into play for decision making. Through the process of analytics, as you mentioned some more forward thinking; things like predictive analytics, most organizations today are not well positioned around structuring the operational capabilities, such they know how to process the content and process the data that they have, so John spot on. Understand the improvements that are needed.

Sean O’ Grady: My last question is for both of you, but John we are going to begin with you and it is about best practices. So if a carrier is going to try to implement analytics, how can they get the best return?

John Lucker: Getting the return, going back to Linda’s point earlier, has to do with organizational change management and I will tell you a story. Recently, I had a client who I knew was not getting the value that they were supposed to be getting from some of the analytic solutions that were developed and there was a moment with the CEO where we had the opportunity to say what are you going to do about that $20 million you are losing. What $20 million? Well, it gave us the opportunity to talk to them about change management, about the fact that people were reducing prices on certain things, but they were not raising the prices on certain things because it created an uncomfortable moment that they had to have with their customers. So there is an example of change management not allowing the analytics to speak. They were basically breaking the analytics in half and only using a part of it and causing actual damage to the result.
Sean O’ Grady: Linda your final thoughts on maximizing return from the implementation of analytics.

Linda Pawczuk: I think there are two points I would like to add. One is, as John talked again about change, but specifically you have got to get good at your data and as John mentioned, it is not about hiding behind having pure and clean data, but you have to get good at understanding the source and being patient to go through the process of converting that information and also in the process of validating the information in the backside. We hear so often, it is all about faster, faster, faster; of course we understand that. There is a lot of economic pressure, there is a lot of competitive pressure, but the reality is this faster, faster, faster does not necessarily mean better outcomes, so be patient. The second point that I think is important too and particularly as there are such rich information found in very unstructured places today as well, like in e-mail system, like in other capabilities that are in an organization, which lend well to various forms of analytics. So in bringing out some real value, you have to look beyond the obvious sources in the traditional capabilities and traditional systems that you have had in the past, so look beyond what is just in front of you.

Sean: Patience and keep your eyes open. You have been listening to Linda Pawczuk, a principal in Deloitte Consulting LLP and leader of the Insurance Technology practice. You have also been listening to John Lucker, a principal in Deloitte Consulting LLP and the leader of the Advanced Analytics & Modeling National Practice. If you would like to learn more about Linda, John, or any of the topics we discussed on this broadcast, you can find them and many more on our website. It is www.deloitte.com/insightsus

For all the good folks here in Insights, I am Sean O’Grady. We will you see next time.