“Dynatrace improved our response time to rate policies four times over.”
Leader – Service-Oriented Architecture Team, Westfield Insurance
Westfield Insurance is a part of Westfield Group, an insurance and banking group of businesses that’s been headquartered in Westfield Center, Ohio for more than 166 years. The company has $3.7 billion in combined assets and $1.4 billion in written premium annually. Westfield provides commercial insurance in 21 states, personal insurance in 10 states and surety services to customers in 31 states. As one of the nation’s 50 largest property and casualty insurance groups, the organization is represented by a network of more than 1,000 independent insurance agencies.
As an insurance provider, Westfield relies on a highly complex IT infrastructure that includes more than 200 enterprise web services in production working together with a range of middleware tools. Every step in maintaining policy and claims processes includes many individual transactional components that must all work as efficiently as possible to ensure the fastest and best end-user experience.
During the year, any slight dip in speed at any point in their web based policy quoting and issuance systems used by their network of independent agents resulted in slow response times—and was a big challenge for the IT team. It took time and required checking with many people from several different groups.
Bemiller says: “When we had issues in the past, we would receive calls from agencies telling us that our quoting application is slow. To troubleshoot, we would end up having to call four to six people—sometimes more. Since so many components were involved in each process, we didn’t know if the problem was caused by bad code in the application; or, if it was a services or a middleware performance issue; or, a policy processing system delay; or, even infrastructure, memory, CPU problems.”
“So, we would get a lot of people together in a room, work through triaging the issues and troubleshoot them until we figure them out,” he continues. “Then came what we call IMPACT.”
“IMPACT was a multi-year, multi-million dollar IT program of at least 10 individual projects. Its main focus was replacing our legacy claims management system with a new package,” says Bemiller. “To keep the functionality that our claims department needed, this required a significant amount of integration as well as the implementation of a significant number of other new systems, either developed in-house or additionally purchased.”
Westfield reduced the time it took to finish a claim by nearly half—in just a few weeks. Altogether, Westfield implemented more than 120 new enterprise services as a part of its IMPACT program. As a result, the company had to significantly extend its ESB capability to be able to handle more attachments and increased message content. Today the organization has integrated its management system with more than 20 external vendors that handle everyday insurance services for customers—such as renting a car or doing a damage estimate evaluation.
Westfield also implemented single sign-on technology so that cross system authentication tools worked faster and easier. It also created a new operational data store (ODS) for all claim-related policy data to verify customers’ coverage levels. However, with all of this updating and streamlining, the IT environment became even more complex and interconnected.
“Just as an example, when a customer calls in a claim, we refer to that as a first notice of loss,” explains Bemiller. “When this happens, it starts a series of activities that have to happen in tight coordination to record data about the claim and set it up.”
“During just that first notice of loss transaction, we use services that access the policy operational data store to validate coverage,” he says. “We create an electronic claim folder and add documents to that claim folder—all through services—that interact with the new electronic content management tools. We have multiple vendor integrations to help provide customers with rental cars or other services. With all of these steps happening at once with single sign-on and within just this one simple transaction, it was incredibly difficult to find the cause of any slow process. We couldn’t easily determine if the problem was a services, policy, ODS, ESB, or vendor issue. We had to call everyone to track down the problem.”
“To troubleshoot, we would end up having to call four to six people—sometimes more.”
“I knew that we needed an Application Performance Management (APM) solution for rapid issue triage and root cause isolation in our distributed environment,” Bemiller says. Although some performance statistics were available, Westfield needed deep level analysis, problem detection and alerting for every single transaction to ensure the level of service that supported their entire business.
“Dynatrace provides the answer to the five dimensions of APM as Gartner describes them,” says Bemiller. “It gives us deep-dive component monitoring; runtime application architecture modeling discovery, modeling and display; end-user experience monitoring; business transaction management; and, analytics and reporting.”
“The deep-dive Component monitoring, using ‘PurePaths’ is the most valuable part of Dynatrace—and it’s far more than just monitoring, says Bemiller. “Every single transaction happening in our production environment is recorded, providing a significant amount of data.”
“You can see SQL statements and values. You can see the XML messages being passed back and forth. You see the execution time for every single method happening in the call. In a few clicks you can actually get even more in-depth detail.”
“Right out of the box, Dynatrace visually shows the flow of information through the entire architecture for a single PurePath,” he says. “This didn’t require any configuration on our part to show you critical metrics at each hop along the way.”
The transaction management capabilities in Dynatrace allow Bemiller to take multiple system transactions and roll them up into one meaningful process for either business customers or support teams. Dynatrace shows how many times the process is completed in a day and gives the team a way to drill down to see successes, failures, response, times, what transactions are taking the longest. “This is a perfect tool to find, improve and streamline any process bottlenecks. It makes sense of any problems we face,” he adds.
“Analytics reporting with Dynatrace gives us historical data and enables us to create dashboards that we use to proactively monitor our enterprise every day, Bemiller says. “For example, we use dashboards for our middleware components that can quickly identify any slowness and then, pinpoint exactly what caused the delay so that we can eliminate it. We never had that kind of visibility before Dynatrace.”
“I knew that we needed an Application Performance Management (APM) solution for rapid issue triage and root cause isolation in our distributed environment.”
Using Dynatrace, Westfield reduced the time it took to finish a claim by nearly half—in just a few weeks.
“Now we see the queries that are being run, you can see how long they’re taking, you can see how many times they’re being run. You can see all of that information, so you fix the bottlenecks and then everybody’s happy, Bemiller says.
With Dynatrace, Westfield has replaced what he calls “troubleshooting by committee” with a command center process for their newly implemented claims management system. “Originally we had members of every team in a war-room-type scenario waiting for problems,” he says.” Today the we have four or five people in that room helping troubleshoot problems and Dynatrace is up on a monitor at all times and we’ve been using it to keep improving all the time.”